AAPOR 2022

Jozkowski, K.N., Crawford, B.L., Turner, R.C., McClelland, S.I., Lo, W. (2022, May 11 - May 13). Conceptualizing and Measuring the Complexity of Abortion Attitudes in the United States. American Association of Public Opinion Research (AAPOR) Annual Meeting. Chicago, IL.

According to findings from several national polls and surveys, attitudes toward abortion in the United States have remained relatively stable since the 1970's. Despite these trends, abortion legislation has increased dramatically. Given such shifts in the abortion landscape, it is important to consider how abortion attitudes are measured. Abortion attitudes have been primarily assessed via two broad frameworks--the extent that abortion should be legal and is moral. Within these frameworks, people's attitudes are often contextual. Factors such as gestational age and circumstances regarding the pregnancy and the people involved in the pregnancy influence opinions. There may also be variability in strength of endorsement or support for abortion across these frameworks and contexts. For example, some people hold strong beliefs about abortion under some conditions but may hold less strong convictions under other circumstances. Additionally, some people may feel ambivalent about abortion in general or ambivalent across frameworks and circumstances associated with abortion. Given these findings, we argue that people's attitudes toward abortion are complex and multi-dimensional; current measures are not designed to capture such nuance. In this panel, we provide evidence for this complexity across five studies. First, we present findings from in-depth interviews, highlighting salient frameworks and contexts relevant to abortion. Next, we compare people's attitudes across two dominate frameworks—morality and legality—demonstrating how people may hold completing perspectives. Third, within the legality framework, we demonstrate how people's attitudes may vary based on whether they are asked about abortion as being legal versus illegal. Fourth, we provide evidence of complexity in terms of how people identify with regard to pro-life and pro-choice labels. And finally, we demonstrate how complexity in attitudes may vary depending on item wording and response format. We will conclude our panel by proposing ideas for capturing such complexity in abortion attitude measures going forward.

Jozkowski, K.N., Dennis, B., Crawford, B.L., Turner, R.C., & Lo, W. (2022, May 11 - May 13) Complexity in Abortion Attitudes: Findings from in-depth interviews with US adults. American Association of Public Opinion Research (AAPOR) Annual Meeting. Chicago, IL.

We theorize people's attitudes toward abortion are complex and multi-dimensional. By complex, we mean people may feel conflicted in their stance toward abortion such that their opinions may be contextual (e.g., abortion should be legal in instances of rape, but subsequent abortions should be illegal) or their strength of endorsement may vary based on circumstances (e.g., people feel strongly in some circumstances, but ambivalent or indifferent in other circumstances). We also theorize that an aspect of complexity may manifest in people potentially holding competing views of abortion due to the multi-dimensional nature of abortion attitudes. That is, people may feel conflicted in their stance based on the framework through which they are being asked to consider abortion. Representing competing frameworks--some people may hold certain beliefs about the morality of abortion as a particular framework and those beliefs may conflict with people's perspective on whether abortion should be legal or the agency of abortion seekers. Current mechanisms to assess abortion attitudes tend to primarily focus on legality as a framework and often lack assessments of contextual factors or ability to assess strength of endorsement. Such limitations may yield conceptualizations of abortion that lack nuance. To address these limitations, we conducted in-depth qualitative interviews with English (N = 110) and Spanish-speaking (N = 60) adults from NORC's AmeriSpeak panel to generate constructs that underlie people's attitudes toward abortion; we are currently analyzing data. In this presentation, we will discuss relevant (1) frameworks for conceptualizing abortion beyond legality, (2) contextual factors people perceive as relevant in terms of shaping their attitudes, and (3) other important factors relevant to attitude formation. Data from these interviews will be integrated with findings from other formative studies to identify common themes and constructs that will drive item development for new abortion attitude scales and polling items.

Turner, R.C., Crawford, B.L., Jozkowski, K.N., Lo, W. (2022, May 11 - May 13).When Reverse-Oriented Item Formats are Not Polar Opposites: An Experiment Comparing Attitudinal Responses when Asking if Abortion Should be Legal vs Illegal. American Association of Public Opinion Research (AAPOR) Annual Meeting. Chicago, IL.

Research has indicated that reverse-oriented questions are not necessarily contradictory (Spector et al, 1997; Tourangeau & Rasinski, 1988). In this study, we conducted an experiment of wording effects to investigate participants' levels of abortion support or opposition based on question orientation: should abortion be legal versus should abortion be illegal? We randomly assigned two sets of abortion circumstance questions to a volunteer sample (Qualtrics© n=2,042) with one group receiving the illegal format followed by the legal format, and the second group receiving the reverse order. Administering both sets to all participants allowed for an experimental comparison between the initial item wording formats, with comparisons to the second format to investigate order effects. Results indicate no significant order effect, however there was a significant wording effect. The impact of the wording effect was different for the types of abortion circumstances. Two types of circumstances were used to create composite scores (health-related and socioeconomic-related) as they have been demonstrated to function as two correlated, but multidimensional constructs (Muthen, 1981). For the health-related composite score, participants' responses to the legal format indicated significantly higher support for abortion than when they responded to the illegal format. In other words, participants were more likely to say yes when asked if abortion should be legal than to say no when asked if abortion should be illegal in circumstances such as health endangerment, rape, and genetic diseases. Conversely, participants indicated lower support for abortion in socioeconomic circumstances (e.g., unmarried, cannot afford, does not want more children) when asked if abortion should be legal (compared to illegal). Participants were less likely to say yes that abortion should be legal in socioeconomic circumstances than to say no when asked if abortion should be illegal. Results indicate item wording impacts may differ based on construct location.

Crawford, B.L., Jozkowski, K.N., Turner, R.C., Lo, W. (2022, May 11 - May 13). Assessing Abortion Complexity Across the Dimensions of Legality and Morality. American Association of Public Opinion Research (AAPOR) Annual Meeting. Chicago, IL.

Abortion is a complex multi-dimensional social issue. Findings from public opinion polls and survey research often frame abortion attitudes in terms of support or opposition to abortion across various dimensions (e.g., legality, morality). Another way to evaluate abortion attitudes is by examining abortion complexity, instead of support or opposition. Abortion complexity refers to a value free assessment of abortion attitudes that is not a measure of support or opposition but rather a measure of the extent that someone may be conflicted or ambivalent about abortion. Stated differently, abortion complexity can be thought of as the opposite of polarization. Abortion complexity has primarily been assessed within one dimension (e.g., complexity of abortion legality across different scenarios). In the current presentation, we examine abortion complexity across dimensions by identifying the number of scenarios where a participant has disagreement between similar items dealing with the morality and legality of abortion. Specifically, participants were provided with two sets of 10 scenario specific abortion questions. One set of questions asked about the legality of abortion and the second pertained to the morality of abortion. For our complexity measure, we summed the number of instances when a participant's answers for legality and morality did not match. Data were obtained using a probability-based panel designed to be nationally representative (n=1,025). Approximately 25% of the sample showed no complexity across the 10 items, with approximately 20% being complex on at least four of the items. In our presentation, we will present overall findings regarding legal – moral complexity within the sample and discuss which socio-demographic characteristics are the best predictors of legal – moral complexity. We will conclude with a discussion of the relevance of abortion complexity when assessing abortion attitudes.

McClelland, S., Baker, M.R., & Jozkowski, K.N. (2022, May 11 - May 13). Abortion Identities: Bringing Qualitative Insights to pro-choice and pro-life categories. American Association of Public Opinion Research (AAPOR) Annual Meeting. Chicago, IL.

Abstract: In two survey studies, we asked participants to report their abortion identity using a novel procedure which combined a quantitative measure with a qualitative follow up question. We asked participants to evaluate two questions: "How 'pro-life are you?" and "How 'pro-choice' are you?" on a self-anchored scale (McClelland, 2017), where they first indicated their level of agreement with each item and then provided qualitative data on how they defined each abortion idenity. Study 1 (N=72) was conducted in-person using paper and pencil; Study 2 (N=1,060) was conducted online using a sample from Prolific. Both studies were conducted with samples purposively constructed to represent diverse abortion and political identities, as well as gender and racial diversity. Across both studies, we found two patterns: (1) participants relied on extremely different definitions of pro-life and pro-choice when endorsing their level of agreement with each identity, indicating that these identities are more complex and also are less stable than most measures can capture and (2) the majority of participants endorsed at least some level of agreement of each identity, indicating more overlap between the two than most survey items allow for. These findings draw attention to the complexity of abortion identities which are often considered only in binary forms, and to the potential for mis-measurement of abortion identity in national polls where binary options are most commonly used.

Lo, W., Turner, R.C., Crawford, B.L., Jozkowski, K.N. (2022, May 11 - May 13). Comparing response labelling between agreement and judgement anchors: Using a multi-group structural equation model approach. American Association of Public Opinion Research (AAPOR) Annual Meeting. Chicago, I

The summative rating scale has been widely adapted in social science research after Likert designed his 5-point scale to quantify and measure underlying attitudes. The well-delineated guidelines for scale development have also been established and recommended by scholars (e.g., Krosnick & Presser, 2010). However, there is limited research investigating whether the response anchor labeling produces different response patterns, which may lead to measurement bias. Therefore, the purpose of this study is to compare two of the most frequently used anchors (Casper et al. 2019): the context labels of agreement (e.g., strongly agree – strongly disagree) and judgement (e.g., definitely legal to definitely illegal), and evaluate if both anchors could be used interchangeably while measuring people's attitudes. A total of 1,345 participants were recruited through the Prolific panel and randomly assigned to one of four sets of online surveys. The survey adopted 6 GSS items related to abortion circumstances with 2 underlying factors. These sets differed in the number of response categories (i.e., 4 and 5) and the context labeling (i.e., agreement and judgement). To evaluate psychometric properties between group differences in anchor labelling under each response category, the multi-group structural equation model (Thompson & Green, 2013) was used for comparing latent variable means and reaching coherent conclusions at the construct level. The procedure sequentially assesses configural, metric, and scalar invariance in factor means between groups. The results indicated there were no significant differences between agreement and judgement anchors in either 4 or 5 response categories. The full paper presents a discussion of the findings of these studies within the context of existing research and common practice.

Crawford, B.L., Jozkowski, K.N., Turner, R.C., Lo, W. (2022, May 11 - May 13). Dimensions of Abortion Attitudes and Their Relationship to Abortion Focused Policies and Politicians. American Association of Public Opinion Research (AAPOR) Annual Meeting. Chicago, IL.

Abortion is a complex social issue that yields attitudes and opinions that span multiple dimensions. These dimensions may include support for abortion generally, policies and laws related to abortion, or even identities related to abortion (e.g., pro-choice/pro-life). However, because different firms and surveys often use different measures, the extent that these dimensions are examined relative to each other is limited. For example, less research focuses on what types of abortion policies someone who identifies as pro-life might support, or the extent that individuals who support the legality of abortion in various circumstances also think it is immoral. Understanding the relationship between these different dimensions helps to contextualize the complexity and nuance of abortion attitudes. Furthermore, given that in 2022 the Supreme Court will rule on abortion laws in Mississippi and Texas that may drastically alter the precedence of Roe v. Wade and the abortion landscape in the United States, it is crucial that we continue to expand our understanding of the complexity of abortion attitudes and how these attitudes may translate to support or opposition for abortion-focused policies and politicians. This presentation uses survey data obtained from both quota and probability-based samples between 2017-2021 (N1 – 3,000; N2 = 919; N3 = 1,583). The surveys included various measures of abortion identity, scenario specific attitudes about the morality and legality of abortion, and support for different abortion-focused policies and politicians. First, we will use frequencies and crosstabs to highlight how the different questions and dimensions relate to one another. Next, we will explore which questions may be better predictors of support or opposition to abortion-focused policies and politicians. We will conclude by discussing why some measures are better predictors than others of abortion focused policies and politicians and make recommendations for key considerations in assessing abortion attitudes moving forward.

Turner, R.C., Jozkowski, K.N., Lo, W., Crawford, B.L. (2022, May 11 - May 13). Demographic and Response Comparisons of Participants Selecting to Use Different Device Types for Survey Completion. American Association of Public Opinion Research (AAPOR) Annual Meeting. Chicago, IL.

Smartphones are increasingly being used for survey administration, and experimental studies have investigated potential differences in responses based on mode of administration (e.g., Antoun, Couper & Conrad, 2017; Tourangeau et al., 2017). Although certain types of questions have been discussed as being more challenging for smartphone users, the mode effects across devices appear small in many circumstances. However, there are also studies that have identified significant differences in characteristics such as response rates, response times, and omissions for open-ended items (Vicente et al., 2009; Wells, Bailey, & Link, 2014). As research on the impacts of devices on survey responses continue, it may also be informative to understand who tends to use the different types of devices when completing surveys. In this study, we investigate responses from two web-based surveys administered using two national panels (2020 IPSOS KnowledgePanel [N=919] and 2021 NORC Amerispeak [N=1025]), allowing participants to select the type of device used for survey completion (computer, smartphone, tablet). Characteristics such as demographics and background variables, item omissions, and relationship to survey construct outcomes are compared for participants choosing different types of devices. Results from the IPSOS dataset indicate that smartphone users tend to be younger with lower educational levels than those using a computer. Democrats and participants with no political affiliation were also more likely than Republicans to use smartphones. Item omission levels were also higher for those using a smartphone. However, device used was not a significant predictor of construct outcomes investigated. A comparison of results across datasets will be made along with a discussion of additional characteristics investigated that were not found to be significant

Bueno, X., Montenegro, M.S., Lo, W., Valdez, D., Crawford, B.L., Turner, R.C. Jozkowski, K.N. (2022, May 11 - May 13). Contexts of Socialization and Abortion Attitudes: A Close Look at Latinx Population in the U.S. by Migrant Generation Status. American Association of Public Opinion Research (AAPOR) Annual Meeting. Chicago, IL.

Many Latin American countries are governed by abortion laws that are more restrictive than those in the United States. When moving to the U.S., integration processes into the new social context might progress at different paces among first-generation migrants while perhaps coexisting with beliefs and views conformed in their country of origin. Similarly, second and subsequent generation migrants -born and socialized in a different context- might adopt views on social issues, such as abortion, of their parents' generation. Given the dual cultural context inherent to families of migrant origin, and under the general assumption that abortion attitudes are, themselves, complex and contextual, we explored and compared people's attitudes toward legal abortion across first, second, and third-plus generations of Latinxs population in the US. A total of 1,183 self-identified Latinxs was selected from a larger data pool collected through Qualtrics using quota-based sampling. We applied a Multiple Indicators Multiple Causes (MIMIC) approach to investigate the effects of Latinxs' generation, sex, education level, religious attendance, and political affiliation on a 7-item measure of abortion attitudes comprising a two-factor model similar to the abortion items from the General Social Survey (i.e., 'hard' and 'soft' reasons; Benin, 1985). Results indicated abortion attitudes across Latinx adults in the US from different migrant generations are not homogeneous, but sensitive to the reasons motivating abortion. There were no significant differences across generations for 'hard' circumstances (i.e., woman health risk, rape, fetal anomaly). However, there were significant differences in abortion support across Latinx migrant generations in relation to abortion motivated by social-related circumstances known as 'soft' reasons. Our findings lend support to previous research which suggests complexity in abortion attitudes by demonstrating differences in attitudes across generations and other demographics. We will discuss implications for research involving culturally diverse populations in light of their migrant origin or ancestry.

Hadfield, J.I., Bueno, X., Ezike, N., Turner, R.C., Lo, W., Jozkowski, K.N., Crawford, B.L., (2022, May 11 - May 13). Are there Racial/Ethnic Disparities in Support for Abortion Access? Exploring the Role of Abortion Law Knowledge and Political Engagement. American Association of Public Opinion Research (AAPOR) Annual Meeting. Chicago, IL.

Given the social and health implications that unequal access to reproductive health has, it is important to explore whether a link exists between race-ethnicity and different factors involved in explaining beliefs towards abortion access. Little is known about whether racial-ethnic disparities within political engagement and abortion law knowledge might impact people's support for abortion access. Thus, this study aims to examine whether and how political engagement and abortion law knowledge, may inform racial-ethnic disparities regarding support for abortion access. We administered a web-based survey in English and Spanish to a national sample of n=2,383 participants that identified as Black/African American, Latinx, or White. The outcome variable asked participants about their position regarding how easy it should be to get an abortion and whether there should be fewer or more abortion restrictions in the US and in their state. We used political engagement, abortion law knowledge, and demographic (e.g. race-ethnicity) variables as predictors. Quantile regression was used to investigate the effect of political engagement and abortion law knowledge on support for abortion access across different racial-ethnic groups. Political engagement was significantly associated with support for abortion access at the 25th and 75th percentiles. Abortion knowledge was significantly associated with support for abortion access at all three percentiles. We found significant differences in support for abortion access between Black and White participants at the 25th percentile with Black respondents providing higher endorsements compared to White respondents. There was also a significant effect on support for abortion access among political affiliations and education levels. We conclude that the association between racial-ethnic groups and support for abortion access entails certain complexity. There are not consistent significant differences among groups, and, as shown, observed differences were reduced only to Black/African Americans (compared to Whites) and in the most polarized segments of the scale distribution.

*DeJarnatt, A., Turner, R.C., Lo, W., Jozkowski, K.N., Crawford, B.L., Buyuker, B. (2022, May 11 - May 13). A comparison of US adult profiles based on attitudes about abortion being moral vs legal. American Association of Public Opinion Research (AAPOR) Annual Meeting. Chicago, IL.

"Prior research indicates that US adults' attitudes about whether they believe abortion is moral does not always align with whether they believe abortion should be legal. Differences in attitudes about the morality and legality of abortion have been demonstrated for participants of different backgrounds such as racial/cultural groups (Dugger, 1998), political parties (Turner et al., 2021), and religious identity (Montenegro et al., 2020). Using a decision tree analytical approach, participant profiles are first developed for identifying demographic, sociological, personal experiences, and attitudinal variables related to beliefs about abortion morality and legality. Analyses are then conducted to identify which background variables differentiate between participant subgroups whose attitudes align on the morality and legality variables (i.e., it is moral and should be legal, it is immoral and should be illegal) versus participant subgroups who differ on these two outcomes (i.e., it is immoral but should be legal). Background variables such as prior reproductive health experiences, attitude about government involvement, and belief in when life begins are also included. The results are intended to provide a more nuanced understanding of how background factors may be related to complexity in adults' attitudes about abortion. National data collections obtained from a GfK panel collected in 2019 (n = 1583) and a nationally representative Amerispeak panel (n = 1025) collected in 2021 are used for cross-comparisons."

*Asamoah, N.A., Turner, R.C., Lo, W., Crawford, B.L., Jozkowski, K.N., McClelland, S.I. (2022, May 11 - May 13). A comparison of national historical and personal retrospective trends of perceptions about abortion. American Association of Public Opinion Research (AAPOR) Annual Meeting. Chicago, IL.

When using General Social Survey data (GSS, Smith et al., 2018) for attitudes about abortion legality from 1972 to 2018, trends indicate that US adults are generally supportive of abortion being legal in medical-related circumstances (e.g., health endangerment, birth defects) and split on support of abortion being legal for socioeconomic circumstances (e.g., not married, cannot afford more children). From 1972 to 1996, there are also relatively consistent age-related trends where younger age groups (18-34 years and 35-49 years) are the most supportive of abortion legality and the oldest age group (65 years and older) is least supportive. However, these trends appear disrupted from 1996 to 2018, with the exception of the oldest age group typically less supportive of socioeconomic circumstances. In this study, participants marked responses on a 5-point Likert-type scale to indicate whether they are supportive of or opposed to abortion over four timeframes in their life: now, one year ago, in early adulthood (18-25 years of age), and in adolescence (13-17 years of age). Data using a GfK panel in 2019 (n=1583) and an AmeriSpeak panel in 2021 (n=968) are presented. Relevant age groups indicate significantly larger percentages of participants that are more supportive of abortion now than in their adolescence. In regard to changes in attitude since young adulthood, approximately one-third of the two older age groups (≥ 45 years) indicate they are either more or less supportive with approximately equal proportions of change in each direction. For the two younger age groups (< 45 years) a significantly larger percent indicate they are more supportive of abortion now than in their young adulthood, and even one year prior. Further comparisons will be presented to investigate how the retrospective results that indicate more US adults believe they are becoming more supportive of abortion, aligns with historical data trends.

Baker, M.R., Jozkowski, K.N., McClelland, S.I. (2022, May 11 - May 13). The role of racism and sexism in attitudes towards abortion in White, Latinx, and Black individuals. American Association of Public Opinion Research (AAPOR) Annual Meeting. Chicago, IL.

Attitudes towards abortion play significant historical and contemporary roles in U.S. politics. While religious and political opinions are important predictors of abortion attitudes, we argue that other social attitudes—specifically "new and subtle" forms of anti-Black racism and sexism—could be predictive of abortion attitudes as well. In Study 1, we demonstrate the novel finding that individuals who endorsed more symbolic racism and modern sexism did, indeed, endorse more opposition to abortion in a sample of Black, Latinx, and White U.S. residing individuals (N=1,060). These relationships were true above and beyond religious factors, indicating the opposition to abortion is not simply a religious attitude as might be presumed based on prior literature and theory. In Study 2, we conducted a conceptual replication of these relationships using the 2012 ANES data (N=3,860) in order to: (a) test robustness of these novel findings using different measures of symbolic racism, modern sexism, and abortion; and (b) extend these findings by controlling for political ideology (liberal‒conservative) in addition to religious factors. Preliminary results showed that the relationships between more subtle racism and sexism and opposition to abortion replicated in the ANES online sample of Black and White voters and partially replicated in the sample of Latinx voters. These results confirm our hypotheses that racist and sexist attitudes relate to individuals abortion attitudes for both the religious and the non-religious. Future analyses for the presentation will include examining these relationships alongside political ideology. These two studies demonstrate that abortion attitudes are related to more than just religious or political attitudes. Without this insight, researchers and policy makers risk interpreting data from opinion polls without a full picture of what individuals might be drawing on when voting, answering poll questions, or supporting political candidates.

Jozkowski, K.N., Crawford, B.L., LaRoche, K.J., Handley, M.N., Turner, R.C., Lo, W. (2022, May 11 - May 13). Consequences for Illegal Abortion: Who should Face Penalties and What Should they Be? American Association of Public Opinion Research (AAPOR) Annual Meeting. Chicago, IL.

According to findings from several national polls and surveys, attitudes toward abortion in the United States have remained relatively stable since the 1970's. Despite these trends, abortion legislation has increased dramatically. Given such shifts in the abortion landscape, it is important to consider how abortion attitudes are measured. Abortion attitudes have been primarily assessed via two broad frameworks--the extent that abortion should be legal and is moral. Within these frameworks, people's attitudes are often contextual. Factors such as gestational age and circumstances regarding the pregnancy and the people involved in the pregnancy influence opinions. There may also be variability in strength of endorsement or support for abortion across these frameworks and contexts. For example, some people hold strong beliefs about abortion under some conditions but may hold less strong convictions under other circumstances. Additionally, some people may feel ambivalent about abortion in general or ambivalent across frameworks and circumstances associated with abortion. Given these findings, we argue that people's attitudes toward abortion are complex and multi-dimensional; current measures are not designed to capture such nuance. In this panel, we provide evidence for this complexity across five studies. First, we present findings from in-depth interviews, highlighting salient frameworks and contexts relevant to abortion. Next, we compare people's attitudes across two dominate frameworks—morality and legality—demonstrating how people may hold completing perspectives. Third, within the legality framework, we demonstrate how people's attitudes may vary based on whether they are asked about abortion as being legal versus illegal. Fourth, we provide evidence of complexity in terms of how people identify with regard to pro-life and pro-choice labels. And finally, we demonstrate how complexity in attitudes may vary depending on item wording and response format. We will conclude our panel by proposing ideas for capturing such complexity in abortion attitude measures going forward.

Lo, W., Turner, R.C., Crawford, B.L., Jozkowski, K.N. (2022, May 11 - May 13). Using autoregressive cross lagged exploratory structural equation model to detect wording effects between scales with even and odd response options. American Association of Public Opinion Research (AAPOR) Annual Meeting. Chicago, IL.

Likert (1932) proposed the well-known 5-point summative rating scale to quantify attitude measurement. A couple of decades later, Travers (1951) credits Paul Horst as the first to propose and develop the forced-choice technique idea (i.e., an even number of responses) which intended to remove the effect of attraction to select the middle option, which was especially useful in the personality scale. The general guideline has been well-documented for selecting even or odd numbers of responses (e.g., Krosnick & Presser, 2010). However, we only found a few studies investigating the response option with the wording effect under the repeated measure design. Therefore, we recruited 1,338 participants and randomly assigned them to one of four sets of response options with 4, 5, 6, and 7 categories. The survey adopted 6 General Social Survey (GSS) items related to abortion circumstances with two underlying factors. Then, the same set of questions was provided again by adding the gestational length (i.e., in 4 months pregnant) at the end of each question. Since all participants responded to the original GSS items without the gestational length, they might mark responses differently after the extra statement was provided. We used the autoregressive cross lagged exploratory structural equation modeling (ESEM) for further analyses. This new technique (Morin et al, 2014) uses the latent variables to predict other latent variables, whereas the traditional analytic procedures would aggregate data together, causing a loss of information. The full paper presents a discussion of the findings of these studies within the context of existing research and common practice.