Rising Income Inequality and the Perceived Social Structure of Economic Opportunity

Download data and study materials from OSF

Principal investigators:

Pia Dietze

University of California, Irvine

Email: pdietze@uci.edu

Homepage: https://www.piadietze.com/

Leslie McCall

The Graduate Center, City University of New York

Email: lmccall@gc.cuny.edu

Homepage: https://www.gc.cuny.edu/people/leslie-mccall

Jennifer A. Richeson

Yale University

Email: jennifer.richeson@yale.edu

Homepage: https://psychology.yale.edu/people/jennifer-richeson


Sample size: 1760

Field period: 07/28/2023-11/17/2023

Abstract
We are examining whether Americans believe that equal opportunity exists for groups marginalized by race, ethnicity, gender, or social class, and whether these beliefs are affected when rising economic inequality is foregrounded in an experimental condition. In past studies, we found that Americans perceive a clear social structure of economic opportunity — a structure of dis/advantage across class, race, ethnicity, and gender groups. Our results also showed that these perceptions can be exacerbated when individuals are exposed to information about rising inequality. In our proposed follow-up study, we will test if our findings replicate with 1) simplified stimuli and 2) a modification in our questionnaire in which we ask about perceptions of 'unfair' dis/advantages (not simply dis/advantages as in our prior studies). In addition, we will test a potential mechanisms underlying perceptions of a social structure of economic opportunity — perceptions that different dimensions of inequality are interconnected — and we will include two additional outcome measures (i.e., affirmative action support/opposition for racial/ethnic minorities and low-income individuals).
Hypotheses

Our main hypothesis, consistent with our prior studies, is that the inequality treatment results in perceptions of unequal opportunities for an outcome composite of low-economic-status groups (women, Blacks, Latinxs, and people from low-income families), such that reading about inequality of outcomes decreases perceptions of “unfair advantages” for the four low-economic-status groups.

A second hypothesis is that a composite of four items assessing our proposed mechanism, viewing inequalities across dimensions as interconnected, statistically mediates the treatment effect that individuals exposed to information about rising economic inequality will perceive low-economic-status groups as having fewer advantages in getting ahead.

A third hypothesis is that the inequality informational treatment will increased support for affirmative action programs by both class and race/ethnicity.

Experimental Manipulations
Respondents are randomized into one of two manipulations. The treatment condition presents a short (approximately 300 words), non-partisan, descriptive, news story describing the rise in income and earnings inequality in the United States and is accompanied by a graph charting the growth in incomes across quintiles and for the top one percent. There is no mention of racial or gender inequality in the text. The control condition includes statistical information of a similar nature and in a similar format and style, accompanied by a graph, but the topic is baseball statistics. These conditions are a simplified version of those presented in McCall, Burk, Laperierre, and Richeson (2017).
Outcomes

MAIN OUTCOME (composite of responses to question when low-economic-status groups are the target):

We have some questions about opportunities for getting ahead. Specifically, the questions are about unfair advantages that might provide some people opportunities to get ahead. Please indicate to what extent the following groups have unfair advantages in terms of getting ahead using the scale provided.

To what extent do ___________________ have unfair advantages in getting ahead?
Target group wording: people from upper-income families, people from lower-income families, men, women, White Americans, Black Americans, Hispanic Americans, Asian Americans.
Response categories: 1 = Not At All, 2 = A Little Extent, 3= Some Extent, 4 = A Large Extent, 5 = A Very Large Extent.

Source: Authors’ adaptation of ISSP questions to separate social class, gender, and racial/ethnic groups. Groups appeared in randomized order within class, gender, and race/ethnicity categories, with the class categories presented first, followed randomly by gender or racial/ethnic categories.

MEDIATOR OUTCOME (composite of four items):

“The struggle for income equality in America should be closely related to the struggle of other disadvantaged groups”
“The discrimination that low-income individuals have experienced is similar to that of other disadvantaged groups”
“There are other people who experience injustice and indignities similar to low-income individuals”
”Low-income individuals will be more successful in achieving their goals if they form coalitions with other disadvantaged groups”
Response categories: 1 = Strongly Agree; 2 = Somewhat Agree; 3 = Slightly Agree; 4 = Neither Agree Nor Disagree; 5 = Slightly Disagree; 6 = Somewhat Disagree; 7 = Strongly Disagree.
Source: Authors’ adaptation of validated scale developed to measure an ideological sub-component of Black identity (Sellers, Rowley, Chavous, Shelton, & Smith, 1997).

POLICY OUTCOMES:

“Do you favor, oppose, or neither favor nor oppose allowing universities to increase the numbers of racial minorities studying at their schools by considering race along with other factors when choosing students.”
“Do you favor, oppose, or neither favor nor oppose allowing universities to increase the numbers of poor and low-income students stuying at their schools by considering family income along with other factors when choosing students.”
Response categories: 1 = Favor a great deal; 2 = Favor a moderate amount; 3 = Favor a little; 4 = Neither Favor Nor Oppose; 5 = Oppose a little; 6 = Oppose a moderate amount; 7 = Oppose a great deal.
Source: Authors’ adaptation of questions from the 2016 ANES.

Summary of Results
Results are mostly null for the main (first) hypothesis for the pooled sample of all respondents. Some support is found for the last hypothesis concerning policy preferences. Analysis of the second hypothesis has not been undertaken yet.