search

Watts-Lab / surveys

Library of surveys for deliberation experiments
MIT License
3 stars 4 forks source link

Implement affective polarization measure #20

Closed JamesPHoughton closed 3 months ago

JamesPHoughton commented 2 years ago

One of the outcomes we hope to measure is a change in affective polarization regarding individuals who express different opinions. A useful reading is Druckman and Levendusky 2019: https://paperpile.com/shared/38cfuM

This will require some templating to insert the appropriate out-group, discussed here: https://github.com/Watts-Lab/surveys/issues/3

JamesPHoughton commented 2 years ago

Also see surveys in:

Levendusky, Matthew S. 2018. “Americans, Not Partisans: Can Priming American National Identity Reduce Affective Polarization?” The Journal of Politics 80 (1): 59–70.

referenced in first paragraph of pg 62

JamesPHoughton commented 2 years ago

In conversation with @isaacgat, we decided that there are several ways to think about this, each of which measures something subtly different, depending on who we ask participants to consider.

  1. ask about the generalized other party (or unnamed cross-partisans), and measure whether the discussion changed the way they think about the out-group as a whole, and whether their experience in the deliberation will improve their approach to that group in the future. This is useful in contexts where we hope that deliberation can be a tool for reducing cross-party acrimony in general.
  2. ask about the other people in the discussion group in general, without calling out specific names. This measures something possibly akin to the team viability survey.
  3. ask about a particular alter from the deliberation that the participant identifies themselves. We could ask them to think about the person they most disagreed with, and the person they most agreed with. This measures willingness to work with particular individuals, emphasizing the most contentious relationships and those that are most likely to cause problems for the group's work in the future.
  4. ask about two uniquely assigned individuals in the group, such that every person is rated by two others, assigned randomly. This is similar to above, but gets a balanced sample of the relationships across the group.
  5. ask about two uniquely assigned individuals, who are known cross-partisans.

At some point down the road we may wish to measure more than one. For the purposes of the summer, we want an additional outcome that is:

I would suggest then that we think about implementing number 3 - ask the participants to identify the person they disagreed with most, and the person they agreed with most.

JamesPHoughton commented 3 months ago

implemented as partisan feeling thermometers and trait ratings. Could also do a social distance measure in the future: https://github.com/Watts-Lab/surveys/issues/153