Who cares about human rights? Public opinion about human rights foreign policy

ABSTRACT

Literature on foreign aid and human rights often presupposes that constituents favor using foreign policy to promote human rights abroad and lead elected policymakers to pursue such policies to retain electoral support. This assumption, although frequently asserted, has not been empirically evaluated. And there are reasons to be skeptical about whether public opinion supports human rights foreign policy compared to other policy objectives. This article explores US public opinion about human rights, by asking two questions: Does the public think human rights should factor in foreign aid decisions and does the context—the strategic or economic relationship between the donor and recipient—affect this? This article uses results from a nationwide experimental survey to evaluate these questions. I find that the majority of respondents support cutting aid to punish human rights violators and that this depends minimally on the importance of the recipient.

View correction statement:

Corrigendum

Notes on contributor

Michelle Giacobbe Allendoerfer is an Assistant Professor of Political Science at George Washington University with a primary appointment in the Women's Leadership Program, a freshman live-and-learn community. Her research focuses on international human rights. She received her PhD from the University of Michigan in 2010.

Notes

1. Scholars have also examined the relationship between foreign aid and human rights in cross-national datasets, evaluating Organisation for Economic Cooperation and Development (OECD) donor countries foreign aid patterns. Results, much like in the US context, are largely mixed. In two studies based on case studies, Barratt (2008) and Gillies (1996) argue that donors only punish violators when their own interests—economic or strategic—are not at stake. Results in quantitative work are also mixed; some scholars find that some donors sometimes condition foreign aid on human rights (Neumayer 2003). In a recent article, Nielsen (2013) attempts to explain the mixed findings and argues that OECD donors use foreign aid selectively, reducing economic aid to recipients that violate human rights while not cutting other types of aid (social and humanitarian).

2. Because the survey will include deception—manipulation of Chad's human rights record and the US relationship with Chad—the experiment ended with a debrief informing the respondents of this fact.

3. The full survey is available in the online appendix at the author's website.

4. An important distinction in the framing literature is between equivalency frames (e.g., the difference between 95% employment and 5% unemployment) and emphasis frames (for a review of this literature, see Chong and Druckman 2007). This study is more closely related to the latter type of frame, although not perfectly as the difference between prompts are less about the emphasis of the two dimensions (human rights and recipient importance) but also the content of those dimensions. That said, the framing literature is relevant in thinking about how the survey prompts could affect public opinion about foreign aid by framing the issue of foreign aid in terms of human rights.

5. In this survey, I do not explore how alternative ways of framing human rights abuses affects public opinion or behavior. See McEntire et al. (2015) for work that evaluates the effectiveness of different types of frames.

6. The Amnesty International prompt is adapted from existing language of Amnesty International releases to most accurately reflect the language and tone of such news releases. This one is largely drawn from a 2013 release on Sudan (Amnesty International 2013).

7. Respondents were randomly assigned to one of six treatment groups, using Qualtrics randomization function. Due to an error in the original survey, one treatment was included twice and another omitted. I released the survey again with the originally omitted treatment only. In other words, for Treatment Group D, respondents were not randomly assigned. All respondents on June 10, 2014 were in the same group. There is no a priori reason to believe that all respondents on this given day would have some characteristic that would damage the inferences drawn.

8. The results are substantively similar when political ideology is used in place of political-party affiliation.

9. Concerns about external validity remain if there is reason to believe that MTurk samples behave differently than other samples in ways that affect replicability of these results. In a comparison of three sampling methods — undergraduate students, MTurk, and YouGov—Krupnikov and Levine (2014) do find that MTurk samples can produce different results than the other samples on experiments that require more “buy-in” (i.e., reading an article or trusting information from the experimenter) from the respondent. As my study requires little “buy-in” from the respondent and is most similar to their simple experiment, similar results should be produced across sampling methods.

10. A complete table with probit results is available in Appendix II (Table A2).

11. A complete table with all coefficients is available in Appendix II (Table A3).

12. Full probit and multinomial results tables are available in the appendix.

13. This discussion of policy-specific knowledge is not complete without acknowledging that prior to reading the treatment prompts, I inform the respondents of the actual percent of the federal budget the United States allocates to foreign aid. According to Gilens (2001), providing this policy-specific knowledge may reduce opposition to foreign aid. This may account for the insignificant findings when this measure is used to capture political knowledge.