Individual
McClelland, S., Carrillo, D., Dutcher, H., Jozkowski, K.N. (2019, August 8-11). What's in the item counts too: Qualitative analysis of abortion measures in U.S. research. American Psychological Association (APA) Annual Meeting. Chicago, IL.
Researchers’ decisions when designing survey items often have enormous (but invisible) power in documenting and even shaping public opinion. This is especially true when assessing controversial issues such as abortion. It is, therefore, crucial to examine how researchers ask questions in survey designs. In this paper, we turn to the items, scales, and measures used to measure abortion attitudes as our object of analysis. We assess the machinery itself – survey items -- and ask how they structure and determine the kinds of information gathered, about whom, and with what biases in place.
This paper details how an item bank was used to study the gendered and racialized imagery woven into public opinion measures about abortion attitudes. We used systematic review methods to search peer-reviewed articles published in the past decade (2008-2018) that measured attitudes towards abortion with U.S. respondents. From this search, we developed an item bank comprised of 456 items, drawn from 81 studies, which included large, national studies with probability-based samples, as well as smaller studies with non-probability samples. Three trained coders content coded and analyzed the data at the item and scale level; procedures included the construction of a detailed codebook, discussion of interpretations, and consensus building for final coding decisions.
Qualitative insights from analysis of this item bank included how “women” or “a woman” who is considering an abortion was consistently described in survey items (e.g., “Abortion is acceptable if the woman cannot take care of her child”). Qualitative analysis of items that reference a women who seeks abortion care highlight how survey respondents are asked to evaluate their abortion attitude with negative associations of both the woman who would be a mother, as well as the woman who would make an abortion decision (and should not be trusted with either decision). This leaves abortion attitudes – and their evaluation – often tied to negative feelings about women in general, women of color who are already associated with poor mothering, and positions respondents as moral arbiters in issues of reproduction. Similarly, moral evaluations were evoked in 29 percent of items, including, for example, describing abortion as “available,” “accessible,” or “allowed” as compared to “legitimate,” “acceptable,” or “elective.” These and other quantitative and qualitative patterns help to demonstrate several levels of meaning that are embedded within measures of abortion attitudes.
Research has the potential to affect how ideas circulate in contemporary debates in the U.S., how resources are distributed, laws imagined and passed, and how justice claims are made or forgotten. Qualitative analysis of quantitative survey measures offers an important aspect of understanding the power of research in shaping how and what we think and the policies we live with (and without). Developing a comprehensive item bank – and content analysis methods used to analyze item structure and wording – is an innovative contribution to the field of research methods, abortion research, as well as other subfields such as feminist methods and reproductive justice research.