Texas Politics - Polling go back

1. #UTTribPoll: The Latest Analysis from the UT/Texas Tribune Poll
1.1 Most Recent UT/Texas Tribune Survey: February, 2014
2. Mood of the State
2.1 Trends
2.3 Most Important Problem Archive
3. Assessments of Political Leaders: October, 2009
3. Assessments of Political Leaders
4. 2014 Elections
5. Education Opinions
6. Immigration Opinions
7. Social Issue Opinions
8. National Policy Issue Opinions
9. Methodology, Data, and Graphical Archive
9.1 July 2008
9.2 October 2008
9.3 February-March 2009
9.4 June 2009
9.5 October 2009
9.6 February 2010
9.7 May 2010
9.8 September 2010
9.9 October 2010
9.10 February 2011
9.11 May 2011
9.12 October 2011
9.13 February 2012
9.14 May 2012
9.15 October 2012
9.16 February 2013
9.17 June 2013
9.18 October 2013
9.19 February 2014
10. About These Polls

1. #UTTribPoll: The Latest Analysis from the UT/Texas Tribune Poll

GOP Candidates Bet on Border Security

December 12, 2013: Republican candidates in Texas have figured out how to talk about immigration without stepping on political land mines: They talk about border security instead.

Obama Weighs Heavily on Texas Democrats

December 5, 2013: There is a reason so many Texas Republicans are mentioning the Democratic president in their commercials: He's unpopular with Republicans and moderates, and some Democrats have their reservations, too.

Women Through the Looking Glass

November 21, 2013: As a group, women have neither followed the March Hare to the Tea Party nor signed up for Wendy Davis' trip to Wonderland, leaving the campaigns to ponder their place on the electoral chessboard.

Cornyn and the Cruz Effect

November 15, 2013: Nothing captures the political impact of Ted Cruz's ascension in American politics like a comparison of public sentiment toward him and fellow Sen. John Cornyn in the wake of last month's government shutdown. Cornyn's ratings in the electoral sweet spot where conservatives and Tea Party Republicans overlap have fallen at the same time as Cruz's have risen.

What Early Poll Results Tell Us

November 6, 2013: The latest UT/TT Poll showing a single-digit lead for Greg Abbott over Wendy Davis in the gubernatorial race raised some eyebrows. Adding some context to a survey taken more than a year before Election Day helps provide some clarity on the results.

In Vote, Opposition Isn't Overflowing

October 24, 2013: Political chatter about a grassroots uprising against the water funding measure on the November ballot appears to be overblown. Polling indicates a fair amount of Tea Party support for that constitutional amendment.

Deflecting Immigration Questions

October 16, 2013: Uncomfortable questions about in-state tuition might prompt candidates like Greg Abbott to reach into Rick Perry's bag of tricks for an issue that addresses immigration issues without inflaming the wrong voters.

A Trickle of Votes for a Water Fund

October 15, 2013: Texas voter turnout is low, but for constitutional amendments like the one next month, turnout is often very, very low. So how do you figure out which poll respondents deserve your attention?

Clear Demographics, Unclear Politics

October 10, 2013: Sages in both political parties say they are the natural ideological allies of the rising Hispanic population in Texas. But while the demographic trends are undeniable, the political meaning behind them is cloudy.

Will the Tea Party Press John Cornyn?

October 2, 2013: Some in the Tea Party faction of the Texas GOP are encouraging talk of a challenge to U.S. Sen. John Cornyn, R-Texas, raising a question: Is he vulnerable to a challenge?

Education Could Test Both Parties

September 26, 2013: Education could be a tricky issue for gubernatorial candidates in 2014, with both the Democratic and Republican nominee having to navigate through unexpected cross-currents among their own constituencies.

Lies, Damn Lies and Campaign Polls

September 19, 2013: An ever-expanding niche market of political junkies - and the specialized media that feeds it - finds news in polling results and in conflicts over polling practice. The release of internal polls becomes as much about shaping public opinion as it is about measuring it.

Are Suburban Women Key to Democratic Resurgence?

September 12, 2013: While the Hispanic vote has been the focus of much of the analysis of Democrats' prospects for turning the Republican tide, in the short term, they will almost certainly need to look to suburban women - especially if Wendy Davis is at the top of the ticket.

Inventing Abbott

September 5, 2013: For all the advantages that have lent the feel of an unofficial coronation to his candidacy for governor, Attorney General Greg Abbott remains an undefined figure among many Texas voters, including as many as 40 percent of Republican primary voters.

How Holder May Help Abbott in 2014

August 29, 2013: Public opinion on voting rights in Texas neither paints a dour picture for gubernatorial hopeful Greg Abbott nor presents a clear path forward for Democrats.

How Abortion Could Shape a 2014 Abbott-Davis Contest

August 22, 2013: The return of abortion bills during the special sessions presents opportunities for both Greg Abbott and Wendy Davis to consolidate support and financial backers. But the choices they make could result in a potentially complicated general election dynamic.

The Irresistible Plot Arc of a Two-Party Texas

August 21, 2013: We look at how social media has buoyed Wendy Davis' star and added some turbulence to what must have seemed like a clear path to the governor's mansion for Greg Abbott.

Gridlock on NSA Surveillance Structured by Ambivalence

August 21, 2013: It's not surprising that the political class hasn't rallied to one side or the other on the debate over NSA surveillance; the public is sending oblique messages to its elected officials structured by ambivalence between the countervailing pulls of ideology and partisanship.

Perry's Whack at Obamacare

August 9, 2013: While most Americans do in fact have an opinion on the Affordable Care Act, to say that they understand it - at all, let alone "all too well" - runs contrary to the data currently available.

Joe Straus and the Other Republicans

August 8, 2013: Many statewide Republican candidates are running to the right to position themselves for the primaries, but the Speaker of the House enjoys a rarified position: An office with statewide reach that doesn't appear on the statewide ballot.

Patrick Ad Goes Fishing for GOP Primary Votes

August 5, 2013: Data from the University of Texas/Texas Tribune Poll suggests that the issues Sen. Dan Patrick invokes in the latest ad in his bid for lieutenant governor serve up very inviting bait for conservative voters, the big fish in GOP primary elections.

From Smooth Sailing to Traffic Jam

August 1, 2013: The regular Texas legislative session was notable for bipartisan coalitions and harmony. The special sessions have been notable for partisan battles and stalemates. To understand what's going on, just look at the voters.

Texas Voters, Congress and Immigration

July 18, 2013: We observe that Republican voters in Texas still have immigration and border security atop their lists of most important problems facing the state, and their sway over members of the Texas delegation will be noted by political colleagues and potential opponents alike.

Voter Positions on Ethics Seem Not to Affect Rick Perry's Political Position

The Morning After for Texas Democrats

July 12, 2013: A look at electoral returns and public opinion data helps explain why Democratic exuberance in the days after the Wendy Davis filibuster should be met with caution.

Voter Positions on Ethics Seem Not to Affect Rick Perry's Political Position

July 10, 2013: Texas voters are concerned about public ethics, and about some of the issues that have attached to the governor over the last 12 years - but they're partisan about it, and that has made all the difference for Rick Perry.

Will Dewhurst be Forced to Pay for the End of Session Missteps?

July 5, 2013: Lt. Gov. David Dewhurst currently enjoys the unenviable status of being the least popular of the major statewide Republican elected officials in Texas, according to an analysis of his approval numbers and standing in election matchups.

Does the Supreme Court's Decision to Invalidate DOMA put the GOP in an Immigration-Like Position?

July 3, 2013: A brief analysis of public opinion on gay marriage in Texas puts recent comments by Sen. Ted Cruz and Agriculture Commission and Lt. Gov. hopeful Todd Staples at odds with an electorate growing increasingly comfortable with same-sex marriage.

Abortion Attitudes in Texas Help Explain Special Session Failure of SB 5

July 1, 2013: Despite Texas' place as a reliably conservative state, public opinion on abortion is far from monolithic. As we write here, Texans display attitudes supporting stricter abortion regulations (regardless of question wording), but much less of an appetite for an outright prohibition on the procedure. This disconnect, between the perception of an overwhelmingly conservative Texas and the realities of public opinion, no doubt fueled the surprise legislators felt when overwhelmed by the Senate gallery last week, ultimately assisting in the defeat of SB 5 at the end of the Legislature's first special session.

Perry Looks Strong, But Not Only Option in 2014...or 2016 for Texas GOP

June 27, 2013: As Texans await Rick Perry's electoral plans, we take a look at what our most recent polling says about the political landscape he will likely face here in Texas should he choose to run, finding that though in a strong position, he may not find himself the default choice when GOP voters finally enter the primary voting booth.

Cruz Leads Perry, Rest of Field in Early Sounding of 2016 GOP Presidential Primary

The June 2013 University of Texas/Texas Tribune Poll Found US Senator Ted Cruz leading a field of potential challengers, including Governor Rick Perry, for the 2016 Presidential nomination. This and other results are being released the week of June 17 by The Texas Tribune. Check back here for more graphics and analysis from The Texas Politics Project staff as the week progresses.

Budget Endgames and Public Opinion

May 21, 2013: The tripartisanship evident during this week's final budget negotiations pits Democrats, Republicans, and Tea Partiers with vastly different priorities into a game of chicken. Despite the fact that the public doesn't closely follow the budget process, the broad positions of these groups in the Legislature neatly reflect the preferences of those groups in the electorate according to polling data collected at the beginning of the session focusing on the public's perception of the legislative priorities.

Dueling Attitudes on Guns

May 16, 2013: In the context of the gun control debate currently taking place at both the federal and state levels, to say that the Senate was ignoring public opinion in tabling the background check measure - or that the Legislature is accurately reflecting it in expanding gun access - is to ignore the intricacies and ambivalence in public opinion on gun control.

Immigration Reform and the GOP's Self-Interest

May 8, 2013: While many GOP leaders argue that passing comprehensive immigration reform is in the GOP's best interest, some data suggest that the long-term interest of party strategists and the short-term self-interest of members of Congress are not necessarily in sync.

The Calculus of Women's Health

May 3, 2013: While women's health may not have garnered the attention that it did in the 2011 session, here we discuss the connection between women's health and abortion and how the removal of Planned Parenthood from the women's health calculus may have made passing restrictive abortion legislation more difficult in 2013.

Rough Sailing for Water Policy

April 30, 2013: In the wake of last night's failed attempt to pass a comprehensive water plan - a priority of the governor, among many other political elites - Jim Henson takes a look at what public opinion can tell us about why water funding has faced so much difficulty despite the elite consensus behind the issue.

The Tea Party as a Constituency

April 19, 2013: The power and pull of the Tea Party inside the national, and in this case Texas, GOP is a source of interest for many insiders and some outsiders alike. In our latest piece, we examine how the Tea Party has become an acknowledged but understood constituency within the Texas Republican Party, and as a result, less of a force in this legislative session.

Guest Worker Program the Sweet Spot for Texas GOP

April 16, 2013: As an immigration reform package gets unveiled in the Senate today, our polling data suggests that, at least here in Texas, anything even resembling amnesty is a poison pill for Republican voters.

Tea Party Identifiers Outside the Mainstream on Gun Control Proposals

April 11, 2013: As the Senate opens up debate on gun control legislation, it appears that Tea Party members in particular register extreme opposition to these new proposals, far to the right of most Americans, and even most other Texans. Might this help explain the opposition of some Tea Party standard bearers?

Texans Reluctant to Change Conceal Carry Requirements

April 9, 2013: Texans hold ideologically conservative attitudes on most gun control proposals, but literally conservative attitudes on those proposals that seek to reduce requirements for carrying a concealed weapon, a topic that we explore here.

According to Polling Data, Budget Fights Track Closely to Public Opinion

April 4, 2013: If you're really hip like us, you're probably watching the budget fight on the house floor today (either in person, online, or like me, a little of both). See what public opinion can tell us about the budget process so far and why it might get a little more contentious going forward.

No Evidence of Education Backlash Despite Cuts from Last Session

April 2, 2013: Despite all the rhetoric of the calamity that would befall legislators who supported cutting education by $5.4 billion in 2011, we find little evidence of an electorate overwhelming displeased with the Texas education system.

GOP Disapproval of Obama

March 28, 2013: We just think that this is a particularly nifty, and potentially telling, graphic of Texas GOP attitudes about the president.

Shifting Attitudes on Gay Marriage in Texas

March 26, 2013: Gay marriage attitudes in Texas are shifting with the rest of the nation, and this shift poses a potential problem for the Republicans here in Texas as independents, women, suburbanites, and the youngest age cohorts become increasingly support of same sex unions.

1.1 Most Recent UT/Texas Tribune Survey: February, 2014

The February, 2014 University of Texas/Texas Tribune Poll found Attorney General Greg Abbott with an 11-point advantage over State Senator Wendy Davis in the election for governor to be held in November. The survey also asked about a variety of attitudes likely to inform the 2014 elections, including: education, abortion, and immigration, among others.

The University of Texas/Texas Tribune internet survey of 1,200 registered voters was conducted Feb. 7-17 and has an overall margin of error of +/- 2.83 percentage points. For questions of likely Republican primary voters, the margin of error is +/- 4.56 percentage points; for likely Democratic primary voters, +/- 6.04 percentage points.

2. Mood of the State

Texans' assessment of the direction of the economy and general path of both the US and Texas show continued concern about both personal situations and the state of the national economy in the February 2013 University of Texas/Texas Tribune poll. Views of whether the state was moving in the right direction or was on the wrong track flipped between May and October and then continued that trend in February with a sharp drop in people expressing no opinion. While in October, 23 percent of respondents had no opinion on whether the state was going in the right direction or on the wrong track, in February, that number had dropped to 16 percent. Of those with an opinion, 45 percent say that the state is headed in the right direction while 39 percent say that we're off on the wrong track. This result marked a slight contraction of the net positive judgment since October, as this graphic tracking trends on this item illustrates.

Following a well-established trend, assessments of the direction of the country were more negative than evaluations of Texas, but were slightly more dour in February 2013 than in October. At 62 percent, the wrong track number for February was a slight increase from October's 58 percent, while the right direction number decreased from 31 to 29 percent.

Perceptions of the national economy remained in net negative territory in the October 2012 poll - 42 percent thought the national economy was worse off than a year ago compared with 32 percent who thought it was better off -- but results indicate an increase in both those who say the economy has gotten better and those who say it has gotten worse since May 2012. (See the trends plotted in this feature.)

2.1 Trends

Retrospective Assessments of the National Economy

Texans' views on the state of the national economy are improving slowly and gradually, as the longitudinal chart comparing assessments as of October 2010 to previous assessments shows. Fifty-six percent of respondents answered that the national economy is worse off than it was a year ago, slightly more than (and statistically very close to) September's 52% , but overall, part of a steady decline for the peak of 86% in October 2008.. Conversely, 28% of respondents answered that the national economy is better off than it was a year ago, a marked increase from 2% (!) in the first July 2008 University of Texas poll.

Retrospective Assessments of Personal Economic Situations

No such clear trend can be discerned from respondents with respect to their personal economic situations. Much like in previous polls, large and statistically equal percentages of respondents say they and their families are either economically worse off or about the same as they were one year ago, while a significantly smaller percentage of respondents say they and their families are better off than they were one year ago.

Right Direction / Wrong Track

The primary characteristic of the responses to the "right direction/wrong track" item continues to be the gap between Texans' perception of the trajectories of the United States and of Texas.

2.3 Most Important Problem Archive

In every poll that we have conducted for the Texas Politics Project (and The Texas Tribune ), we have asked respondents to name the most important problem facing the country and the most important problem facing Texas. For the first three polls that were conducted, these questions were open-ended -- meaning that respondents could write in their responses, which were then coded. Since June 2009, we have provided respondents with a long list of possible answers to both questions. In general, the items on the lists have been consistent, though we have occasionally added items that had recently emerged onto the national or state scene (e.g., the gulf oil spill or Texas's budget shortfall) and deleted items that had become irrelevant.

Results for both the national and state questions have been fairly consistent over time, though there have also been some significant changes. With respect to the most important problem facing the country, answer choices related to the economy or unemployment have always been the most popular. During the initial surveys in 2008 and early 2009, answer choices related to energy issues (gas prices, energy independence, etc.) as well as national security issues (the war in Iraq, terrorism, etc.) were also salient to our respondents, but they have since receded in importance. In their place, concerns about federal spending and the national debt, as well as about political corruption and leadership in Washington, have risen dramatically.

With respect to the most important problem facing Texas, answer choices related to the economy have also been very popular, but they have generally been equally popular to (and have at times been eclipsed by) answer choices related to immigration and border security. In recent surveys, the state's budget shortfall and education have also become frequently-chosen answers.

3. Assessments of Political Leaders: October, 2009

President Obama's ratings differed little from his performance in the general election, with 41 approving and 52 disapproving of his overall job performance. The US Congress job approval rating from Texans was the most negative of all of the assessments: 71% disapproved and only 13% approved. In both of these assessments the most negative option, "disapprove strongly", was the most common choice: 44% strongly disapproved of Obama, and 49% strongly disapproved of Congress.

At the state level, where we solicited approval ratings of Governor Perry, Senator Hutchison, and the Texas state legislature, assessments were less negative though still tepid. Governor Perry's approval numbers were at 36% approve (10% strongly/26% somewhat) versus 44% disapprove (26% strongly / 18% somewhat). Senator Hutchison elicited comparable positives with a 39% approval rating (10% strongly / 29% somewhat), but her disapproval rating was only 27% (10% strongly/17% somewhat) with a larger share choosing the neutral option (24%) than in the Governor's case (15%).

The Texas state legislature fared much better than their counterparts in Washington DC: 31% approved of the way the Texas Legislature has been handling its job (3% strongly/28% somewhat). Their state legislature's disapproval the 36% (15% stronglydisapproving/21% somewhat disapproving).

3. Assessments of Political Leaders

The February 2014 UT/Texas Tribune survey finds voters becoming more familiar with and more positively disposed towards gubernatorial candidate Greg Abbott, with the Democrat Davis remaining static in most Texans' minds.

The poll also tested Texans' attitudes towards a number of primary and potential general election candidates, including David Dewhurst, Dan Patrick, Todd Staples, Jerry Patterson and Leticia Van de Putte.

4. 2014 Elections

The February 2014 UT/Texas Tribune Poll took a close look at the upcoming primary elections and the marquee match-ups of the 2014 race. The poll found State Senator and Democratic Gubernatorial Candidate Wendy Davis trailing Attorney General and Republican Gubernatorial Candidate Greg Abbott by 11-points in a general election match-up, with 17 percent of voters still undecided. The poll also found current Lieutenant Governor David Dewhurst running ahead of his three challengers, but facing a likely run off with State Senator Dan Patrick.

5. Education Opinions

While we ask recurring questions to assess Texans' attitudes on the quality of public K-12 schools, the quality of higher education, and the adequacy of education spending here in Texas, in the February 2013 University of Texas/Texas Tribune Poll, we also asked specific questions about proposals that would allow individuals to carry concealed handguns on college campuses and K-12 teachers the ability to carry concealed handguns in their schools.

6. Immigration Opinions

According to our respondents, immigration and border security are the most important problems facing Texas as indicated by our recent surveys. We regularly assess support for comprehensive immigration reform with a pathway to citizenship and whether illegal immigrants should pay in-state or out-of-state tuition when attending Texas colleges and universities. In response to the ongoing debate, in the February 2013 University of Texas / Texas Tribune Survey, we also asked respondents whether they would favor a guest worker program.

7. Social Issue Opinions

In The University of Texas / Texas Tribune Survey, we have regularly asked questions on abortion, same-sex marriage, and respondents' views toward the death penalty, in addition to other social/moral issues that come up from time to time.

8. National Policy Issue Opinions

In the University of Texas/Texas Tribune survey, in addition to our primary focus on Texas, we also ask respondents about pertinent national policy issues to see where Texans stand in today's most important debates, and when possible, how they compare to the nation as a whole. In the October 2013 UT/Texas Tribune survey, we asked a battery of questions about the Affordable Care Act and its constituent parts.

9. Methodology, Data, and Graphical Archive

The sub-sections that follow contain code books, data files, cross tabs, and summaries for all of the polls organized chronologically.

The link on the right side of this page will load an excerpt of an interview with Daron Shaw, the polling director for the Texas Politics project and the UT-Austin/Texas Tribune poll, conducted by Jason Embry of the Austin American-Statesman. For a link to the entire podcast, which also discusses poll results and the 2010 election in Texas, see Embry's "First Reading" blog. (The Statesman compiles blog entries at that URL. If you search the page and can't find the entire podcast, you may need search the site for an archived version.)

9.1 July 2008

The July, 2008 UT-Austin Texas Politics Poll was designed by researchers at UT-Austin and conducted by YouGovPolimetrix, a firm with demonstrated success in internet polling. YouGovPolimetrix accomplishes internet polling through a unique sampling procedure known as "matched random sampling." The firm begins with two lists: (1) a list of all consumers in Texas (covering approximately 95 percent of the adult population), and (2) a list of people who have agreed to take YouGovPolimetrix's surveys. For each list, Polimetrix has an extensive set of demographics.

The sampling procedure then progresses in two stages. First, a random sample of consumers is drawn. For each person drawn from this sample a list of key demographics is recorded. In essence, each individual drawn is represented as a cluster of demographic characteristics, including age, income, education, race, gender, longitude and latitude, etc. Second, YouGovPolimetrix uses a matching algorithm to find the PollingPoint panelist who is the closest match to the person drawn off the consumer file. In this way an entire matched random sample is constructed for all people in the sample.

The current poll of 800 adult Texans has a margin of error of +/- 3.46 percentage points at the 95% confidence level. The poll includes interviews with 677 registered voters, with an attendant margin of error of +/- 3.77 percentage points. Response rates are almost 100% given the matching methodology. The YouGovPolimetrix pool includes people who are much less likely to have access to the Internet or a personal computer. They have been especially assiduous at enlisting people with lower incomes and ethnic and racial minorities, part of an attempt to bolster the representativeness of their samples.The poll was administered by YouGov/Polimetrix. Polimetrix interviewed 899 respondents who were then matched down to a sample of 800 to produce the final data set. The respondents were matched on gender, age, race, education, party identification and political interest. YouGov/Polimetrix then weighted the matched set of survey respondents to known marginals for the general population of Texas from the 2006 American Community Survey. Those marginals are:

Age:

18-34: 33.45%

35-54: 39.08%

55+: 27.53%

Gender:

Male: 48.97%

Female: 51.03%

Race:

White/Other: 57.63%

Black: 10.94%

Hispanic: 31.43%

Education:

HS or less: 49.23%

Some College: 28.49%

College Graduate: 15.23%

Post-graduate: 7.50%

9.2 October 2008

The October 2008 UT-Austin Texas Politics Poll was designed by researchers at UT-Austin and conducted by YouGov/Polimetrix, a firm with demonstrated success in internet polling. YouGov/Polimetrix accomplishes internet polling through a unique sampling procedure known as "matched random sampling." The firm begins with two lists: (1) a list of all adult "consumers" in Texas (covering approximately 95 percent of the adult population), and (2) a list of people who have agreed to take YouGov/Polimetrix's surveys. For each list, Polimetrix has an extensive set of demographics.

The sampling procedure then progresses in two stages. First, a random sample of consumers is drawn. For each person drawn from this sample a list of key demographics is recorded. In essence, each individual drawn is represented as a cluster of demographic characteristics, including age, income, education, race, gender, longitude and latitude, etc. Second, YouGov/Polimetrix uses a matching algorithm to find the PollingPoint panelist who is the closest match to the person drawn off the consumer file. In this way an entire "matched" random sample is constructed for all people in the "drawn" sample.

The October 2008 poll consists of 613 adult Texans, and has a margin of error of +/- 3.98 percentage points at the 95% confidence level. The poll includes interviews with 550 registered voters, with an attendant margin of error of +/- 4.20 percentage points. Response rates are almost 100% given the matching methodology. The YouGov/Polimetrix pool includes people who are much less likely to have access to the Internet or a personal computer. YouGov/Polimetrix has been especially assiduous at enlisting ethnic and racial minorities, as well as people who are less affluent, as part of their attempt to ensure the representativeness of their samples.The poll was administered by YouGov/Polimetrix. Polimetrix interviewed 899 respondents who were then matched down to a sample of 613 to produce the final data set. The respondents were matched on (among other items) gender, age, race, education, party identification and political interest. YouGov/Polimetrix then weighted the matched set of survey respondents to known marginals for the general population of Texas from the 2006 American Community Survey. Those marginals are:

Age:

18-34: 33.45%

35-54: 39.08%

55+: 27.53%

Gender:

Male: 48.97%

Female: 51.03%

Race:

White/Other: 57.63%

Black: 10.94%

Hispanic: 31.43%

Education:

HS or less: 49.23%

Some College: 28.49%

College Grad.: 15.23%

Post-graduate: 7.50%

9.3 February-March 2009

The March 2009 UT-Austin Texas Politics Poll was designed by researchers at UT-Austin and conducted by YouGov/Polimetrix, a firm with demonstrated success in internet polling. YouGov/Polimetrix accomplishes internet polling through a unique sampling procedure known as "matched random sampling." The firm begins with two lists: (1) a list of all adult "consumers" in Texas (covering approximately 95 percent of the adult population), and (2) a list of people who have agreed to take YouGov/Polimetrix's surveys. For each list, Polimetrix has an extensive set of demographics.

The sampling procedure then progresses in two stages. First, a random sample of consumers is drawn. For each person drawn from this sample a list of key demographics is recorded. In essence, each individual drawn is represented as a cluster of demographic characteristics, including age, income, education, race, gender, longitude and latitude, etc. Second, YouGov/Polimetrix uses a matching algorithm to find the PollingPoint panelist who is the closest match to the person drawn off the consumer file. In this way an entire "matched" random sample is constructed for all people in the "drawn" sample.

The March 2009 poll consists of 800 adult Texans, and has a margin of error of +/- 3.46 percentage points at the 95% confidence level. The poll includes interviews with 715 registered voters, with an attendant margin of error of +/- 3.66 percentage points. Response rates are almost 100% given the matching methodology. The YouGov/Polimetrix pool includes people who are much less likely to have access to the Internet or a personal computer. YouGov/Polimetrix has been especially assiduous at enlisting ethnic and racial minorities, as well as people who are less affluent, as part of their attempt to ensure the representativeness of their samples. Surveys were completed between February 24 and March 6, 2009.The poll was administered by YouGov/Polimetrix. Polimetrix interviewed 899 respondents who were then matched down to a sample of 800 to produce the final data set. The respondents were matched on gender, age, race, education, party identification and political interest. YouGov/Polimetrix then weighted the matched set of survey respondents to known marginals for the general population of Texas from the 2006 American Community Survey. Those marginals are:

Age:

18-34: 33.45%

35-54: 39.08%

55+: 27.53%

Gender:

Male: 48.97%

Female: 51.03%

Race:

White/Other: 57.63%

Black: 10.94%

Hispanic: 31.43%

Education:

HS or less: 49.23%

Some College: 28.49%

College Graduate: 15.23%

Post-graduate: 7.50%

Surveys were completed between February 25 and March 6.

9.4 June 2009

The June 2009 UT-Austin Texas Politics Poll was designed by researchers in the UT-Austin Department of Government and conducted by YouGov/Polimetrix, a firm with demonstrated success in internet polling. YouGov/Polimetrix accomplishes internet polling through a unique sampling procedure known as "matched random sampling." The firm begins with two lists: (1) a list of all adult "consumers" in Texas (covering approximately 95 percent of the adult population), and (2) a list of people who have agreed to take YouGov/Polimetrix's surveys. For each list, Polimetrix has an extensive set of demographics.

The sampling procedure then progresses in two stages. First, a random sample of consumers is drawn. For each person drawn from this sample a list of key demographics is recorded. In essence, each individual drawn is represented as a cluster of demographic characteristics, including age, income, education, race, gender, longitude and latitude, etc. Second, YouGov/Polimetrix uses a matching algorithm to find the PollingPoint panelist who is the closest match to the person drawn off the consumer file. In this way an entire "matched" random sample is constructed for all people in the "drawn" sample.

The June 2009 poll consists of 924 adult Texans, and has a margin of error of +/- 3.22 percentage points at the 95% confidence level. The poll includes interviews with 791 registered voters, with an attendant margin of error of +/- 3.66 percentage points. Response rates are almost 100% given the matching methodology. The YouGov/Polimetrix pool includes people who are much less likely to have access to the Internet or a personal computer. YouGov/Polimetrix has been especially assiduous at enlisting ethnic and racial minorities, as well as people who are less affluent, as part of their attempt to ensure the representativeness of their samples. Surveys were completed between June 11 and June 22, 2009.

Polimetrix interviewed 924 respondents who were then matched down to a sample of 800 to produce the final data set. The respondents were matched on gender, age, race, education, party identification and political interest. YouGov/Polimetrix then weighted the matched set of survey respondents to known marginals for the general population of Texas from the 2006 American Community Survey. Those marginals are:

Age:

18-34: 33.45%

35-54: 39.08%

55+: 27.53%

Gender:

Male: 48.97%

Female: 51.03%

Race:

White/Other: 57.63%

Black: 10.94%

Hispanic: 31.43%

Education:

HS or less: 49.23%

Some College: 28.49%

College Graduate: 15.23%

Post-graduate: 7.50%

9.5 October 2009

The October 2009 Texas Tribune/UT-Austin Texas Politics Poll was designed by researchers at UT-Austin and conducted by YouGov/Polimetrix, a firm with demonstrated success in internet polling. YouGov/Polimetrix accomplishes internet polling through a unique sampling procedure known as "matched random sampling." The firm begins with two lists: (1) a list of all adult "consumers" in Texas (covering approximately 95 percent of the adult population), and (2) a list of people who have agreed to take YouGov/Polimetrix's surveys. For each list, Polimetrix has an extensive set of demographics.

The sampling procedure then progresses in two stages. First, a random sample of consumers is drawn. For each person drawn from this sample a list of key demographics is recorded. In essence, each individual drawn is represented as a cluster of demographic characteristics, including age, income, education, race, gender, longitude and latitude, etc. Second, YouGov/Polimetrix uses a matching algorithm to find the PollingPoint panelist who is the closest match to the person drawn off the consumer file. In this way an entire "matched" random sample is constructed for all people in the "drawn" sample.

The October 2009 poll consists primarily of 800 adults who are registered voters in Texas, and has a margin of error of +/-3.46 percentage points at the 95% confidence level. The YouGov/Polimetrix pool includes people who are much less likely to have access to the Internet or a personal computer. YouGov/Polimetrix has been especially assiduous at enlisting ethnic and racial minorities, as well as people who are less affluent, as part of their attempt to ensure the representativeness of their samples. Surveys were completed between October 20 and October 27, 2009. Polimetrix interviewed 1152 respondents who were then matched down to a sample of 800 to produce the final dataset. The respondents were matched ongender, age, race, education, party identification, ideology and political interest.

Polimetrix then weighted the matched set of survey respondents to known marginals for the registered voters of Texas from the 2008 Current Population Survey. Those marginals are shown below.

Age:

18-34: 33.45%

35-54: 39.08%

55+: 27.53%

Gender:

Male: 48.97%

Female: 51.03%

Race: White/Other: 57.63%

Black: 10.94%

Hispanic: 31.43%

Education:

HS or less: 49.23%

Some College: 28.49%

College Graduate: 15.23%

Post-graduate: 7.50%

Because Hank Gilbert was inadvertently left off the list of Democratic gubernatorial nominees in the original round of surveying, YouGov/Polimetrix re-interviewed 269 respondents to the October 2009 survey who had indicated that they intended to vote in the Texas Democratic gubernatorial primary. These respondents were then weighted to the distribution of Democratic gubernatorial primary voters in the initial survey on the following variables: gender, race, age, education, party ID, political interest, ideology, and national origin.

9.6 February 2010

Sampling and Weighting Methodology for the February 2010 Texas Statewide Study

Survey Panel Data

The PollingPoint panel, a proprietary opt-in survey panel, is comprised of 1.6 million U.S. residents who have agreed to participate in YouGov Polimetrix's Web surveys. At any given time, YouGov Polimetrix maintains a minimum of five recruitment campaigns based on salient current events.

Panel members are recruited by a number of methods and on a variety of topics to help ensure diversity in the panel population. Recruiting methods include Web advertising campaigns (public surveys), permission-based email campaigns, partner sponsored solicitations, telephone-to-Web recruitment (RDD based sampling), and mail-to-Web recruitment (Voter Registration Based Sampling).

The primary method of recruitment for the PollingPoint Panel is Web advertising campaigns that appear based on keyword searches. In practice, a search in Google may prompt an active PollingPoint advertisement soliciting opinion on the search topic. At the conclusion of the short survey respondents are invited to join the PollingPoint panel in order to receive and participate in additional surveys. After a double opt-in procedure, where respondents must confirm their consent by responding to an email, the database checks to ensure the newly recruited panelist is in fact new and that the address information provided is valid.

Additionally, YouGov Polimetrix augments their panel with difficult to recruit respondents by soliciting panelists in telephone and mail surveys. For example, in the fall and winter of 2006, YouGov Polimetrix completed telephone interviews using RDD sampling and invited respondents to join the online panel. Respondents provided a working email where they could confirm their consent and request to receive online survey invitations. YouGov Polimetrix also employed registration based sampling, inviting respondents to complete a pre-election survey online. At the conclusion of that survey, respondents were invited to become PollingPoint members and receive additional survey invitations at their email address.

The PollingPoint panel currently has over 55,000 active panelists who are registered voters in Texas. These panelists cover a wide range of demographic characteristics.

Sampling and Sample Matching

Sample matching is a methodology for selection of "representative" samples from non-randomly selected pools of respondents. It is ideally suited for Web access panels, but could also be used for other types of surveys, such as phone surveys. Sample matching starts with an enumeration of the target population. For general population studies, the target population is all adults, and can be enumerated through the use of the decennial Census or a high quality survey, such as the American Community Survey. In other contexts, this is known as the sampling frame, though, unlike conventional sampling, the sample is not drawn from the frame. Traditional sampling, then, selects individuals from the sampling frame at random for participation in the study. This may not be feasible or economical as the contact information, especially email addresses, is not available for all individuals in the frame and refusals to participate increase the costs of sampling in this way.

Sample selection using the matching methodology is a two-stage process.

First, a random sample is drawn from the target population. We call this sample the target sample. Details on how the target sample is drawn are provided below, but the essential idea is that this sample is a true probability sample and thus representative of the frame from which it was drawn. Second, for each member of the target sample, we select one or more matching members from our pool of opt-in respondents. This is called the matched sample. Matching is accomplished using a large set of variables that are available in consumer and voter databases for both the target population and the opt-in panel.

The purpose of matching is to find an available respondent who is as similar as possible to the selected member of the target sample. The result is a sample of respondents who have the same measured characteristics as the target sample. Under certain conditions, described below, the matched sample will have similar properties to a true random sample. That is, the matched sample mimics the characteristics of the target sample. It is, as far as we can tell, "representative" of the target population (because it is similar to the target sample).

When choosing the matched sample, it is necessary to find the closest matching respondent in the panel of opt-ins to each member of the target sample. Polimetrix employs the proximity matching method to find the closest matching respondent. For each variable used for matching, we define a distance function, d(x,y), which describes how "close" the values x and y are on a particular attribute. The overall distance between a member of the target sample and a member of the panel is a weighted sum of the individual distance functions on each attribute. The weights can be adjusted for each study based upon which variables are thought to be important for that study, though, for the most part, we have not found the matching procedure to be sensitive to small adjustments of the weights. A large weight, on the other hand, forces the algorithm toward an exact match on that dimension.

Sampling Frame and Target Sample

YouGov/Polimetrix constructed a national sampling frame from the 2007 American Community Survey, including data on age, race, gender, education, marital status, number of children under 18, family income, employment status, citizenship, state, and metropolitan area. The frame was constructed by stratified sampling from the full 2007 ACS sample with selection within strata by weighted sampling with replacements (using the person weights on the public use file). Data on voter registration status and turnout were matched to this frame using the November 2006 Current Population Survey. Data on interest in politics and party identification were then matched to this frame from the 2007 Pew Religious Life survey, using the following variables for the match: age, race, gender, education, marital status, number of children under 18, family income, employment status, citizenship, state. The target sample of 800 Texas registered voters was selected with stratification by age, race, gender, education, and with simple random sampling within strata.

Weighting

Because matching is approximate, rather than exact, and response rates vary by group, the sample of completed interviews normally shows small amounts of imbalance that can be corrected by post-stratification weighting.

Raking, first proposed by Deming and Stephan (1940), adjusts an initial set of weights to match a known set of population marginals, using a method of iterative proportional fitting (see Bishop, Fienberg and Holland, 1975 for details). In this procedure, the weights are adjusted sequentially to match the marginal distribution of each weight variable. The process proceeds until all marginals are matched. It does not require any information about the joint distribution of the variables (though, if these data are available and believed to be important, they can be employed by defining a marginal distribution involving a cross-classification of two variables).

You Gov Politmetrix calculated post-stratification weights by raking the completed interviews to known marginals for the general population of Texas from the November 2006 Current Population Survey and Pew Religious Life survey for the following variables: age, race, gender, education, and ideology.

9.7 May 2010

Sampling and Weighting Methodology for the February 2010 Texas Statewide Study

Sampling Frame and Target Sample

YouGov/Polimetrix constructed a national sampling frame from the 2007 American Community Survey, including data on age, race, gender, education, marital status, number of children under 18, family income, employment status, citizenship, state, and metropolitan area. The frame was constructed by stratified sampling from the full 2007 ACS sample with selection within strata by weighted sampling with replacements (using the person weights on the public use file). Data on voter registration status and turnout were matched to this frame using the November 2006 Current Population Survey. Data on interest in politics and party identification were then matched to this frame from the 2007 Pew Religious Life survey, using the following variables for the match: age, race, gender, education, marital status, number of children under 18, family income, employment status, citizenship, state. The target sample of 800 Texas registered voters was selected with stratification by age, race, gender, education, and with simple random sampling within strata.

Weighting

Because matching is approximate, rather than exact, and response rates vary by group, the sample of completed interviews normally shows small amounts of imbalance that can be corrected by post-stratification weighting.

Raking, first proposed by Deming and Stephan (1940), adjusts an initial set of weights to match a known set of population marginals, using a method of iterative proportional fitting (see Bishop, Fienberg and Holland, 1975 for details). In this procedure, the weights are adjusted sequentially to match the marginal distribution of each weight variable. The process proceeds until all marginals are matched. It does not require any information about the joint distribution of the variables (though, if these data are available and believed to be important, they can be employed by defining a marginal distribution involving a cross-classification of two variables).

Post-stratification weights are calculated by raking the completed interviews to known marginals for the general population of Texas from the November 2006 Current Population Survey and Pew Religious Life survey for the following variables: age, race, gender, and education.

9.8 September 2010

For the University of Texas / Texas Tribune survey, YouGovPolimetrix interviewed 906 respondents between September 3 and 8, 2010, who were then matched down to a sample of 800 to produce the final dataset. The respondents were matched on gender, age, race, education, party identification, ideology and political interest. YouGovPolimetrix then weighted the matched set of survey respondents to known marginals for the registered voters of Texas from the 2008 Current Population survey and the 2007 Pew Religious Life Survey.

Sampling Frame and Target Sample

YouGovPolimetrix constructed a national sampling frame from the 2007 American Community Survey, including data on age, race, gender, education, marital status, number of children under 18, family income, employment status, citizenship, state, and metropolitan area. The frame was constructed by stratified sampling from the full 2007 ACS sample with selection within strata by weighted sampling with replacements (using the person weights on the public use file). Data on voter registration status and turnout were matched to this frame using the November 2008 Current Population Survey. Data on interest in politics and party identification were then matched to this frame from the 2007 Pew Religious Life survey, using the following variables for the match: age, race, gender, education, marital status, number of children under 18, family income, employment status, citizenship, state. The target sample of 800 Texas registered voters was selected with stratification by age, race, gender, education, and with simple random sampling within strata.

Weighting

Because matching is approximate, rather than exact, and response rates vary by group, the sample of completed interviews normally shows small amounts of imbalance that can be corrected by post-stratification weighting. Raking, first proposed by Deming and Stephan (1940), adjusts an initial set of weights to match a known set of population marginals, using a method of iterative proportional fitting (see Bishop, Fienberg and Holland, 1975 for details). In this procedure, the weights are adjusted sequentially to match the marginal distribution of each weight variable. The process proceeds until all marginals are matched. It does not require any information about the joint distribution of the variables (though, if these data are available and believed to be important, they can be employed by defining a marginal distribution involving a cross-classification of two variables). Post-stratification weights are calculated by raking the completed interviews to known marginals for the general population of Texas from the November 2008 Current Population Survey for the following variables: age, race, gender, and education.

9.9 October 2010

Sampling and Weighting Methodology for the October 2010 Texas Statewide Study

For the survey, YouGovPolimetrix interviewed 914 respondents between October 11 and 19, 2010, who were then matched down to a sample of 800 to produce the final dataset. The respondents were matched on gender, age, race, education, party identification, ideology and political interest. YouGovPolimetrix then weighted the matched set of survey respondents to known marginals for the registered voters of Texas from the 2008 Current Population survey and the 2007 Pew Religious Life Survey.

Sampling Frame and Target Sample

YouGovPolimetrix constructed a national sampling frame from the 2007 American Community Survey, including data on age, race, gender, education, marital status, number of children under 18, family income, employment status, citizenship, state, and metropolitan area. The frame was constructed by stratified sampling from the full 2007 ACS sample with selection within strata by weighted sampling with replacements (using the person weights on the public use file). Data on voter registration status and turnout were matched to this frame using the November 2008 Current Population Survey. Data on interest in politics and party identification were then matched to this frame from the 2007 Pew Religious Life survey, using the following variables for the match: age, race, gender, education, marital status, number of children under 18, family income, employment status, citizenship, state. The target sample of 800 Texas registered voters was selected with stratification by age, race, gender, education, and with simple random sampling within strata.

Weighting

Because matching is approximate, rather than exact, and response rates vary by group, the sample of completed interviews normally shows small amounts of imbalance that can be corrected by post-stratification weighting. Raking, first proposed by Deming and Stephan (1940), adjusts an initial set of weights to match a known set of population marginals, using a method of iterative proportional fitting (see Bishop, Fienberg and Holland, 1975 for details). In this procedure, the weights are adjusted sequentially to match the marginal distribution of each weight variable. The process proceeds until all marginals are matched. It does not require any information about the joint distribution of the variables (though, if these data are available and believed to be important, they can be employed by defining a marginal distribution involving a cross-classification of two variables). Post-stratification weights are calculated by raking the completed interviews to known marginals for the general population of Texas from the November 2008 Current Population Survey for the following variables: age, race, gender, and education.

9.10 February 2011

Sampling and Weighting Methodology for the February 11 Texas Statewide Study

For the February 2011 survey, YouGovPolimetrix interviewed 963 respondents between February 9-18 2011, who were then matched down to a sample of 800 to produce the final dataset. The respondents were matched on gender, age, race, education, party identification, ideology and political interest. YouGovPolimetrix then weighted the matched set of survey respondents to known marginals for the registered voters of Texas from the 2008 Current Population survey and the 2007 Pew Religious Life Survey.

Sampling Frame and Target Sample

YouGovPolimetrix constructed a national sampling frame from the 2007 American Community Survey, including data on age, race, gender, education, marital status, number of children under 18, family income, employment status, citizenship, state, and metropolitan area. The frame was constructed by stratified sampling from the full 2007 ACS sample with selection within strata by weighted sampling with replacements (using the person weights on the public use file). Data on voter registration status and turnout were matched to this frame using the November 2008 Current Population Survey. Data on interest in politics and party identification were then matched to this frame from the 2007 Pew Religious Life survey, using the following variables for the match: age, race, gender, education, marital status, number of children under 18, family income, employment status, citizenship, state. The target sample of 800 Texas registered voters was selected with stratification by age, race, gender, education, and with simple random sampling within strata.

Weighting

Because matching is approximate, rather than exact, and response rates vary by group, the sample of completed interviews normally shows small amounts of imbalance that can be corrected by post-stratification weighting. Raking, first proposed by Deming and Stephan (1940), adjusts an initial set of weights to match a known set of population marginals, using a method of iterative proportional fitting (see Bishop, Fienberg and Holland, 1975 for details). In this procedure, the weights are adjusted sequentially to match the marginal distribution of each weight variable. The process proceeds until all marginals are matched. It does not require any information about the joint distribution of the variables (though, if these data are available and believed to be important, they can be employed by defining a marginal distribution involving a cross-classification of two variables). Post-stratification weights are calculated by raking the completed interviews to known marginals for Texas registered voters from the November 2008 Current Population Survey for the following variables: age, race, gender, and education.

9.11 May 2011

For the May 2011 University of Texas / Texas Tribune survey, Polimetrix interviewed 891 respondents who were then matched down to a sample of 800 to produce the final dataset. The respondents were matched on gender, age, race, education, party identification, ideology and political interest. Polimetrix then weighted the matched set of survey respondents to known marginals for the registered voters of Texas from the 2008 Current Population survey and the 2007 Pew Religious Life Survey.

Sampling Frame and Target Sample

YouGovPolimetrix constructed a national sampling frame from the 2007 American Community Survey, including data on age, race, gender, education, marital status, number of children under 18, family income, employment status, citizenship, state, and metropolitan area. The frame was constructed by stratified sampling from the full 2007 ACS sample with selection within strata by weighted sampling with replacements (using the person weights on the public use file). Data on voter registration status and turnout were matched to this frame using the November 2008 Current Population Survey. Data on interest in politics and party identification were then matched to this frame from the 2007 Pew Religious Life survey, using the following variables for the match: age, race, gender, education, marital status, number of children under 18, family income, employment status, citizenship, state. The target sample of 800 Texas registered voters was selected with stratification by age, race, gender, education, and with simple random sampling within strata.

Weighting

Because matching is approximate, rather than exact, and response rates vary by group, the sample of completed interviews normally shows small amounts of imbalance that can be corrected by post-stratification weighting. Raking, first proposed by Deming and Stephan (1940), adjusts an initial set of weights to match a known set of population marginals, using a method of iterative proportional fitting (see Bishop, Fienberg and Holland, 1975 for details). In this procedure, the weights are adjusted sequentially to match the marginal distribution of each weight variable. The process proceeds until all marginals are matched. It does not require any information about the joint distribution of the variables (though, if these data are available and believed to be important, they can be employed by defining a marginal distribution involving a cross-classification of two variables). Post-stratification weights are calculated by raking the completed interviews to known marginals for Texas registered voters from the November 2008 Current Population Survey for the following variables: age, race, gender, and education.

9.12 October 2011

Sampling and Weighting Methodology for the October 2011 Texas Statewide Study

For the survey, YouGovPolimetrix interviewed 889 respondents between October 19-26 2011, who were then matched down to a sample of 800 to produce the final dataset. The respondents were matched on gender, age, race, education, party identification, ideology and political interest. YouGovPolimetrix then weighted the matched set of survey respondents to known marginals for the registered voters of Texas from the 2008 Current Population survey and the 2007 Pew Religious Life Survey.

Sampling Frame and Target Sample

YouGovPolimetrix constructed a national sampling frame from the 2007 American Community Survey, including data on age, race, gender, education, marital status, number of children under 18, family income, employment status, citizenship, state, and metropolitan area. The frame was constructed by stratified sampling from the full 2007 ACS sample with selection within strata by weighted sampling with replacements (using the person weights on the public use file). Data on voter registration status and turnout were matched to this frame using the November 2008 Current Population Survey. Data on interest in politics and party identification were then matched to this frame from the 2007 Pew Religious Life survey, using the following variables for the match: age, race, gender, education, marital status, number of children under 18, family income, employment status, citizenship, state. The target sample of 800 Texas registered voters was selected with stratification by age, race, gender, education, and with simple random sampling within strata.

Weighting

Because matching is approximate, rather than exact, and response rates vary by group, the sample of completed interviews normally shows small amounts of imbalance that can be corrected by post-stratification weighting.

Raking, first proposed by Deming and Stephan (1940), adjusts an initial set of weights to match a known set of population marginals, using a method of iterative proportional fitting (see Bishop, Fienberg and Holland, 1975 for details). In this procedure, the weights are adjusted sequentially to match the marginal distribution of each weight variable. The process proceeds until all marginals are matched. It does not require any information about the joint distribution of the variables (though, if these data are available and believed to be important, they can be employed by defining a marginal distribution involving a cross-classification of two variables).

Post-stratification weights are calculated by raking the completed interviews to known marginals for Texas registered voters from the November 2008 Current Population Survey for the following variables: age, race, gender, and education.

Survey Panel Data

The YouGov panel, a proprietary opt-in survey panel, is comprised of 1.2 million U.S. residents who have agreed to participate in YouGov Web surveys. At any given time, YouGov maintains a minimum of five recruitment campaigns based on salient current events.

Panel members are recruited by a number of methods and on a variety of topics to help ensure diversity in the panel population. Recruiting methods include Web advertising campaigns (public surveys), permission-based email campaigns, partner sponsored solicitations, telephone-to-Web recruitment (RDD based sampling), and mail-to-Web recruitment (Voter Registration Based Sampling).

The primary method of recruitment for the YouGov Panel is Web advertising campaigns that appear based on keyword searches. In practice, a search in Google may prompt an active YouGov advertisement soliciting opinion on the search topic. At the conclusion of the short survey respondents are invited to join the YouGov panel in order to receive and participate in additional surveys. After a double opt-in procedure, where respondents must confirm their consent by responding to an email, the database checks to ensure the newly recruited panelist is in fact new and that the address information provided is valid.

Additionally, YouGov augments their panel with difficult to recruit respondents by soliciting panelists in telephone and mail surveys. For example, in 2006 and 2010, YouGov completed telephone interviews using RDD sampling and invited respondents to join the online panel. Respondents provided a working email where they could confirm their consent and request to receive online survey invitations. YouGov also employed registration based sampling, inviting respondents to complete a pre-election survey online. At the conclusion of that survey, respondents were invited to become YouGov members and receive additional survey invitations at their email address.

The YouGov panel currently has nearly 20,000 panelists who are residents of Texas. These panelists cover a wide range of demographic characteristics.

Sampling and Sample Matching

Sample matching is a methodology for selection of "representative" samples from non-randomly selected pools of respondents. It is ideally suited for Web access panels, but could also be used for other types of surveys, such as phone surveys. Sample matching starts with an enumeration of the target population. For general population studies, the target population is all adults, and can be enumerated through the use of the decennial Census or a high quality survey, such as the American Community Survey. In other contexts, this is known as the sampling frame, though, unlike conventional sampling, the sample is not drawn from the frame. Traditional sampling, then, selects individuals from the sampling frame at random for participation in the study. This may not be feasible or economical as the contact information, especially email addresses, is not available for all individuals in the frame and refusals to participate increase the costs of sampling in this way.

Sample selection using the matching methodology is a two-stage process. First, a random sample is drawn from the target population. We call this sample the target sample. Details on how the target sample is drawn are provided below, but the essential idea is that this sample is a true probability sample and thus representative of the frame from which it was drawn.

Second, for each member of the target sample, we select one or more matching members from our pool of opt-in respondents. This is called the matched sample. Matching is accomplished using a large set of variables that are available in consumer and voter databases for both the target population and the opt-in panel.

The purpose of matching is to find an available respondent who is as similar as possible to the selected member of the target sample. The result is a sample of respondents who have the same measured characteristics as the target sample. Under certain conditions, described below, the matched sample will have similar properties to a true random sample. That is, the matched sample mimics the characteristics of the target sample.

When choosing the matched sample, it is necessary to find the closest matching respondent in the panel of opt-ins to each member of the target sample. YouGov employs the proximity matching method to find the closest matching respondent. For each variable used for matching, we define a distance function, d(x,y), which describes how "close" the values x and y are on a particular attribute. The overall distance between a member of the target sample and a member of the panel is a weighted sum of the individual distance functions on each attribute. The weights can be adjusted for each study based upon which variables are thought to be important for that study, though, for the most part, we have not found the matching procedure to be sensitive to small adjustments of the weights. A large weight, on the other hand, forces the algorithm toward an exact match on that dimension.

9.13 February 2012

Sampling and Weighting Methodology for the February 2012 Texas Statewide Study

For the survey, YouGov interviewed 909 respondents between February 8-15, 2012, who were then matched down to a sample of 800 to produce the final dataset. The respondents were matched on gender, age, race, education, party identification, ideology and political interest. YouGov then weighted the matched set of survey respondents to known marginals for the registered voters of Texas from the 2008 Current Population survey and the 2007 Pew Religious Landscape Survey.

Sampling Frame and Target Sample

YouGov constructed a national sampling frame from the 2007 American Community Survey, including data on age, race, gender, education, marital status, number of children under 18, family income, employment status, citizenship, state, and metropolitan area. The frame was constructed by stratified sampling from the full 2007 ACS sample with selection within strata by weighted sampling with replacements (using the person weights on the public use file). Data on voter registration status and turnout were matched to this frame using the November 2008 Current Population Survey. Data on interest in politics and party identification were then matched to this frame from the 2007 Pew Religious Life survey, using the following variables for the match: age, race, gender, education, marital status, number of children under 18, family income, employment status, citizenship, state. The target sample of 800 Texas registered voters was selected with stratification by age, race, gender, education, and with simple random sampling within strata.

Weighting

Because matching is approximate, rather than exact, and response rates vary by group, the sample of completed interviews normally shows small amounts of imbalance that can be corrected by post-stratification weighting. Raking, first proposed by Deming and Stephan (1940), adjusts an initial set of weights to match a known set of population marginals, using a method of iterative proportional fitting (see Bishop, Fienberg and Holland, 1975 for details). In this procedure, the weights are adjusted sequentially to match the marginal distribution of each weight variable. The process proceeds until all marginals are matched. It does not require any information about the joint distribution of the variables (though, if these data are available and believed to be important, they can be employed by defining a marginal distribution involving a cross-classification of two variables).

Post-stratification weights are calculated by raking the completed interviews to known marginals for Texas registered voters from the November 2008 Current Population Survey for the following variables: age, race, gender, and education. Survey Panel Data

The YouGov panel, a proprietary opt-in survey panel, is comprised of 1.2 million U.S. residents who have agreed to participate in YouGov Web surveys. At any given time, YouGov maintains a minimum of five recruitment campaigns based on salient current events. Panel members are recruited by a number of methods and on a variety of topics to help ensure diversity in the panel population. Recruiting methods include Web advertising campaigns (public surveys), permission-based email campaigns, partner sponsored solicitations, telephone-to-Web recruitment (RDD based sampling), and mail-to-Web recruitment (Voter Registration Based Sampling).

The primary method of recruitment for the YouGov Panel is Web advertising campaigns that appear based on keyword searches. In practice, a search in Google may prompt an active YouGov advertisement soliciting opinion on the search topic. At the conclusion of the short survey respondents are invited to join the YouGov panel in order to receive and participate in additional surveys. After a double opt-in procedure, where respondents must confirm their consent byr esponding to an email, the database checks to ensure the newly recruited panelist is in fact new and that the address information provided is valid.

Additionally, YouGov augments their panel with difficult to recruit respondents by soliciting panelists in telephone and mail surveys. For example, in 2006 and 2010, YouGov completed telephone interviews using RDD sampling and invited respondents to join the online panel. Respondents provided a working email where they could confirm their consent and request to receive online survey invitations. YouGov also employed registration based sampling, inviting respondents to complete a pre-election survey online. At the conclusion of that survey, respondents were invited to become YouGov members and receive additional survey invitations at their email address. The YouGov panel currently has nearly 20,000 panelists who are residents of Texas. These panelists cover a wide range of demographic characteristics.

Sampling and Sample Matching

Sample matching is a methodology for selection of "representative" samples from non-randomly selected pools of respondents. It is ideally suited for Web access panels, but could also be used for other types of surveys, such as phone surveys. Sample matching starts with an enumeration of the target population. For general population studies, the target population is all adults, and can be enumerated through the use of the decennial Census or a high quality survey, such as the American Community Survey. In other contexts, this is known as the sampling frame, though, unlike conventional sampling, the sample is not drawn from the frame. Traditional sampling, then, selects individuals from the sampling frame at random for participation in the study. This may not be feasible or economical as the contact information, especially email addresses, is not available for all individuals in the frame and refusals to participate increase the costs of sampling in this way.

Sample selection using the matching methodology is a two-stage process. First, a random sample is drawn from the target population. We call this sample the target sample. Details on how the target sample is drawn are provided below, but the essential idea is that this sample is a true probability sample and thus representative of the frame from which it was drawn. Second, for each member of the target sample, we select one or more matching members from our pool of opt-in respondents. This is called the matched sample. Matching is accomplished using a large set of variables that are available in consumer and voter databases for both the target population and the opt-in panel.

The purpose of matching is to find an available respondent who is as similar as possible to the selected member of the target sample. The result is a sample of respondents who have the same measured characteristics as the target sample. Under certain conditions, described below, the matched sample will have similar properties to a true random sample. That is, the matched sample mimics the characteristics of the target sample.

When choosing the matched sample, it is necessary to find the closest matching respondent in the panel of opt-ins to each member of the target sample. YouGov employs the proximity matching method to find the closest matching respondent. For each variable used for matching, we define a distance function, d(x,y), which describes how "close" the values x and y are on a particular attribute. The overall distance between a member of the target sample and a member of the panel is a weighted sum of the individual distance functions on each attribute. The weights can be adjusted for each study based upon which variables are thought to be important for that study, though, for the most part, we have not found the matching procedure to be sensitive to small adjustments of the weights. A large weight, on the other hand, forces the algorithm toward an exact match on that dimension.

9.14 May 2012

For the survey May 2012 , YouGov interviewed 909 respondents between May 7-13 2012, who were then matched down to a sample of 800 to produce the final dataset. The respondents were matched on gender, age, race, education, party identification, ideology and political interest. YouGov then weighted the matched set of survey respondents to known marginals for the registered voters of Texas from the 2008 Current Population survey and the 2007 Pew Religious Landscape Survey.

Sampling Frame and Target Sample

YouGov constructed a national sampling frame from the 2007 American Community Survey, including data on age, race, gender, education, marital status, number of children under 18, family income, employment status, citizenship, state, and metropolitan area. The frame was constructed by stratified sampling from the full 2007 ACS sample with selection within strata by weighted sampling with replacements (using the person weights on the public use file). Data on voter registration status and turnout were matched to this frame using the November 2008 Current Population Survey. Data on interest in politics and party identification were then matched to this frame from the 2007 Pew Religious Life survey, using the following variables for the match: age, race, gender, education, marital status, number of children under 18, family income, employment status, citizenship, state. The target sample of 800 Texas registered voters was selected with stratification by age, race, gender, education, and with simple random sampling within strata.

Weighting Because matching is approximate, rather than exact, and response rates vary by group, the sample of completed interviews normally shows small amounts of imbalance that can be corrected by post-stratification weighting.

Raking, first proposed by Deming and Stephan (1940), adjusts an initial set of weights to match a known set of population marginals, using a method of iterative proportional fitting (see Bishop, Fienberg and Holland, 1975 for details). In this procedure, the weights are adjusted sequentially to match the marginal distribution of each weight variable. The process proceeds until all marginals are matched. It does not require any information about the joint distribution of the variables (though, if these data are available and believed to be important, they can be employed by defining a marginal distribution involving a cross-classification of two variables).

Post-stratification weights are calculated by raking the completed interviews to known marginals for Texas registered voters from the November 2008 Current Population Survey for the following variables: age, race, gender, and education.

Survey Panel Data

The YouGov panel, a proprietary opt-in survey panel, is comprised of 1.2 million U.S. residents who have agreed to participate in YouGov Web surveys. At any given time, YouGov maintains a minimum of five recruitment campaigns based on salient current events.

Panel members are recruited by a number of methods and on a variety of topics to help ensure diversity in the panel population. Recruiting methods include Web advertising campaigns (public surveys), permission-based email campaigns, partner sponsored solicitations, telephone-to-Web recruitment (RDD based sampling), and mail-to-Web recruitment (Voter Registration Based Sampling).

The primary method of recruitment for the YouGov Panel is Web advertising campaigns that appear based on keyword searches. In practice, a search in Google may prompt an active YouGov advertisement soliciting opinion on the search topic. At the conclusion of the short survey respondents are invited to join the YouGov panel in order to receive and participate in additional surveys. After a double opt-in procedure, where respondents must confirm their consent by responding to an email, the database checks to ensure the newly recruited panelist is in fact new and that the address information provided is valid.

Additionally, YouGov augments their panel with difficult to recruit respondents by soliciting panelists in telephone and mail surveys. For example, in 2006 and 2010, YouGov completed telephone interviews using RDD sampling and invited respondents to join the online panel. Respondents provided a working email where they could confirm their consent and request to receive online survey invitations. YouGov also employed registration based sampling, inviting respondents to complete a pre-election survey online. At the conclusion of that survey, respondents were invited to become YouGov members and receive additional survey invitations at their email address.

The YouGov panel currently has nearly 20,000 panelists who are residents of Texas. These panelists cover a wide range of demographic characteristics.

Sampling and Sample Matching

Sample matching is a methodology for selection of "representative" samples from non-randomly selected pools of respondents. It is ideally suited for Web access panels, but could also be used for other types of surveys, such as phone surveys. Sample matching starts with an enumeration of the target population. For general population studies, the target population is all adults, and can be enumerated through the use of the decennial Census or a high quality survey, such as the American Community Survey. In other contexts, this is known as the sampling frame, though, unlike conventional sampling, the sample is not drawn from the frame. Traditional sampling, then, selects individuals from the sampling frame at random for participation in the study. This may not be feasible or economical as the contact information, especially email addresses, is not available for all individuals in the frame and refusals to participate increase the costs of sampling in this way.

Sample selection using the matching methodology is a two-stage process. First, a random sample is drawn from the target population. We call this sample the target sample. Details on how the target sample is drawn are provided below, but the essential idea is that this sample is a true probability sample and thus representative of the frame from which it was drawn.

Second, for each member of the target sample, we select one or more matching members from our pool of opt-in respondents. This is called the matched sample. Matching is accomplished using a large set of variables that are available in consumer and voter databases for both the target population and the opt-in panel.

The purpose of matching is to find an available respondent who is as similar as possible to the selected member of the target sample. The result is a sample of respondents who have the same measured characteristics as the target sample. Under certain conditions, described below, the matched sample will have similar properties to a true random sample. That is, the matched sample mimics the characteristics of the target sample.

When choosing the matched sample, it is necessary to find the closest matching respondent in the panel of opt-ins to each member of the target sample. YouGov employs the proximity matching method to find the closest matching respondent. For each variable used for matching, we define a distance function, d(x,y), which describes how "close" the values x and y are on a particular attribute. The overall distance between a member of the target sample and a member of the panel is a weighted sum of the individual distance functions on each attribute. The weights can be adjusted for each study based upon which variables are thought to be important for that study, though, for the most part, we have not found the matching procedure to be sensitive to small adjustments of the weights. A large weight, on the other hand, forces the algorithm toward an exact match on that dimension.

9.15 October 2012

Sampling and Weighting Methodology for the October 2012 Texas Statewide Study

For the survey, YouGov interviewed 912 respondents between Oct 15-21 2012, who were then matched down to a sample of 800 to produce the final dataset. The respondents were matched on gender, age, race, education, party identification, ideology and political interest. YouGov then weighted the matched set of survey respondents to known marginals for the registered voters of Texas from the 2008 Current Population survey and the 2007 Pew Religious Landscape Survey.

Sampling Frame and Target Sample

YouGov constructed a national sampling frame from the 2007 American Community Survey, including data on age, race, gender, education, marital status, number of children under 18, family income, employment status, citizenship, state, and metropolitan area. The frame was constructed by stratified sampling from the full 2007 ACS sample with selection within strata by weighted sampling with replacements (using the person weights on the public use file). Data on voter registration status and turnout were matched to this frame using the November 2008 Current Population Survey. Data on interest in politics and party identification were then matched to this frame from the 2007 Pew Religious Life survey, using the following variables for the match: age, race, gender, education, marital status, number of children under 18, family income, employment status, citizenship, state. The target sample of 800 Texas registered voters was selected with stratification by age, race, gender, education, and with simple random sampling within strata.

Weighting

Because matching is approximate, rather than exact, and response rates vary by group, the sample of completed interviews normally shows small amounts of imbalance that can be corrected by post-stratification weighting.

Raking, first proposed by Deming and Stephan (1940), adjusts an initial set of weights to match a known set of population marginals, using a method of iterative proportional fitting (see Bishop, Fienberg and Holland, 1975 for details). In this procedure, the weights are adjusted sequentially to match the marginal distribution of each weight variable. The process proceeds until all marginals are matched. It does not require any information about the joint distribution of the variables (though, if these data are available and believed to be important, they can be employed by defining a marginal distribution involving a cross-classification of two variables).

Post-stratification weights are calculated by raking the completed interviews to known marginals for Texas registered voters from the November 2008 Current Population Survey for the following variables: age, race, gender, and education.

Survey Panel Data

The YouGov panel, a proprietary opt-in survey panel, is comprised of 1.2 million U.S. residents who have agreed to participate in YouGov Web surveys. At any given time, YouGov maintains a minimum of five recruitment campaigns based on salient current events.

Panel members are recruited by a number of methods and on a variety of topics to help ensure diversity in the panel population. Recruiting methods include Web advertising campaigns (public surveys), permission-based email campaigns, partner sponsored solicitations, telephone-to-Web recruitment (RDD based sampling), and mail-to-Web recruitment (Voter Registration Based Sampling).

The primary method of recruitment for the YouGov Panel is Web advertising campaigns that appear based on keyword searches. In practice, a search in Google may prompt an active YouGov advertisement soliciting opinion on the search topic. At the conclusion of the short survey respondents are invited to join the YouGov panel in order to receive and participate in additional surveys. After a double opt-in procedure, where respondents must confirm their consent by responding to an email, the database checks to ensure the newly recruited panelist is in fact new and that the address information provided is valid.

Additionally, YouGov augments their panel with difficult to recruit respondents by soliciting panelists in telephone and mail surveys. For example, in 2006 and 2010, YouGov completed telephone interviews using RDD sampling and invited respondents to join the online panel. Respondents provided a working email where they could confirm their consent and request to receive online survey invitations. YouGov also employed registration based sampling, inviting respondents to complete a pre-election survey online. At the conclusion of that survey, respondents were invited to become YouGov members and receive additional survey invitations at their email address.

The YouGov panel currently has nearly 20,000 panelists who are residents of Texas. These panelists cover a wide range of demographic characteristics.

Sampling and Sample Matching

Sample matching is a methodology for selection of "representative" samples from non-randomly selected pools of respondents. It is ideally suited for Web access panels, but could also be used for other types of surveys, such as phone surveys. Sample matching starts with an enumeration of the target population. For general population studies, the target population is all adults, and can be enumerated through the use of the decennial Census or a high quality survey, such as the American Community Survey. In other contexts, this is known as the sampling frame, though, unlike conventional sampling, the sample is not drawn from the frame. Traditional sampling, then, selects individuals from the sampling frame at random for participation in the study. This may not be feasible or economical as the contact information, especially email addresses, is not available for all individuals in the frame and refusals to participate increase the costs of sampling in this way.

Sample selection using the matching methodology is a two-stage process. First, a random sample is drawn from the target population. We call this sample the target sample. Details on how the target sample is drawn are provided below, but the essential idea is that this sample is a true probability sample and thus representative of the frame from which it was drawn.

Second, for each member of the target sample, we select one or more matching members from our pool of opt-in respondents. This is called the matched sample. Matching is accomplished using a large set of variables that are available in consumer and voter databases for both the target population and the opt-in panel.

The purpose of matching is to find an available respondent who is as similar as possible to the selected member of the target sample. The result is a sample of respondents who have the same measured characteristics as the target sample. Under certain conditions, described below, the matched sample will have similar properties to a true random sample. That is, the matched sample mimics the characteristics of the target sample.

When choosing the matched sample, it is necessary to find the closest matching respondent in the panel of opt-ins to each member of the target sample. YouGov employs the proximity matching method to find the closest matching respondent. For each variable used for matching, we define a distance function, d(x,y), which describes how "close" the values x and y are on a particular attribute. The overall distance between a member of the target sample and a member of the panel is a weighted sum of the individual distance functions on each attribute. The weights can be adjusted for each study based upon which variables are thought to be important for that study, though, for the most part, we have not found the matching procedure to be sensitive to small adjustments of the weights. A large weight, on the other hand, forces the algorithm toward an exact match on that dimension.

9.16 February 2013

Sampling and Weighting Methodology for the February 2013 Texas Statewide Study

For the survey, YouGov interviewed 1420 respondents between February 15-25, 2013, who were then matched down to a sample of 1200 to produce the final dataset. The respondents were matched on gender, age, race, education, party identification, ideology and political interest. YouGov then weighted the matched set of survey respondents to known marginals for the registered voters of Texas from the 2008 Current Population survey and the 2007 Pew Religious Landscape Survey.

Sampling Frame and Target Sample

YouGov constructed a national sampling frame from the 2007 American Community Survey, including data on age, race, gender, education, marital status, number of children under 18, family income, employment status, citizenship, state, and metropolitan area. The frame was constructed by stratified sampling from the full 2007 ACS sample with selection within strata by weighted sampling with replacements (using the person weights on the public use file). Data on voter registration status and turnout were matched to this frame using the November 2008 Current Population Survey. Data on interest in politics and party identification were then matched to this frame from the 2007 Pew Religious Life survey, using the following variables for the match: age, race, gender, education, marital status, number of children under 18, family income, employment status, citizenship, state. The target sample of 1200 Texas registered voters was selected with stratification by age, race, gender, education, and with simple random sampling within strata.

Weighting

Because matching is approximate, rather than exact, and response rates vary by group, the sample of completed interviews normally shows small amounts of imbalance that can be corrected by post-stratification weighting.

The matched cases were weighted to the sampling frame using propensity scores. The matched cases and the frame were combined and a logistic regression was estimated for inclusion in the frame. The propensity score function included age, years of education, gender, race/ethnicity, and interest in politics. The propensity scores were grouped into deciles of the estimated propensity score in the frame and post-stratified according to these deciles. Weights larger than 7 were trimmed and the final weights normalized to equal sample size.

The margin of error of the weighted data is 3.3%.

Survey Panel Data

The YouGov panel, a proprietary opt-in survey panel, is comprised of 1.2 million U.S. residents who have agreed to participate in YouGov Web surveys. At any given time, YouGov maintains a minimum of five recruitment campaigns based on salient current events.

Panel members are recruited by a number of methods and on a variety of topics to help ensure diversity in the panel population. Recruiting methods include Web advertising campaigns (public surveys), permission-based email campaigns, partner sponsored solicitations, telephone-to-Web recruitment (RDD based sampling), and mail-to-Web recruitment (Voter Registration Based Sampling).

The primary method of recruitment for the YouGov Panel is Web advertising campaigns that appear based on keyword searches. In practice, a search in Google may prompt an active YouGov advertisement soliciting opinion on the search topic. At the conclusion of the short survey respondents are invited to join the YouGov panel in order to receive and participate in additional surveys. After a double opt-in procedure, where respondents must confirm their consent by responding to an email, the database checks to ensure the newly recruited panelist is in fact new and that the address information provided is valid.

Additionally, YouGov augments their panel with difficult to recruit respondents by soliciting panelists in telephone and mail surveys. For example, in 2006 and 2010, YouGov completed telephone interviews using RDD sampling and invited respondents to join the online panel. Respondents provided a working email where they could confirm their consent and request to receive online survey invitations. YouGov also employed registration based sampling, inviting respondents to complete a pre-election survey online. At the conclusion of that survey, respondents were invited to become YouGov members and receive additional survey invitations at their email address.

The YouGov panel currently has nearly 20,000 panelists who are residents of Texas. These panelists cover a wide range of demographic characteristics.

Sampling and Sample Matching

Sample matching is a methodology for selection of "representative" samples from non-randomly selected pools of respondents. It is ideally suited for Web access panels, but could also be used for other types of surveys, such as phone surveys. Sample matching starts with an enumeration of the target population. For general population studies, the target population is all adults, and can be enumerated through the use of the decennial Census or a high quality survey, such as the American Community Survey. In other contexts, this is known as the sampling frame, though, unlike conventional sampling, the sample is not drawn from the frame. Traditional sampling, then, selects individuals from the sampling frame at random for participation in the study. This may not be feasible or economical as the contact information, especially email addresses, is not available for all individuals in the frame and refusals to participate increase the costs of sampling in this way.

Sample selection using the matching methodology is a two-stage process. First, a random sample is drawn from the target population. We call this sample the target sample. Details on how the target sample is drawn are provided below, but the essential idea is that this sample is a true probability sample and thus representative of the frame from which it was drawn.

Second, for each member of the target sample, we select one or more matching members from our pool of opt-in respondents. This is called the matched sample. Matching is accomplished using a large set of variables that are available in consumer and voter databases for both the target population and the opt-in panel.

The purpose of matching is to find an available respondent who is as similar as possible to the selected member of the target sample. The result is a sample of respondents who have the same measured characteristics as the target sample. Under certain conditions, described below, the matched sample will have similar properties to a true random sample. That is, the matched sample mimics the characteristics of the target sample.

When choosing the matched sample, it is necessary to find the closest matching respondent in the panel of opt-ins to each member of the target sample. YouGov employs the proximity matching method to find the closest matching respondent. For each variable used for matching, we define a distance function, d(x,y), which describes how "close" the values x and y are on a particular attribute. The overall distance between a member of the target sample and a member of the panel is a weighted sum of the individual distance functions on each attribute. The weights can be adjusted for each study based upon which variables are thought to be important for that study, though, for the most part, we have not found the matching procedure to be sensitive to small adjustments of the weights. A large weight, on the other hand, forces the algorithm toward an exact match on that dimension.

9.17 June 2013

Sampling and Weighting Methodology for the May-June 2013 Texas Statewide Study

For the survey, YouGov interviewed 1359 respondents between May 30-June13, 2013, who were then matched down to a sample of 1200 to produce the final dataset. The respondents were matched on gender, age, race, education, party identification, ideology and political interest. YouGov then weighted the matched set of survey respondents to known characteristics of registered voters of Texas from the 2010 Current Population survey and the 2007 Pew Religious Landscape Survey.

Sampling Frame and Target Sample

YouGov constructed a national sampling frame from the 2007 American Community Survey, including data on age, race, gender, education, marital status, number of children under 18, family income, employment status, citizenship, state, and metropolitan area. The frame was constructed by stratified sampling from the full 2007 ACS sample with selection within strata by weighted sampling with replacements (using the person weights on the public use file). Data on voter registration status and turnout were matched to this frame using the November 2010 Current Population Survey. Data on interest in politics and party identification were then matched to this frame from the 2007 Pew Religious Life survey, using the following variables for the match: age, race, gender, education, marital status, number of children under 18, family income, employment status, citizenship, state. The target sample of 1200 Texas registered voters was selected with stratification by age, race, gender, education, and with simple random sampling within strata.

Weighting

Because matching is approximate, rather than exact, and response rates vary by group, the sample of completed interviews normally shows small amounts of imbalance that can be corrected by post-stratification weighting. The matched cases were weighted to the sampling frame using propensity scores.

The matched cases and the frame were combined and a logistic regression was estimated for inclusion in the frame. The propensity score function included age, years of education, gender, race/ethnicity, and interest in politics. The propensity scores were grouped into deciles of the estimated propensity score in the frame and post-stratified according to these deciles. Weights larger than 7 were trimmed and the final weights normalized to equal sample size.

The margin of error of the weighted data is 3.3%.

Survey Panel Data

The YouGov panel, a proprietary opt-in survey panel, is comprised of 1.2 million U.S. residents who have agreed to participate in YouGov Web surveys. At any given time, YouGov maintains a minimum of five recruitment campaigns based on salient current events.

Panel members are recruited by a number of methods and on a variety of topics to help ensure diversity in the panel population. Recruiting methods include Web advertising campaigns (public surveys), permission-based email campaigns, partner sponsored solicitations, telephone-to-Web recruitment (RDD based sampling), and mail-to-Web recruitment (Voter Registration Based Sampling).

The primary method of recruitment for the YouGov Panel is Web advertising campaigns that appear based on keyword searches. In practice, a search in Google may prompt an active YouGov advertisement soliciting opinion on the search topic. At the conclusion of the short survey respondents are invited to join the YouGov panel in order to receive and participate in additional surveys. After a double opt-in procedure, where respondents must confirm their consent by responding to an email, the database checks to ensure the newly recruited panelist is in fact new and that the address information provided is valid.

Additionally, YouGov augments their panel with difficult to recruit respondents by soliciting panelists in telephone and mail surveys. For example, in 2006 and 2010, YouGov completed telephone interviews using RDD sampling and invited respondents to join the online panel. Respondents provided a working email where they could confirm their consent and request to receive online survey invitations. YouGov also employed registration based sampling, inviting respondents to complete a pre-election survey online. At the conclusion of that survey, respondents were invited to become YouGov members and receive additional survey invitations at their email address.

The YouGov panel currently has nearly 20,000 active panelists who are residents of Texas. These panelists cover a wide range of demographic characteristics.

Sampling and Sample Matching

Sample matching is a methodology for selection of "representative" samples from non-randomly selected pools of respondents. It is ideally suited for Web access panels, but could also be used for other types of surveys, such as phone surveys. Sample matching starts with an enumeration of the target population. For general population studies, the target population is all adults, and can be enumerated through the use of the decennial Census or a high quality survey, such as the American Community Survey. In other contexts, this is known as the sampling frame, though, unlike conventional sampling, the sample is not drawn from the frame. Traditional sampling, then, selects individuals from the sampling frame at random for participation in the study. This may not be feasible or economical as the contact information, especially email addresses, is not available for all individuals in the frame and refusals to participate increase the costs of sampling in this way.

Sample selection using the matching methodology is a two-stage process. First, a random sample is drawn from the target population. We call this sample the target sample. Details on how the target sample is drawn are provided below, but the essential idea is that this sample is a true probability sample and thus representative of the frame from which it was drawn.

Second, for each member of the target sample, we select one or more matching members from our pool of opt-in respondents. This is called the matched sample. Matching is accomplished using a large set of variables that are available in consumer and voter databases for both the target population and the opt-in panel.

The purpose of matching is to find an available respondent who is as similar as possible to the selected member of the target sample. The result is a sample of respondents who have the same measured characteristics as the target sample. Under certain conditions, described below, the matched sample will have similar properties to a true random sample. That is, the matched sample mimics the characteristics of the target sample.

When choosing the matched sample, it is necessary to find the closest matching respondent in the panel of opt-ins to each member of the target sample. YouGov employs the proximity matching method to find the closest matching respondent. For each variable used for matching, we define a distance function, d(x,y), which describes how "close" the values x and y are on a particular attribute. The overall distance between a member of the target sample and a member of the panel is a weighted sum of the individual distance functions on each attribute. The weights can be adjusted for each study based upon which variables are thought to be important for that study, though, for the most part, we have not found the matching procedure to be sensitive to small adjustments of the weights. A large weight, on the other hand, forces the algorithm toward an exact match on that dimension.

9.18 October 2013

Sampling and Weighting Methodology for the October 2013 Texas Statewide Study

For the survey, YouGov interviewed 1618 respondents between October 18-29, 2013, who were then matched down to a sample of 1200 to produce the final dataset. The respondents were matched on gender, age, race, education, party identification, ideology and political interest. YouGov then weighted the matched set of survey respondents to known characteristics of registered voters of Texas from the 2010 Current Population survey and the 2007 Pew Religious Landscape Survey.

Sampling Frame and Target Sample

YouGov constructed a national sampling frame from the 2010 American Community Survey, including data on age, race, gender, education, marital status, number of children under 18, family income, employment status, citizenship, state, and metropolitan area. The frame was constructed by stratified sampling from the full 2010 ACS sample with selection within strata by weighted sampling with replacements (using the person weights on the public use file). Data on voter registration status and turnout were matched to this frame using the November 2010 Current Population Survey. Data on interest in politics and party identification were then matched to this frame from the 2007 Pew Religious Life survey, using the following variables for the match: age, race, gender, education, marital status, number of children under 18, family income, employment status, citizenship, state. The target sample of 1200 Texas registered voters was selected with stratification by age, race, gender, education, and with simple random sampling within strata.

Weighting

Because matching is approximate, rather than exact, and response rates vary by group, the sample of completed interviews normally shows small amounts of imbalance that can be corrected by post-stratification weighting. The matched cases were weighted to the sampling frame using propensity scores. The matched cases and the frame were combined and a logistic regression was estimated for inclusion in the frame. The propensity score function included age, years of education, gender, race/ethnicity, and interest in politics. The propensity scores were grouped into deciles of the estimated propensity score in the frame and post-stratified according to these deciles. Weights larger than 7 were trimmed and the final weights normalized to equal sample size.

The margin of error of the weighted data is 3.1%.

Survey Panel Data

The YouGov panel, a proprietary opt-in survey panel, is comprised of 1.2 million U.S. residents who have agreed to participate in YouGov Web surveys. At any given time, YouGov maintains a minimum of five recruitment campaigns based on salient current events.

Panel members are recruited by a number of methods and on a variety of topics to help ensure diversity in the panel population. Recruiting methods include Web advertising campaigns (public surveys), permission-based email campaigns, partner sponsored solicitations, telephone-to-Web recruitment (RDD based sampling), and mail-to-Web recruitment (Voter Registration Based Sampling). The primary method of recruitment for the YouGov Panel is Web advertising campaigns that appear based on keyword searches. In practice, a search in Google may prompt an active YouGov advertisement soliciting opinion on the search topic. At the conclusion of the short survey respondents are invited to join the YouGov panel in order to receive and participate in additional surveys. After a double opt-in procedure, where respondents must confirm their consent by responding to an email, the database checks to ensure the newly recruited panelist is in fact new and that the address information provided is valid.

Additionally, YouGov augments their panel with difficult to recruit respondents by soliciting panelists in telephone and mail surveys. For example, in 2006 and 2010, YouGov completed telephone interviews using RDD sampling and invited respondents to join the online panel. Respondents provided a working email where they could confirm their consent and request to receive online survey invitations. YouGov also employed registration based sampling, inviting respondents to complete a pre-election survey online. At the conclusion of that survey, respondents were invited to become YouGov members and receive additional survey invitations at their email address.

The YouGov panel currently has nearly 20,000 active panelists who are residents of Texas. These panelists cover a wide range of demographic characteristics.

Sampling and Sample Matching

Sample matching is a methodology for selection of "representative" samples from non-randomly selected pools of respondents. It is ideally suited for Web access panels, but could also be used for other types of surveys, such as phone surveys. Sample matching starts with an enumeration of the target population. For general population studies, the target population is all adults, and can be enumerated through the use of the decennial Census or a high quality survey, such as the American Community Survey. In other contexts, this is known as the sampling frame, though, unlike conventional sampling, the sample is not drawn from the frame. Traditional sampling, then, selects individuals from the sampling frame at random for participation in the study. This may not be feasible or economical as the contact information, especially email addresses, is not available for all individuals in the frame and refusals to participate increase the costs of sampling in this way.

Sample selection using the matching methodology is a two-stage process. First, a random sample is drawn from the target population. We call this sample the target sample. Details on how the target sample is drawn are provided below, but the essential idea is that this sample is a true probability sample and thus representative of the frame from which it was drawn.

Second, for each member of the target sample, we select one or more matching members from our pool of opt-in respondents. This is called the matched sample. Matching is accomplished using a large set of variables that are available in consumer and voter databases for both the target population and the opt-in panel. The purpose of matching is to find an available respondent who is as similar as possible to the selected member of the target sample. The result is a sample of respondents who have the same measured characteristics as the target sample. Under certain conditions, described below, the matched sample will have similar properties to a true random sample. That is, the matched sample mimics the characteristics of the target sample.

When choosing the matched sample, it is necessary to find the closest matching respondent in the panel of opt-ins to each member of the target sample. YouGov employs the proximity matching method to find the closest matching respondent. For each variable used for matching, we define a distance function, d(x,y), which describes how "close" the values x and y are on a particular attribute. The overall distance between a member of the target sample and a member of the panel is a weighted sum of the individual distance functions on each attribute. The weights can be adjusted for each study based upon which variables are thought to be important for that study, though, for the most part, we have not found the matching procedure to be sensitive to small adjustments of the weights. A large weight, on the other hand, forces the algorithm toward an exact match on that dimension.

9.19 February 2014

Sampling and Weighting Methodology for the October 2013 Texas Statewide Study

For the survey, YouGov interviewed 1327 respondents between February 7-16, who were then matched down to a sample of 1200 to produce the final dataset. The respondents were matched on gender, age, race, education, party identification, ideology and political interest. YouGov then weighted the matched set of survey respondents to known characteristics of registered voters of Texas from the 2010 Current Population survey and the 2007 Pew Religious Landscape Survey.

Sampling Frame and Target Sample

YouGov constructed a national sampling frame from the 2010 American Community Survey, including data on age, race, gender, education, marital status, number of children under 18, family income, employment status, citizenship, state, and metropolitan area. The frame was constructed by stratified sampling from the full 2010 ACS sample with selection within strata by weighted sampling with replacements (using the person weights on the public use file). Data on voter registration status and turnout were matched to this frame using the November 2010 Current Population Survey. Data on interest in politics and party identification were then matched to this frame from the 2007 Pew Religious Life survey, using the following variables for the match: age, race, gender, education, marital status, number of children under 18, family income, employment status, citizenship, state. The target sample of 1200 Texas registered voters was selected with stratification by age, race, gender, education, and with simple random sampling within strata.

Weighting

Because matching is approximate, rather than exact, and response rates vary by group, the sample of completed interviews normally shows small amounts of imbalance that can be corrected by post-stratification weighting.

The matched cases were weighted to the sampling frame using propensity scores. The matched cases and the frame were combined and a logistic regression was estimated for inclusion in the frame. The propensity score function included age, years of education, gender, race/ethnicity, and interest in politics. The propensity scores were grouped into deciles of the estimated propensity score in the frame and post-stratified according to these deciles. Weights larger than 7 were trimmed and the final weights normalized to equal sample size.

The margin of error of the weighted data is 3.58%.

The margin of error for Republican primary voters is 5.64%. The margin of error for Democratic primary voters is 6.04%.

Survey Panel Data

The YouGov panel, a proprietary opt-in survey panel, is comprised of 1.2 million U.S. residents who have agreed to participate in YouGov Web surveys. At any given time, YouGov maintains a minimum of five recruitment campaigns based on salient current events.

Panel members are recruited by a number of methods and on a variety of topics to help ensure diversity in the panel population. Recruiting methods include Web advertising campaigns (public surveys), permission-based email campaigns, partner sponsored solicitations, telephone-to-Web recruitment (RDD based sampling), and mail-to-Web recruitment (Voter Registration Based Sampling). The primary method of recruitment for the YouGov Panel is Web advertising campaigns that appear based on keyword searches. In practice, a search in Google may prompt an active YouGov advertisement soliciting opinion on the search topic. At the conclusion of the short survey respondents are invited to join the YouGov panel in order to receive and participate in additional surveys. After a double opt-in procedure, where respondents must confirm their consent by responding to an email, the database checks to ensure the newly recruited panelist is in fact new and that the address information provided is valid.

Additionally, YouGov augments their panel with difficult to recruit respondents by soliciting panelists in telephone and mail surveys. For example, in 2006 and 2010, YouGov completed telephone interviews using RDD sampling and invited respondents to join the online panel. Respondents provided a working email where they could confirm their consent and request to receive online survey invitations. YouGov also employed registration based sampling, inviting respondents to complete a pre-election survey online. At the conclusion of that survey, respondents were invited to become YouGov members and receive additional survey invitations at their email address.

The YouGov panel currently has nearly 20,000 active panelists who are residents of Texas. These panelists cover a wide range of demographic characteristics.

Sampling and Sample Matching

Sample matching is a methodology for selection of "representative" samples from non-randomly selected pools of respondents. It is ideally suited for Web access panels, but could also be used for other types of surveys, such as phone surveys. Sample matching starts with an enumeration of the target population. For general population studies, the target population is all adults, and can be enumerated through the use of the decennial Census or a high quality survey, such as the American Community Survey. In other contexts, this is known as the sampling frame, though, unlike conventional sampling, the sample is not drawn from the frame. Traditional sampling, then, selects individuals from the sampling frame at random for participation in the study. This may not be feasible or economical as the contact information, especially email addresses, is not available for all individuals in the frame and refusals to participate increase the costs of sampling in this way.

Sample selection using the matching methodology is a two-stage process. First, a random sample is drawn from the target population. We call this sample the target sample. Details on how the target sample is drawn are provided below, but the essential idea is that this sample is a true probability sample and thus representative of the frame from which it was drawn.

Second, for each member of the target sample, we select one or more matching members from our pool of opt-in respondents. This is called the matched sample. Matching is accomplished using a large set of variables that are available in consumer and voter databases for both the target population and the opt-in panel. The purpose of matching is to find an available respondent who is as similar as possible to the selected member of the target sample. The result is a sample of respondents who have the same measured characteristics as the target sample. Under certain conditions, described below, the matched sample will have similar properties to a true random sample. That is, the matched sample mimics the characteristics of the target sample.

When choosing the matched sample, it is necessary to find the closest matching respondent in the panel of opt-ins to each member of the target sample. YouGov employs the proximity matching method to find the closest matching respondent. For each variable used for matching, we define a distance function, d(x,y), which describes how "close" the values x and y are on a particular attribute. The overall distance between a member of the target sample and a member of the panel is a weighted sum of the individual distance functions on each attribute. The weights can be adjusted for each study based upon which variables are thought to be important for that study, though, for the most part, we have not found the matching procedure to be sensitive to small adjustments of the weights. A large weight, on the other hand, forces the algorithm toward an exact match on that dimension.

10. About These Polls

The UT-Austin Texas Politics Poll was launched in July of 2008, marking a major step forward in the measurement of public opinion for Texas.

In October, 2009, researchers in the Department of Government welcomed The Texas Tribune, a new non-profit, nonpartisan public media organization based in Austin, as collaborators in the project.

At the time the UT Poll was launched, Texas did not have a regularly occurring, non-partisan poll from which data are made available for public use (other large states, such as California and New York, have these sorts of polls). The UT Poll fills this gap by providing a much-needed measure of public opinion among adult citizens of Texas.

The UT-Austin/Texas Tribune poll is notable not just for its substantive contribution but for its research design as well. It is conducted in conjunction with YouGov/Polimetrix, a company that is well-known for its innovative internet-based survey techniques. Unlike most political surveys, the UT-Austin/Texas Tribune Poll will be conducted entirely online. Internet-based polling is rapidly becoming a popular alternative to telephone-based polling; as many have observed, telephone polls are becoming somewhat problematic because of the growth of cellphone-only households and caller ID screening in residential landlines. Thus, the UT-Austin/Texas Tribune Poll represents a cutting-edge approach to ascertaining popular opinion in Texas.

The poll is designed to provide both educational and research resources to students, educators, and the general public. The survey data is and will continue to be made available to the public and to researchers. Graphs and charts are designed to make data not only available but accessible to everyone interested in Texas politics and government.

Contact:

Dr. James Henson

Department of Government and Liberal Arts Instructional Technology Services

The University of Texas at Austin

512 471-0090

j.henson@austin.utexas.edu

Professor Daron Shaw

Government Department

University of Texas at Austin

512 232-7275

dshaw@austin.utexas.edu

Texas Politics:
© 2009, Liberal Arts Instructional Technology Services
University of Texas at Austin
3rd Edition - Revision 103
go back