Who Do We Picture When We Picture the Poor? Race and the Welfare State, Part 1

POSC 459 — Social Welfare Politics and Policy
Week 12 · Monday, November 9

Gilens 1999, ch. 3, skim ch. 5 · Howard 2007, ch. 9

Today

  1. Last week's bridge
  2. Who gets pictured
  3. Gilens: why oppose welfare?
  4. Howard: "Race Still Matters"
  5. Take a position

Last week's bridge

The film: "Go back to your Section 8 housing"

Soss and Schram: "lack motivation," still tied to opposing welfare spending after 1996

Today: how did welfare come to be seen through race?

Week 11 film · Soss and Schram, p. 117

Share of the poor who are Black

27% actual · 50% guessed

Gilens's figures, 1999 · median answer in one survey

Gilens, p. 68; repeated p. 102

Who gets pictured

Time, Newsweek, U.S. News Black
Poor people pictured, 1950–92 53.4%
Actual poor, average for the same years 29.3%
Poor people pictured, 1967–92 57%

1967–92: "about twice the true proportion"

Gilens, pp. 113–14 · newsmagazines, not today's feeds

Gilens's puzzle

Support: helping the poor; welfare in principle

Oppose: welfare spending

Favor cutting spending on
"welfare" 63%
"people on welfare" 71%

Gilens, pp. 60, 63–65

Who gets labeled "undeserving"?

Most people on welfare... Agree
are "taking advantage of the system" 66%
"could get along without it if they tried" 61%
who can work "try to find jobs" 31%

Gilens, table 3.1, p. 62 · surveys, 1994–95 and earlier

Beliefs about need track spending preferences

"Most welfare recipients could get by without it if they really tried"

Respondent Cut welfare Increase welfare
Strongly agrees 59% 18%
Strongly disagrees 22% 43%

Gilens, figure 3.1, pp. 64–65 · National Race and Politics Study, 1991

Where race comes in

"Welfare" as a "race-coded" topic

Rated Black Americans as Cut welfare Increase welfare
Hard-working (1 of 7) 35% 47%
Lazy (7 of 7) 63% 15%

Gilens, p. 67; figure 3.2, pp. 68–69 · General Social Survey, 1990 and 1994

Which racial attitude?

How to read it: rows are Gilens's labels for beliefs measured in a 1986 survey of white respondents. Columns are outcomes: cutting food stamps, negative feelings toward welfare recipients, and the two combined. Each number is points of opposition, 0–100, between one end of the belief and the other, with the other five held constant.

White respondents' belief Cut food stamps Negative feelings Combined
"Blacks lack work ethic" 20.1** 14.3*** 17.2***
"Civil rights leaders push too fast" 14.2* 4.3 11.7**
Opposition to affirmative action 14.8* 8.5 10.4*
"Helping blacks is not government's job" 11.2* 1.4 4.1
Negative feelings toward Black people 0.7 8.7 3.1
"Blacks are innately inferior" −4.3 5.1 −0.2

Top right: 17 points. Bottom two rows: no statistically clear link. Stars mark results that would be unusual if there were no association; they aren't proof of cause.

Gilens, table 3.2, p. 77 · 1986 NES · 357–381 people · * p < .05, ** p < .01, *** p < .001

What chapter 3 can't settle

  • Surveys, one moment at a time
  • 1986 NES: white respondents, Ns 357–381
  • Food stamps stand in for welfare
  • Black respondents differ: 33% cut, 38% increase (whites: 53%, 17%)
  • Harder statistical test is chapter 4, not assigned

Gilens, pp. 72, 76, 79

The jump

Years Poor people pictured who are Black
1950–59 18%
1964 27%
1965 49%
1966 53%
1967 72%

Actual share of the poor, early to late 1960s: 27% to 30%

Gilens, pp. 114, 117

Same pattern, by topic

Stories on Pictured Black
Underclass 100%
Urban problems 84%
Welfare, antipoverty programs 54%
Children 51%
Medical care 28%
Hunger 25%
Old-age assistance 0%

Gilens, table 5.2, p. 128 · 1950–92 · selected topics

Schneider and Ingram: contested political labels

Power Positive construction Negative construction
Strong Advantaged: the elderly, business, veterans Contenders: the rich, big unions, minorities
Weak Dependents: children, mothers, disabled Deviants: criminals, drug addicts, gangs

Their map of how officials may perceive groups. Minorities' placement is contested.

"the poor were constructed as lazy or shiftless and were often believed to be minorities"

Schneider and Ingram 1993, figure 1, p. 336; p. 345 · not assigned

Howard, chapter 9: "Race Still Matters"

The common view: race has faded from social policy; welfare is "the big exception"

Howard: "has not disappeared, nor is it limited to welfare"

How: the states, about 2000; benefits, not poverty

Form: a student and a teacher work it out

Howard 2007, pp. 178–81

Howard's regressions: TANF, five models

How to read it: each column is one regression across the states, predicting the maximum monthly TANF benefit in 2000, holding income, education, urbanization, poverty, and party constant. Population and officials figures are dollars per percentage point. A dash: not in that model.

Model 1 Model 2 Model 3 Model 4 Model 5
Black population −7.458** — — — —
Black + Hispanic population — −4.568* — — —
Racial attitudes — — 6.474** — 6.195**
Black elected officials (% of all officials) — — — −12.835** −10.760*
Adjusted R² .552 .448 .524 .509 .590
N 49 49 44 49 44

Model 1: one more percentage point of Black population, $7.46 less a month. Attitudes: higher means more tolerant; Howard's units are unclear, but least to most tolerant is about $250 a month. Adjusted R²: share of variation explained. N: states.

Howard, table 9.1, p. 187 · stars: unusual if no association, not proof of cause · * p < .05, ** p < .01

Past welfare: what the states show

How to read it: each row summarizes comparisons across the states; SCHIP uses simple correlations without controls. A negative figure means states with more Black or Hispanic residents pay less, holding income, education, urbanization, need, and party constant. The last column asks whether a state's racial attitudes showed a reliable link to that program's benefits.

Program Per 1 point more minority population, with controls Racial attitudes significant?
TANF, maximum benefit −$7.46** a month (Black) Yes
Medicaid, elderly −$126* a year per patient (Black + Hispanic) No
Medicaid, children and disabled Not significant Not reported
Unemployment insurance −$1.72** a week (Black + Hispanic) No
Workers' comp Fewer of 19 "essentials" of a program No
SCHIP Black share r = .24, Hispanic share r = −.23; neither significant Not reported

Howard, tables 9.1–9.3 and pp. 182–84, 191 · state benefits, 1999–2001 · * p < .05, ** p < .01

Where Howard complicates Gilens

On welfare: "very consistent" with Gilens

Outside welfare: larger minority shares go with lower benefits; the attitude measure leaves the mechanism open

The student: "I'm not yet sure how race matters"

Howard, pp. 188–89

How much of the welfare state's shape does race explain?

"Race matters" is where the argument starts.

Next: "Race explains X and not Y, and here is how you would tell."

Week 12 prompt

Next week

This week: nothing due

Wed 11/11: Veterans Day, no class

Mon 11/16: Michener 2020; Mettler 1998, ch. 1, 5–6

Fri 11/20, 11:59 p.m.: Discussion Paper 5

By 2:15 you should be able to say

The poor: 27% actual, 50% guessed (Gilens, 1999)

Opposition turns on who seems deserving

Among whites, a work-ethic stereotype predicts it best

News pictures jumped in 1965 and followed tone

Howard: the pattern reaches past welfare; attitudes don't say why

Race explains X and not Y: how would you tell?