Read on Twitter
THREAD. What to know about public opinion on abortion. It's all DISCONNECT right now. Btwn past and present. Btwn conventional wisdom and reality. Btwn opinion and policy.
1) Context: About 65% to 77% of voters support abortion rights and access across many different measures. (Note: non-voters tend to be more split).
2) We've seen MAJOR shifts among pro-choice voters over the past several years - the significance of which is neither well understood nor reported:
3) FYI: We have not seen evidence of change in these areas:
4) Lots of DISCONNECT between conventional wisdom and reality, for instance:
5) More on party: Democratic voters are very unified, Republicans more split. We've seen this for about three years now. Also: A survey of Iowa caucus voters from last summer shows abortion is the top "must have" in their list
6) Conventional wisdom: This issue is about views toward life and ending a life. Reality: This issue is also very highly correlated with beliefs about WOMEN:
7) BTW: Is there a patriarchy of polling? I think, yes. Would a male pollster think about analyzing views toward abortion by beliefs and attitudes of women? 
Would a male pollster prioritize this issue and analysis over every other relevant policy issue? (Would male editors?)
Would a male pollster prioritize this issue and analysis over every other relevant policy issue? (Would male editors?)
8) Big DISCONNECT between media coverage / conventional wisdom and reality. It is very easy to conclude from some polling data that abortion is "polarizing" "divisive" "nuanced" "complex" and people "struggle" with their views. @UpshotNYT https://www.nytimes.com/2019/06/08/upshot/politicians-draw-clear-lines-on-abortion-their-parties-are-not-so-unified.html
9) A few weeks after this article came out, I had the opportunity to ask this question in survey of likely 2020 voters:
10) Do people think abortion is a complex issue, based on each person's circumstances? Absolutely. Do most voters struggle with their position on abortion? They do not.
11) Brilliant researchers and journalists rely on survey questions from very well-respected sources: Gallup, the GSS, etc. But, many of these measures lack "construct validity." They are not measuring what we think they are measuring. For example:
12) Here's another simple example: Gallup finds forever the public is "divided" between "pro-choice" and "pro-life" people. But, actually, almost four in ten don't identify with either label.
13) But Gallup, GSS, et al. keep asking the same questions because they want to track change over time. Measuring change is important, but that doesn't mean what you are tracking is valid or accurate.
14) Why aren't there better survey questions on abortion? MANY reasons, including #7 above. Another reason is that public pollsters who write and repeat flawed questions do not conduct qualitative research (focus groups, etc.). But qual is where the LEARNING is. For example:
15) If you understand public opinion from qualitative research, you'll know NOBODY knows ANYTHING about abortions later in pregnancy. So you'll know these %s aren't quite right - from a @HarvardChanSPH poll https://cdn1.sph.harvard.edu/wp-content/uploads/sites/21/2016/08/STAT-Harvard-Poll-August-2016-Zika.pdf
16) And you'll ask this - which gives some context - and as a result, provides a truer sense of where the public is: 
@HarvardChanSPH
@HarvardChanSPH
17) You'll also be able to spot survey questions that really get things wrong - like this one from Knights of Columbus, which produced headlines claiming most Americans want substantial abortion restrictions.
18) We replicated this and found a majority of respondents said they do not want politicians to pass new laws that reflect their response in the KoC question. https://view.publitas.com/perryundem-research-communication/2019-abortion-policy/page/18
19) Based on the data he's using, @Nate_Cohn, understandably, ends his article with this:
20) I would too. But he's not looking at other data, such as (73% don't struggle with their views and):
21) And actually, the more I learn, the simpler and MORE straightforward I think opinion is. If I had to use one question to best summarize where voters are on this issue right now, it'd be this one:
Bottom line: If you are reading polling data or an article that feels familiar and does not challenge conventional wisdom, check yourself. It is at best incomplete, and at worst, flat out wrong.
PS: Pollsters who are repeat Qs from the 1970s (e.g., "Should a woman be able to have an abortion for any reason?") might reflect on this (patriarchal?) language. Women's "reasons" have been irrelevant in our legal framework under Roe for the past 48 years....Or...
We could start asking these questions:
