Oral
Jozkowski, K.N., Crawford, B.L., Hawbaker, A., Parker, E., Golzarri Arroyo, L., Turner, R.C., & Lo, W. (2024, May 15- May 17). Examining complexity in people’s attitudes toward abortion legality and alignment with state legislation. Association of Public Opinion Research (AAPOR) Annual Meeting, Atlanta GA.
National polls and surveys use the same questions over time to assess people’s abortion attitudes. These questions often assess slightly difference aspects of attitudes toward abortion legality. For instance, Pew has asked whether abortion should be legal in all/most cases or illegal in all/most cases whereas the General Social Survey has asked whether abortion should be legal considering six specific circumstances (e.g., rape, life endangerment). However, abortion is not simply legal or illegal; across states there are degrees of legislative restriction, resulting in differential ease or difficulty in accessing abortion. In that vein, Gallup and Harris have asked whether abortion laws should be made more or less strict and whether abortion should be easier or more difficult to obtain, respectively. We included these commonly asked abortion legality items on one survey, administered to US adults (n=894) via IPSOS probability panel. Using Latent Class Analysis, we identified three profiles of participants: (1) 35.0%--abortion should be illegal/more restricted, (2) 35.1%--abortion should be legal/laws should reflect the status quo, and (3) 29.9%--abortion should be legal/more available. Then, using multinomial logistic regression, we predicted latent class membership using states grouped by pre-Dobbs trigger laws and post-Dobbs viability laws. Although most participants believed abortion should be legal, most wanted abortion to be restricted (i.e., status quo; more strict), demonstrating complexity in attitudes toward legal abortion. Additionally, we found that states with the most restrictive laws comprise the largest proportion of people who think abortion should be illegal/more restricted whereas the least restrictive states comprise similar proportions of people from all three latent classes. Our findings suggest that claims of policy misalignment with public opinion may not be universally accurate. Instead, alignment may depend on geographic location and the policies within states. Implications for abortion attitude measurement will be discussed.
Cary, K.M., Jozkowski, K.N., Greer, K.M., Hoffacker, C., Turner, R., Crawford, B. & Lo, W. (2024, May 15- May 17). Who were you thinking about? Participant reports of individuals they were picturing while responding to an abortion attitudes survey. Association of Public Opinion Research (AAPOR) Annual Meeting, Atlanta GA.
People’s attitudes toward abortion may vary based on who they consider (e.g., friends, themselves, strangers) when responding to abortion survey questions. For instance, people may have more permissive attitudes when thinking of oneself or someone they know seeking abortion compared with strangers. We used mixed-methods to explore who participants reported thinking of when completing abortion-related survey items across two studies. In both studies, participants were randomly assigned to think of either (1) themselves/themselves and their partner or (2) a woman, when responding to abortion attitude items.
Study 1: Participants (N=608) were recruited via NORC’s AmeriSpeak panel to complete an abortion attitudes survey in which they responded to a check-all-that-apply item indicating who they were thinking of throughout the survey; 44% of participants selected multiple options. Participants selected “myself” (46.1%), followed by “general woman” (36.3%), “no one specific” (23.4%), “intimate partner” (23.2%), “family member” (15.1%), “teenager” (14.3%), “friend” (12.5%), “someone I know” (11.7%), and “future intimate partner” (7.6%). Responses to abortion attitude questions did not vary across condition.
Study 2: During cognitive interviews, participants (N=47) were asked who they were picturing while completing the survey. Approximately, 45% of participants mentioned multiple individuals. Specifically, 45% referred to women in hypothetical situations (e.g., young women, women experiencing rape), 40% reported thinking of someone they know (e.g., partner, family member, friend, someone who had an abortion), 38% reported thinking of a general woman, and 31% indicated thinking of themselves. Participants seemed more accepting/understanding when talking about women they knew personally, whereas 9% referenced women they heard of through anecdotes, typically expressing more criticism of such women.
Findings highlight the difficulty in teasing apart differences in response patterns as, regardless of survey condition, respondents considered a variety of women concurrently while responding to abortion-related survey items, and acceptance of abortion varied by who respondents considered.
Asamoah, N.A., Turner, R.C., Lo, W.J., Crawford, B.L., & Jozkowski, K.N. (2024, May 15- May 17)Investigating Changes in Self-Described and Perceived Statewide Abortion Identity Pre and Post the Dobbs Decision. Association of Public Opinion Research (AAPOR) Annual Meeting, Atlanta GA.
Abortion identity (e.g., pro-life, pro-choice) is complex and multifaceted and can be influenced by many factors such as family background and community. Considering the diverse landscape of abortion laws across the United States on the state level and the significant legal changes brought about by the overturning of Roe v Wade due to the Supreme Court’s decision on the Dobbs case, we are interested in how a person’s perception of the identity of most people in their state may be related to their own self-described abortion identity and to the Guttmacher Institute’s state abortion policy ratings before and after the Dobbs decision. The study uses longitudinal data collected in 2022 from an IPSOS panel consisting of 792 participants to allow for comparisons of these relationships.
Preliminary findings indicated there was an increase in the percentage of participants who claimed to be equally pro-choice and pro-life and a decrease in the percentage of those who claimed to be neither. A comparison of self-described abortion identity and the perception of others’ abortion identity revealed that while a smaller percentage of respondents describe themselves as equally pro-choice and pro-life, this was the category respondents most often thought the majority of people in their state identified as.
Further analyses will be conducted to ascertain if abortion identity and the perception of other’s abortion identity significantly impact responses to abortion legality and if these results differ since the overturning of Roe v Wade. In other words, are participants’ attitudes about abortion legality influenced by whether they perceive the attitudes of their community to be similar or different from their abortion stance? In summary, our research will explore the intricate dimensions of abortion identity, its relationship with perceptions of state sentiment, and the implications for abortion legality, especially in the wake of recent legal developments.
Cary, K. M., Crawford, B., Jozkowski, N. & Turner, R. (2024, May 15- May 17). Abortion Attitudes as Predictors of Abortion-related Helping Behaviors. Association of Public Opinion Research (AAPOR) Annual Meeting, Atlanta GA.
Cowan et al. (2021) identified that people’s willingness to provide different types of abortion-related support could represent an alternative way to capture people’s “abortion sentiments” than the 7-item abortion legality assessment in the General Social Survey (GSS). To explore the extent to which support intentions may relate to abortion attitudes, we examined how attitudes toward abortion acceptability, rights, and restrictions predicted willingness to provide various types of support to a family member or friend seeking abortion.
We used a Qualtrics opt-in panel to recruit a non-probability, quota-based sample. Participants (N=1,702) were asked whether they would offer various types of support to a close family member or friend seeking abortion. Responses were combined into a four-category variable: no support, only emotional support, only instrumental support (i.e., arrangements, direct costs), and both emotional and instrumental support. Particpants completed demographic items and indicated attitudes toward abortion acceptability, rights, and restrictions. Proportional raking was used to weight the data to demographics based on the 2018 American Community Survey.
A multinomial logistic regression model predicting helping behaviors using abortion attitudes, controlling for demographic variables, revealed participants who were pro-life, did not believe abortion is acceptable, and desired to remove abortion rights and implement more restrictions were most likely to indicate offering no support compared with other groups. Participants who were pro-choice, believed abortion is acceptable, and wanted more abortion rights and fewer restrictions were more likely to indicate instrumental or both emotional and instrumental support compared to other groups. Likelihood of providing emotional support did not vary across abortion attitudes. Women (compared to men) were more likely to indicate both emotional and instrumental support. Interestingly, willingness to provide any type of support did not vary across political affiliation.These findings advance 2018 GSS results by examining helping behaviors in relation to other assessments of abortion attitudes.
Crawford, B., Jozkowski, N., Fordyce, E., Avripas, S., Turner, R.; Lo, W.J., Stern, M., & English, N. (2024, May 15- May 17). The Salience of Abortion As a Political Issue after the Dobbs v. Jackson Decision. Association of Public Opinion Research (AAPOR) Annual Meeting, Atlanta GA.
Despite abortion consistently being at the heart of election and political discourse for several decades, the salience of abortion to voters is complicated. For example, although people may identify abortion as an important issue in the abstract, when asked to rank its importance relative to other issues, abortion consistently ranks as one of the least important. However, the 2022 ruling in Dobbs v. Jackson and subsequent overturning of Roe v. Wade substantially changed the abortion policy landscape in the U.S. by removing the constitutional right of a pregnant person to choose to have an abortion up until the point of viability and allowing states to pass near total abortion bans from the beginning of a pregnancy. Because of this shift, politicians and voters have a larger and more direct effect on abortion policies. Specifically, since the Dobbs v. Jackson decision, state politicians have proposed and passed pre-viability bans (including near total bans) and voters in several states have had a direct say on abortion policies through state ballot initiatives. Given these changes, we examine if and how the salience of abortion has shifted after the Dobbs v. Jackson decision.
Eight waves of cross-sectional data were collected using NORC’s Amerispeak Omnibus (N=1,003-1070), with the first wave of data collection launching the day before the Dobbs decision and the last wave launching approximately one year after the decision. Participants were asked to rank the importance of abortion relative to five other social issues. In our presentation, we will highlight that the salience of abortion increased after the Dobbs decision. We will also discuss how these increases vary across socio-demographic groups, ideologies, and states. We will conclude by discussing how these shifts in salience align with recent election results and the potential future implications for voters, politicians, and elections.
Poster
Bueno, X.; Hawbaker, A.; Hoffacker, C.; Crawford, B.; Turner, R.; Lo, W.J. & Jozkowski, K.N. (2024, May 15- May 17): ‘Some of the Questions Got Me Thinking...’: Survey Effects on the Development of Abortion Attitudes.American Association of Public Opinion Research (AAPOR) Annual Meeting, Atlanta, GA.
Although research indicates responses to abortion questions can be impacted by survey design, the extent to which abortion attitudes are fixed or malleable in a measurable period of time (e.g., while answering a survey) is underexplored. In this mixed-method study, we assessed whether English- and Spanish-speaking survey participants changed their response to abortion attitude items while completing web-based surveys. We analyzed two online surveys where the same question was asked at the beginning and end of each survey. Study 1, a 2021 survey administered via of Qualtrics© to 2,204 US adults, asked whether abortion should be ‘legal in all cases,’ ‘legal in most cases,’ ‘illegal in most cases,’ or ‘illegal in all cases’ at both the beginning and end of the survey, followed by an open-ended question that asked why the participant did or did not change their response. Similarly, Study 2, a 2021 survey of 1,025 US adults from NORC’s Amerispeak© panel asked participants how they identify with the terms pro-choice and pro-life on a 9-point scale. Results from Study 1 indicate that around 25% of respondents changed their response (n=578). Among those who changed their answer, 31% went from legality (in all or most cases) to illegality (in all or most cases) while 21% changed from illegality to legality. Open-ended analysis showed that 26% of the 406 respondents who changed their views and provided a valid open-ended response indicated that the survey itself affected their views toward abortion (i.e. “[my answer] changed because the survey made me think more about it…”). Results from Study 2 indicate that 38% of survey respondents modified how they identified with abortion labels within the 9-point gradient. Further analysis will explore the nuances behind initial findings and make recommendations for survey development.
Wang, X.; Turner, R.; Asamoah, N,; & Lo, W.J. (2024, May 15- May 17): Comparing 4-Point and 5-Point Scales with Equal and Unequal Intervals in Survey Response Options: A Study on Acculturation Measurement .American Association of Public Opinion Research (AAPOR) Annual Meeting, Atlanta, GA.
Research has explored the relationship between the design of rating scale and survey data quality. Some studies recommended factual questions with fewer response points and without midpoint ratings tend to yield more reliable responses. Additionally, linguistic labeling with equal intervals has been shown to result in stronger correlations compared to scales with unequal intervals. However, there is ongoing debate about the optimal point-scale type and anchor terms for survey questions.
This study aimed to evaluate a 5-point versus a 4-point scale on measuring acculturation levels among participants in both Hispanic and non-Hispanic cultural groups in the U.S. The Bidimensional Acculturation Scale (BAS) was employed, with 12 items divided into 'frequency' and 'ability' subgroups. The 5-point format introduced an additional anchor, 'Seldom,' within the 'frequency' subgroup and 'Acceptable' within the 'ability' subgroup. Data was collected in Qualtrics from Feb to May 2021, consisting of 2,050 participants in the English version.
Results revealed that participants displayed a positively skewed distribution in their responses, with the 4-point format often leading to higher percentages of the most extreme options. However, this pattern shifted in the 'frequency' subgroup due to the addition of 'Seldom,' leading to higher selection rates for 'Sometimes' in the 5-point format. Composite scores between the 4-point and 5-point formats did not show statistically significant differences when considering all items together. However, significant differences emerged when analyzing the 'frequency' and 'ability' subgroups separately.
In conclusion, the study suggests that adding an anchor point (i.e., 5-point) may significantly impact response patterns and composite scores. It also raises questions about the positioning of added points in survey response options and their implications for data quality and psychometric properties. Further research is needed to explore the optimal characteristics of response scales for various survey contexts and topics.
Mena-Meléndez, L.; Lohrmann, C.; Hoffacker, C.; Crawford, B.; Turner, R.; Lo, W.J.; Jozkowski, K. (2024, May 15 - May 17): Exploring People’s Attitudes Towards Punishment Endorsements for a Woman and a Healthcare Provider Who Had or Who Provided an “Illegal” Abortion. American Association of Public Opinion Research (AAPOR) Annual Meeting, Atlanta, GA.
Since the overturn of Roe v. Wade in 2022, many states have banned abortion (or are likely to do so), codifying specific legal consequences for those seeking or providing an “illegal” abortion. The current study is part of a multi-study effort to assess attitudes toward whether—and how—different actors should be punished if abortion were to be “illegal.” Using 2021 survey data collected from English- (n=2,204) and Spanish-speaking (n=285) US adults via Qualtrics’ national panels, we explored participants’ open-ended responses (n=4,978 total) for why there should be a penalty/no penalty for a woman (n=2,489) and for a healthcare provider (n=2,489) who had/provided an “illegal” abortion. We aimed to answer: (1) Why do respondents think there should be a penalty/no penalty for a woman?; (2) Why do respondents think there should be a penalty/no penalty for a healthcare provider?; and (3) Are there language differences across responses? Preliminary findings indicate that for a woman, people think there should be no penalty because “it is the woman’s body and decision” and “abortion is not murder.” For the provider, there should be no penalty because “providers are just doing their job,” “it is the woman’s body and decision,” and “abortion is not murder.” For both the woman and provider, people think there should be a penalty because “abortion is illegal, a crime, and breaking the law,” “there should be consequences for breaking the law,” and “abortion is murder.” We also observe differences across language subgroups for English and Spanish respondents. Findings suggest that the people’s attitudes vary by actor (e.g., woman vs. provider) and by survey language (e.g., English vs. Spanish). We will discuss implications for measuring public opinion regarding abortion legality and consequently punishments associated with “illegal” abortion in a post-Dobbs v. Jackson landscape, including implications for multi-language survey administration.
Turner, R.C., Jozkowski, K.N., Lo, W.-J., Crawford, B.C., & Bueno, X. (2024, May 15-17). Are the impacts of contextual questions on attitudes about abortion temporary or persistent? Poster presentation at the American Association of Public Opinion Research annual conference, Atlanta, GA.
In 2022, we conducted an experiment using a nationally representative sample [IPSOS KnowledgePanel, n = 1014] to investigate whether asking contextual questions about pregnancy and abortion [e.g., 1) how far into a pregnancy people typically find out they are pregnant, 2) how far into a pregnancy people typically obtain abortions, 3) when specific fetal milestones occur, and 4) how safe pregnancy and abortion are for a woman], influenced how participants responded to questions about when, during a pregnancy, abortion should be legal. One-half of the sample received the abortion legality questions prior to the contextual questions about pregnancy and abortion, and the other half received them after. Results indicated that item order did not impact responses to pregnancy, abortion, fetal milestone, and safety questions. However, there were significant impacts on responses to abortion legality questions with shorter timelines for abortion legality indicated when they were presented prior to pregnancy and abortion contextual questions. We repeated the experiment 4-5 months later with the same participants [n = 792], to investigate whether the impacts of pregnancy and abortion contextual questions are temporary or persistent. The use of both waves of data provides a four-group design for comparison. It is important to note that the goal of our surveys is to measure attitudes and not change attitudes; however, it may be informative to know if helping people think through components of a decision, prior to asking them to make a decision, results in answers that are more stable over time. The results of our study are designed to inform researchers about whether the use of the contextual variables provided assistance with cognitive processing that resulted in longer-term attitude development when measuring abortion attitudes, or whether the contextual variables seemed to have a temporary survey-constrained impact on responses.
Lo, W.-J., Turner, R.C., Crawford, B.C., Jozkowski, K.N., Bueno, X., & Mena-Melendez, L. (2024, May 15-17). Does Providing the Midpoint Improve Measurement Quality? A Comparison between Binary and 5-Point Likert-Type Scale Using Latent Factor Construct. Association of Public Opinion Research (AAPOR) Annual Meeting, Atlanta GA.
In the past, researchers often measured attitudes and perceptions using a binary scale. However, after new technology is implemented in the 21st century, it is no longer a barrier for researchers who want to include more response categories in their surveys. This research delves into the comparative analysis of binary scales and 5-point Likert-type scales concerning the reliability and validity of measuring people's attitudes toward abortion. Although this study is grounded in the recognition of the nuanced nature of abortion attitudes, the primary goal is investigating the potential influence of scale types on the measurement reliability and validity.
Employing a multi-wave design, this study explores the test-retest reliability of survey instruments employing two-point and five-point scales, adapting Green's (2003) foundational work on a study for assessing the internal consistency of measures over time under the latent construct. In addition, since both binary and 5-point scales are provided to each participant using modified GSS abortion items, we could investigate people who chose the midpoint while their responses were Yes/No in the previous set of questions and examine latent constructs (i.e., validity) across multiple time points. The study employs a diverse sample to ensure comprehensive insights into the reliability and validity of both binary and Likert-type scales. Participants are surveyed at multiple waves, allowing for the examination of temporal stability and consistency in their responses.
The findings aim to contribute to the methodological discourse surrounding scale selection in attitude measurement. By comparing the psychometric properties of binary and Likert-type scales, this study seeks to inform researchers and practitioners about the implications of scale choice for studying attitudes toward sensitive and multifaceted topics like abortion. The implications of this research extend to the refinement of survey instruments, promoting more accurate and reliable assessments of individuals' attitudes toward abortion over time.
Jozkowski, K.N. Avripas, S., Crawford, B.L., English, N., Fordyce, E., Lo, W.-J., Mena-Meléndez, L., Stern, M. & Turner, R.C. (2024, May 15- May 17) Examining Trends in People's Attitudes Toward Legal Abortion Before and in the Year After the Dobbs v. Jackson Women’s Health Organization Decision. Association of Public Opinion Research (AAPOR) Annual Meeting, Atlanta GA.
The Dobbs v. Jackson Women’s Health Organization decision resulted in significant changes to the abortion legislative landscape. As a result, 14 states restricted abortion throughout pregnancy (with rare exceptions) and several other states restricted abortion at different weeks’ gestation (e.g., six weeks; fifteen weeks). Additionally, in some states, ballot initiatives brought the issue of abortion legality/constitutionality directly to voters. Advocates and media have claimed that people’s attitudes toward abortion have changed in response to Dobbs v. Jackson. Using trend data collected just prior to the Dobbs v. Jackson decision and in the year that followed, we examined potential changes in people’s attitudes toward legal abortion.
Eight waves of cross-sectional data were collected using IPSOS’ KnowledgePanel (Nwave1=1,1014) and NORC’s AmeriSpeak Omnibus (Nwaves2-8=1,003-1070). Participants were asked eight items regarding whether abortion should be legal considering different circumstances (e.g., pregnancy resulted from rape; pregnant woman wants to focus on her career) at different weeks’ gestation—never, six, fifteen, and twenty-two weeks, and throughout pregnancy. Disregarding weeks’ gestation, more participants endorsed legal abortion for some circumstances (e.g., rape) than others (e.g., focus on career). Disregarding circumstance, more participants endorsed legal abortion at earlier weeks’ gestation than later. Most notably, we found little to no variation in attitudes over time, across circumstance and weeks’ gestation. There were slight, but not statistically significant, differences in attitudes at Wave 1 (pre-Dobbs), compared with Wave 2 (collected immediately following the decision), with greater endorsement of legal abortion at Wave 2. However, this slight increase dropped by Wave 3 and persisted through all eight waves. Our findings suggest that people’s attitudes did not change post-Dobbs. However, the context in which we are considering these attitudes changed dramatically. We will discuss implications for our findings in the context of state ballot initiatives and other changes to the abortion legislative landscape.