Factors Relating to Misanthropy in Contemporary American Society

Misanthropy is higher among those with lower socioeconomic status, subgroups toward the social periphery such as racial and ethnic minorities and Southerners, people suffering recent negative life events, those who do not attend church and fundamentalists, and members of more recent cohorts. For blacks, misanthropy is decreased by more social contact with whites, but the misanthropy of whites is not affected by contact with blacks. There has been some recent increase in misanthropy, but claims that misanthropy has shown a large and long-term rise are not well supported by the time series.

In both academic and popular discourse over the past year and a half, considerable attention has been devoted to the issue of social capital in contemporary America (Elshtain, 1996; Putnam, 1995a; Putnam, 1995b; Putnam, 1995c; Putnam, 1996; Fukuyama, 1995; Brehm and Rahn, 1995; Uslaner, 1995; Wright, 1995; Samuelson, 1996; Hong, 1996; Ladd, 1996; Stengel, 1966). Robert Putnam and other proponents of the social capital school have argued that social capital is a set of both beliefs and behaviors relating to interpersonal relations (e.g., trust in other people) and social connectedness (e.g., membership in voluntary associations and interaction with neighbors) that are essential for the harmonious functioning of society. Without such attitudinal predispositions and social ties to facilitate social exchange, individuals become isolated and suspicious of other citizens, groups, and government itself and social, economic, and political life functions poorly. In effect, social capital is both a glue that bonds society together and a lubricant that permits the smooth running of society’s interactions (both interpersonal and among people, groups, and organizations). Moreover, the social capital school contends that social capital has been declining for the past 20–30 years and that it has been depleted to a dangerously low level.

In this paper we examine one key element of social capital, the degree of misanthropy in contemporary American society. We consider how the level of misanthropy has changed over the past 40 years and what are the factors that predict misanthropy.

Measuring Misanthropy

Misanthropy is measured by a 3-item scale adapted from Morris Rosenberg’s original 5-item misanthropy index (Rosenberg, 1956; Robinson, Shaver, and Wrightsman, 1991) (see below).

Generally speaking, would you say that most people can be trusted or that you can’t be too careful in dealing with people? (TRUST)

Do you think most people would try to take advantage of you if they got a chance, or would they try to be fair? (FAIR)

Would you say that most of the time people try to be helpful, or that they are mostly just looking out for themselves? (HELPFUL).

As conceptualized by Rosenberg (1956), the scale taps “faith in people,” “attitudes towards human nature,” or an “individual’s view of humanity.” The misanthropic believe that people in general are untrustworthy, unfair, and unhelpful. As Luhmann (1979) characterizes them they have a lack of faith in human nature, a negative and pessimistic outlook on people, and feel besieged and isolated, suffering from “negative atomism.” Similarly, Lewis and Weigert (1985a; 1985b) describe them as lacking both interpersonal and system trust and suffering from “social atomism.”[1]

The Rosenberg scale has been a core battery on the General Social Surveys (GSSs) of the National Opinion Research Center since 1972. The GSSs are full-probability samples of the adult household population of the United States using in-person interviews (Davis and Smith, 1994). The 3 misanthropy items on the GSS were recoded so that 1 expresses the judgement that people are trustworthy, fair, or helpful, 2 equals “depends,” and 3 means that you “can’t be too careful in dealing with people,” that people “try to take advantage of you,” or that people “are mostly just looking out for themselves.” Scores run from 3 for someone who considers people to be trustworthy, fair, and helpful to 9 for those who consider people untrustworthy, unfair, and unhelpful. Across all years 26.5% saw people in the most favorable light, 22.3% in the most negative light and the mean score was 5.85.

The 3 items form a good, short scale with a Cronabach’s alpha of .67 and an average interitem correlation of .40. The misanthropy items are distinct from, but correlated with, items taping related concepts. They differentiate themselves from other items on the GSS measuring anomia (ANOMIA5, ANOMIA6, ANOMIA7), views about the nature of the world (WORLD1, WORLD4), and confidence in leaders of institutions (CONFINAN, CONBUS, CONCLERG, CONEDUC, CONFED, CONLABOR, CONPRESS, CONMEDIC, CONTV, CONJUDGE, CONSCI, CONLEGIS, CONARMY). The misanthropy questions form their own 3-item factor in all comparisons with each alternative set of items either singly or altogether.[2]

However, the misanthropy scale is related to these other scales as one would expect (Table 1). The misanthropic are more anomic, have negative views of the nature of the world, and have less confidence in the leaders of most institutions. The relationships with anomia and world views are moderate in magnitude and consistent in direction. The associations with low confidence in institutional leaders are more modest and there are two exceptions. First, the relationships are modest because attitudes towards people in general are both conceptually and usually empirically distinct from attitudes toward institutions and elites. Second, the exceptions are that misanthropy is associated with having slightly more confidence in the people running labor unions and television. Since previous research has identified these as two nonestablishment institutions that often had low or even negative associations with confidence in the other 11 institutions (Smith, 1981; Lipset and Schneider, 1983), this pattern is not unexpected.

Trends In Misanthropy

The social capital school argues that social capital in general and interpersonal trust in particular has been declining. As Putnam (1995a) has written, “Americans are also less trusting. The proportion of Americans saying that most people can be trusted fell by more than a third between 1960, when 59 percent chose that alternative, and 1993, when only 37 percent did.”

But trends in misanthropy in general and trust in particular are difficult to reliably ascertain.[3] First, responses are very sensitive to both wording and context. Regarding question wording, a 1983 GSS experiment found that 57% thought that “most people can be trusted” (Table 2A), but only 36.5% believed that “most people can be trusted” vs “you can’t be too careful in dealing with people” (Table 3). Also, a nonexperimental comparison of the variant wording used in 1960 and 1978 (Table 2B) suggests that it gathers 4–8 percentage points more trust than the standard GSS wording (Tables 3 and 4). In addition, No Opinion levels on the three measures on the self-administered Monitoring the Future (MTF) surveys of high school seniors are much higher than for the same items on the interviewer administered GSS (23–35% in the former vs 3–7% in the latter, see Tables 4 and 5). While these differences could result from the different populations being sampled, it is likely that the much higher nonresponse on MTF results from the fact that “Don’t Know, undecided” is a precoded, middle option on MTF, but that “Depends” and “Don’t Know” are unread responses on the GSS.[4]

---

Table 1. Attitudinal Associates of Misanthropy (Pearson’s r)

A. Anomia
Lot of average man (ANOMIA5)	2.263a
Unfair to bring child into the world (ANOMIA6)	2.286
Public officials not interested in average man (ANOMIA7)	2.260
B. World view
World evil (WORLD1)	2.261
Human nature perverse (WORLD4)	.273
C. Confidence in institutions
Banks and financial institutions (CONFINAN)	.105
Major companies (CONBUS)	.164
Organized religion (CONCLERG)	.098
Education (CONEDUC)	.067
Executive branch of the federal government (CONFED)	.142
Organized labor (CONLABOR)	2.042
The press (CONPRESS)	.041
Medicine (CONMEDIC)	.091
TV (CONTV)	2.024
U.S. Supreme Court (CONJUDGE)	.147
Scientific community (CONSCI)	.176
Congress (CONLEGIS)	.081
Military (CONARMY)	.014

Source. GSS, 1972–1994.

Anomia. Now I’m going to read you several more statements. Some people agree with a statement, others disagree. As I read each one, tell me whether you more or less agree with it, or more or less disagree.

A. In spite of what some people say, the lot (situation/condition) of the average man is getting worse, not better. (ANOMIA5)

B. It’s hardly fair to bring a child into the world with the way things look for the future. (ANOMIA6)

C. Most public officials (people in public office) are not really interested in the problems of the average man. (ANOMIA7)

World View. People have different images of the world and human nature. We’d like to know the kinds of images you have. Here is a card with sets of contrasting images. On a scale of 1–7 where would you place your image of the world and human nature between the two contrasting images.

Look at the first set of contrasting images. If you think that “The world is basically filled with evil and sin,” you would place yourself at 1. If you think “There is much goodness in the world which hints as God’s goodness” you would place yourself at 7. If you think things are somewhere between these two you would place yourself at 2, 3, 4, 5, or 6.

A. The world is basically filled with evil and sin ... There is much goodness in the world which hints at God’s goodness (WORLD1)

B. Human nature is basically good ... Human nature is fundamentally perverse and corrupt. (WORLD4)

Confidence. I am going to name some institutions in this country. As far as the people running these institutions are concerned, would you say you have a great deal of confidence, only some confidence, or hardly any confidence at all in them?

^a All correlations are statistically significant at least at the .01 level except with confidence in the military which is not significant at the .05 level.

Table 2. Trends in Trust of People

A. Do you think most people can be trusted?

Date	% Yes	% No	% DK	N
3/1948	65	30	4	1289
8/1952	68	30	2	1297
11/1953	57	39	4	1232
1/1954	62	34	4	1250
11/1954	65	32	3	1201
4/1957	75	22	3	1279
10/1964	77	21	2	1975
3/1983	56	41	3	790
3/1983	57	40	3	790

B. Some people say that most people can be trusted.

Others say you can’t be too careful in your dealing(s) with people. How do you feel about it? ^b

Year	Trust	Not trust	DK/other	N
1960	55.3	40.4	4.2	970
1978	48	51	1	1635

Source. National Opinion Research Center.

^a The first GSS figure does not weight for number of eligible respondents in the household, while the second figure does. The latter is more accurate, but the former is probably more comparable to the other figures in the series.

^b Plural in 1960; singular in 1978.

---

Regarding question order, both the trust and helpful items have experienced large context effects of, respectively, 7.7 and 9.4 percentage points (Table 3) (Smith, 1983; 1990; 1991a). These items are probably especially prone to context effects because they call for global assessments of people in general based presumably on one’s entire life experience. Making judgments based on such massive, cognitive retrievals are difficult and open to variability. Sampling of ones own memories on such broad topics tend to be biased rather than complete or random. Questionnaire context is one factor that biases the cognitive processing and in turn influences the summary judgments (Smith, 1991a; 1991b).

The Trust1 series (See Table 5, Adjusted) is preceded by an item on whether one’s life is exciting and a ranking of five work values in all years and then by an item on getting ahead in life and two sexual morality questions in 1976–1994. The Trust2 is preceded by three or four crime attitude items and then in 1978 to 1994 by items on political ideology, equalizing wealth, divorce laws, and legalizing marijuana. 1972 does not match either of these contexts. Helpful1 is preceded by three items on personal finances and a measure of subjective social class. Three anomia items come next in 1973 and 1976 and two batteries on the use of violence in 1980–1994. In 1976 the hit items come immediately before the anomia items. Helpful2 is preceded by media use items: viewing TV and reading newspaper in 1975 and 1986 to 1994 and radio listening, viewing TV, and reading the newspaper in 1978 and 1983. From 1978 to 1994 these are preceded by items of smoking, drinking, and socializing. 1972 does match either of these contexts.

Table 3. Non-NORC Trends in Misanthropy

Year	Trust	Not trust	DK/other	N
1964	53.4	44.7	1.9	1446
1966	52.9	45.6	1.5	1284
1968	55.2	43.2	1.6	1343
1971	48.5	50.0	1.5	2164
1972	45.8	52.4	1.8	2179
1974	46.6	52.1	1.3	2486
1976	51.3	45.9	2.8	2400
1978	47.9	51.9	0.2	3630
1979	43	56	2	1635
1981	43	54	3	2325
1981	47	50	3	1729
1983	40	55	5	1207
1990	49	49	3	1839
1991	34	63	3	600
1992	44.7	54.5	0.8	2244
1995	35	63	2	1514

Year	Helpful	Not helpful	DK/other	N
1964	54.3	41.3	4.4	1445
1966	51.9	45.7	2.4	1285
1968	58.2	45.7	3.2	1344
1971	54.8	41.7	3.5	2164
1972	46.9	50.7	4.2	2174
1974	50.7	46.5	2.8	2450
1976	51.9	43.8	4.3	2394
1978	57.9	41.5	0.6	3605
1979	41	56	2	1635
1983	49	45	6	1207
1991	43	48	9	600
1992	58.7	39.2	2.1	2229
1994	37	58	6	600
1995	49	48	3	1514

Year	Fair	Not fair	DK/other	N
1964	67.3	28.6	4.2	1443
1968	66.8	30.1	3.1	1342
1971	65.9	31.5	2.6	2164
1972	58.9	36.8	4.3	2179
1974	57.6	39.5	2.8	2473
1976	59.9	35.5	4.6	2390
1977	58	35	7	1447
1978	67.0	32.0	1.0	3604
1979	56	41	3	1207
1983	56	36	8	1207
1995	50	48	2	1514

Source. NES/SRC, 1964, 1966, 1972, 1974, 1976, 1992; QOL/SRC, 1971, 1978; Gallup, 1981, 1981, 1990, 1994; Audits and Surveys, 1983; Princeton Survey Research Associates, 1991, 1995.

Note. Wording: Trust: Generally speaking, would you say that most people can be trusted or that you can’t be too careful in dealing with people? Helpful: Would you say that most of the time people try to be helpful, or that they are mostly just looking out for themselves? “Generally speaking” precedes question in 1979. “Just” omitted in 1991. Fair: Do you think most people would try to take advantage of you if they got a/the chance, or would they try to be fair?

Table 4. GSS Trends in Misanthropy

A. Unadjusted

Year	Trust	Not trust	DK/other	N
1972	46.3	50.0	3.8	1598
1973	46.8	50.4	2.8	1499
1975	39.7	56.1	4.2	1479
1976	44.0	52.6	3.4	1494
1978	39.9	55.7	4.4	1528
1980	44.3	51.9	3.7	1463
1983	36.5	59.2	4.3	801
1984	48.9	48.7	2.4	1462
1986	37.7	59.5	2.7	1466
1987	43.6	52.8	3.6	1460
1988	39.5	56.1	4.4	990
1989	41.1	55.5	3.3	1018
1990	38.4	57.4	4.2	892
1991	38.9	56.0	5.1	1019
1993	35.3	60.6	4.1	1061
1994	34.4	61.3	4.3	1976

Year	Helpful	Not helpful	DK/other	N
1972	45.9	47.9	6.2	1586
1973	46.7	49.6	3.7	1496
1975	56.7	36.9	6.3	1477
1976	42.0	51.9	6.1	1493
1978	59.4	35.5	5.1	1523
1980	48.8	46.8	4.5	1459
1983	57.4	38.5	4.1	1586
1984	52.2	44.1	3.7	1466
1986	56.3	38.4	5.2	1458
1987	47.6	47.8	4.6	1456
1988	49.6	46.1	4.3	987
1989	50.9	44.2	4.9	1014
1990	51.6	43.0	5.5	884
1991	49.3	44.5	6.2	1012
1993	52.2	41.9	5.6	1052
1994	46.9	46.6	6.4	1977

Year	Fair	Not fair	DK/other	N
1972	59.0	35.2	5.8	1590
1973	57.5	37.6	4.9	1496
1975	61.7	31.3	7.0	1474
1976	59.4	36.6	4.0	1490
1978	64.2	30.4	5.3	1516
1980	61.5	34.2	4.3	1454
1983	58.9	35.8	5.2	1584
1984	63.0	34.1	2.9	1468

TABLE 4—Continued

A. Unadjusted—Continued

Year	Fair	Not fair	DK/other	N
1986	63.0	33.3	3.7	1456
1987	58.7	37.6	3.7	1446
1988	60.6	33.7	5.7	985
1989	59.2	35.3	5.4	1016
1990	58.0	36.0	6.0	887
1991	57.1	37.5	5.4	1009
1993	55.8	37.3	7.0	1055
1994	53.5	39.6	6.8	1969

B. Adjusted

Trust 1	Trust 2	Helpful 1	Helpful 2
1972		46.3	45.9
1973	46.8		46.7
1975		39.7		56.7
1976	44.0		42.0
1978		39.9		59.4
1980	44.3		48.8
1983		36.5		57.4
1984	48.9		52.2
1986		37.7		56.3
1987	43.6		47.6
1988	44.2	34.7	44.5	54.7
1989	47.8	34.2	48.8	53.1
1990	42.0	35.1	47.6	55.2
1991	43.5	34.6	46.0	52.4
1993	38.4	32.3	48.7	56.3
1994	36.6	32.0	39.8	54.5

Note. Trust1 is preceded by an item on whether ones life is exciting and a ranking of five work values in all years and then by an item on getting ahead in life and two sexual mortality questions in 1976–1994.

Table 5. Trends in Misanthropy among High School Seniors

A. Trust

1975	34.5	39.8	25.7	3013
1976	31.5	37.8	30.7	2953
1977	32.7	38.9	28.4	3117
1978	31.3	40.6	28.2	3683
1979	31.2	42.6	26.3	3285
1980	31.0	41.6	27.3	3219
1981	32.0	40.2	27.8	3534
1982	28.3	44.1	27.6	3584
1983	27.7	43.7	28.6	3344
1984	27.5	46.6	25.9	3223
1985	28.6	45.5	25.9	3222
1986	25.1	46.8	28.1	3088
1987	24.4	49.6	26.0	3309
1988	23.3	50.6	26.1	3316
1989	20.3	55.5	24.2	2785
1990	19.7	54.2	26.1	2583
1991	20.2	55.0	24.9	2544
1992	18.3	58.9	22.8	2657

B. Helpful

Year	Helpful	Looks out for self	Don’t know	N
1975	32.5	37.5	30.0	3008
1976	31.7	37.3	31.0	3010
1977	33.9	34.7	31.4	3177
1978	33.0	36.0	31.0	3754
1979	31.6	39.1	29.3	3345
1980	32.9	35.9	31.2	3285
1981	33.2	36.6	30.1	3591
1982	32.4	38.0	29.7	3651
1983	34.0	34.9	31.1	3414
1984	34.0	35.0	31.1	3274
1985	34.4	36.1	29.6	3275
1986	29.8	37.4	32.8	3155
1987	30.0	38.5	31.5	3340
1988	28.2	40.2	31.6	3363
1989	27.0	40.6	32.4	2858
1990	25.6	40.4	34.0	2616
1991	26.7	41.9	31.4	2563
1992	24.5	42.2	33.3	2674

C. Fair

Year	Fair	Take advantage	Don’t know	N
1975	28.2	39.1	32.7	3011
1976	27.9	38.1	34.0	3004

TABLE 5—Continued

C. Fair—Continued

Year	Fair	Take advantage	Don’t know	N
1977	30.4	35.2	34.4	3170
1978	30.1	36.7	33.1	3755
1979	28.8	39.0	32.2	3343
1980	27.8	38.7	33.5	3276
1981	29.2	36.5	34.3	3588
1982	26.7	39.7	33.6	3651
1983	26.6	37.9	35.5	3409
1984	26.3	40.9	32.8	3278
1985	27.6	39.6	32.8	3277
1986	27.0	40.5	32.6	3150
1987	25.0	41.3	33.7	3338
1988	22.1	44.4	33.5	3360
1989	20.5	45.8	33.6	2842
1990	19.8	45.8	34.4	2612
1991	20.3	46.1	33.6	2557
1992	17.5	52.0	30.5	2672

Source. Monitoring the Future.

Note. Wording: Trust: Generally speaking, would you say that most people can be trusted or that you can’t be too careful in dealing with people? Helpful: Would you say that most of the time people try to be helpful or that they are mostly just looking out for themselves? Fair: Do you think most people would try to take advantage of you if they got a chance or would they try to be fair? “Don’t know, undecided” is a precoded, middle option for all three items.

Trust2 series is preceded by three or four crime attitude items and then in 1978 to 1994 by items on political ideology, equalizing wealth, divorce laws, and legalizing marijuana. The focusing on crime and victimization may account for the lower trust expressed in the latter context.

The Helpful1 series is preceded by three items on personal finances and a measure of subjective social class. Three anomia items come next in 1973 and 1976 and two batteries on the use of violence in 1980–1994. In 1976 the hit items come immediately before the anomia items. The Helpful2 series is preceded by media use items: viewing TV and reading newspaper in 1975 and 1986 to 1994 and radio listening, viewing TV, and reading the newspaper in 1978 and 1983. From 1978 to 1994 these are preceded by items of smoking, drinking, and socializing. The anomia and violence items in the former series may be causing the lower ratings of helpfulness in that context.

Because of these context effects on the GSS the times series are reported separately for each context (Table 5, Adjusted).

In sum, the sensitivity of the misanthropy items to variations in wordings and context means that comparisons across different wordings (as Putnam did)[5] are unwise and that even intersurvey comparisons of identical wordings are suspect because of possible context effects.

In part undoubtedly due to these measurement factors, trends in misanthropy are complex and somewhat contradictory. Looking at the individual components first, we see considerable variation, but no clear trend, in trust for the oldest series from 1948 to 1983 (Table 6). The series from the mid-1960s on (NORC 1960–1978; non-NORC 1964–1995; and GSS 1973–1994 and 1975–1994) do point to a decline in trust of about 0.3 to 0.4 percentage points per annum, but when the non-NORC series are broken down into subseries by program (NES, QOL, Gallup, and PSRA) the pattern is much less consistent. Next, for being helpful there are either constant fits (GSS 1973–1994) or nondirectional change (GSS 1975–1994 and non-NORC 1964–1995) for the main, adult times series with only the student MTF series showing a decline. Finally, for being fair the series agree that a decline is occurring, but differ a great deal on its magnitude from 20.16 percentage points per annum on the 1972–1994 GSS to 20.66 percentage points on the 1975–1992 MTF. Thus, with the exception of the reasonably consist MTF student series (Easterlin and Crimmins, 1991), the trends for the individual misanthropy items are complex and only partly consistent.

Looking at the composite scale on the GSS shows a modest increase in misanthropy.[6] Itis significantly related to year, but the correlation is only .033 and the mean rose from a low of 5.6 in 1978 to a high of 6.1 in 1994 (Table 7). Furthermore, close inspection of the data indicates the even this modest trend became established only since 1991.[7] Thus, at most one can talk about only a modest and fairly recent increase in misanthropy.[8]

Predictors of Misanthropy

Based on a review of the literature on misanthropy and related concepts, we formulated a number of hypotheses about the factors that influence misanthropy.

First, we expect that misanthropy will decrease with socioeconomic status (Bahr and Martin, 1983; Brehm and Rahn, 1995; DeMaris and Yang, 1994; Grabb, 1980; Huang and Anderson, 1991; Robinson, Shaver, and Wrightsman, 1991; Smith, 1985). The better-off people are materially and the higher their social standing, the more likely they will view the world and other people in a favorable light. Specifically, misanthropy should decline as household income and respondent’s education increase.

Table 6 Models of Trends^a

Time	Model	Per annum	r2
A. Trust
1948–83	NCNL	(1.0007)	.03
1960–78	SLT	2.0041	1.0
1964–95	SLC	2.0041	.59
1964–92 (NES)	SLC	2.0031	.54
1971–78 (QOL)	C	(2.0009)	1.0
1981–90 (Gal.)	SLC	1.0048	.57
1991–95 (PSRA)	C	(1.0025)	1.0
1972–94 (GSS)	SLC	2.0039	.40
1973–94 (GSS)	SLC	2.0028	.33
1975–94 (GSS)	SLC	2.0041	.89
1975–92 (MTF)	SLC	2.0092	.93
B. Helpful
1964–95	NCNL	(2.0013)	.19
1964–92 (NES)	SLC	1.0019	.11
1971–78 (QOL)	SLT	1.0044	1.0
1991–95 (PSRA)	SLT	1.0015	1.0
1973–94 (GSS)	NCNL	(2.0002)	.00
1975–94 (GSS)	C	(2.0023)	.45
1975–92 (MTF)	SLC	2.0046	.60
C. Fair
1964–95	SLC	2.0052	.66
1964–76 (NES)	SLC	2.0082	.78
1971–78 (QOL)	C	(1.0016)	1.0
1972–94 (GSS)	SLC	2.0019	.21
1975–92 (MTF)	SLC	2.0066	.79

Sources. Data from Tables 2–6.

^a To categorize each trend a series of models are fitted to the data points. First, the constant model is tried that assumes that all data points are random variations around a stable level. An estimate of the pooled or average level is made and a test is made if the data points vary significantly from it. If the model is rejected, the linear model that all data points are random variations around a linear fit is tried. Four outcomes are possible: (1) constant, (2) significant linear component, (3) significant linear trend, and (4) nonconstant, nonlinear. The constant model is accepted when there is no significant variation around the constant or pooled proportion. The significant linear component model is accepted when (a) the constant model is rejected, (b) the linear model is rejected, but (c) the linear fit is significantly better than the constant fit. The significant linear trend is accepted when (a) the constant model is rejected and (b) there is no significant variation around the linear model. The nonconstant, nonlinear model is accepted when (a) the constant model is rejected, (b) the linear model is rejected, and (c) the improvement between the constant and linear models is not significant. r² measures the fit between the trend and time. Parentheses denote changes that are not statistically significant.

Table 7. Trends in GSS Misanthropy Scale

Year	Mean	N
1972	5.81	1559
1973	5.87	1485
1975	5.65	1455
1976	5.95	1484
1978	5.57	1510
1980	5.78	1442
1983	5.88	789
1984	5.63	1454
1986	5.74	1444
1987	5.87	1433
1988	5.87	976
1989	5.83	1007
1990	5.88	879
1991	5.91	1001
1993	5.96	1043
1994	6.12	1953

Second, misanthropy should decrease with upward social mobility. Improvements in social standing, both inter- and intra-generationally, should reduce negative evaluations.

Third, misanthropy should increase as negative life events or traumas occur (Brehm and Rahn, 1995; House and Wolf, 1978; Norris and Kaniasty, 1991; Smith, 1976). Negative experiences, especially those caused by other people, should lead to unfavorable evaluations of people. Specifically, pessimism should increase with experiences of (a) criminal victimization (having been robbed or burglarized in the last year), (b) interpersonal violence (having been hit or shot at), (c) illness and hospitalization, (d) unemployment, and (e) deaths in the family.

Fourth, misanthropy should increase with disruptive family situations especially those involving divorce (Brehm and Rahn, 1995; Southworth and Schwarz, 1987; Yoder and Nichols, 1980). It should be higher among those who (a) were children of divorced parents, (b) have been divorced themselves, (c) are currently divorced or separated, and (d) have never been married. While it is possible to see divorce as just another miscellaneous negative life event, we hypothesis that divorce will have a distinct and especially strong impact on shaping judgments about human nature since it concerns broken commitments involving very close, interpersonal relationships.

Fifth, misanthropy should be greater in large metropolitan areas where more people are strangers and the environment is perceived as more threatening and impersonal (House and Wolf, 1978; Robinson, Shaver, and Wrightsman, 1991; Wilson, 1985).

Sixth, misanthropy should be higher among cultural groups and minorities that have been discriminated against and isolated from the majority culture (Calhoun and Cann, 1994; DeMaris and Yang, 1994; Robinson, Shaver, and Wrightsman, 1991; Terrell and Barrett, 1979; Thomas and Hughes, 1986; Wilson, 1985). Specifically, negative assessments should be higher among blacks and Hispanics and perhaps among Jews and recent immigrants. In addition, social contact between the minorities and the majority should reduce misanthropy coming from exclusion (Smith, 1996).[9] However, for the majority social contact with minorities should have minimal impact. The social contact hypothesis can be tested in reference to blacks and whites, but not other racial groups.

Seventh, religious belief should reduce misanthropy (Bahr and Martin, 1983; Luhmann, 1979; Schoenfeld, 1978). In addition, those who attend church more should be less misanthropic (Bahr and Martin, 1983; Schoenfeld, 1979). However, among the religious those with fundamentalist beliefs, which emphasize the sinful nature of humans and a stern and authoritarian God, should be more misanthropic than those with a liberal religious orientation, which emphasizes human goodness and a compassionate and caring God (Bahr and Martin, 1983; Robinson, Shaver, and Wrightsman, 1991; Schoenfeld, 1979).

Eight, misanthropy should be lower among older adults (Brehm and Rahn, 1995; DeMaris and Yang, 1994; Robinson, Shaver, and Wrightsman, 1991).[10] This may result from life course adjustments, as adults adapt to their surroundings and find friends, employers, etc., whom they can trust. For example, job satisfaction improves with age as people eventually find jobs that met their particular talents, temperament, and expectations. Alternatively, lower misanthropy among older adults might be a function of cohort. Older adults were raised during a period in which society and ones fellow citizens were more civil (e.g., less crime, less divorce, more people living outside of large metropolitan areas). To the extent that their assessments of human nature were forged by a Mannheimian generational socializing process, they should be less misanthropic than more recent generations socialized in more troubled situations and more impersonal environments.

Ninth, misanthropy should be higher among men than women (DeMaris and Yang, 1994; Lagace and Gassenheimer, 1989; Robinson, Shaver, and Wrightsman, 1991; Terrell and Barrett, 1991).[11] Traditionally men see the world in more competitive and conflictual terms, while women are more cooperative and nurturing. Of course, these traditional gender roles and perspectives have been changing in recent decades and these distinctions have attenuated.

Tenth, misanthropy should be greater in the South (Ellison, 1991; Robinson, Shaver, and Wrightsman, 1991). Higher levels of violence, greater intergroup hostility, lower standards of living, isolation from national centers of power, and its status as a defeated, subordinate region should tend to make Southerners more suspicious of people than non-Southerners are.

Eleventh, misanthropy might be greater among recent movers. Movers are outsiders in their new communities and may have few affirming ties to friends and neighbors. On the other hand, movers maybe more adaptive and cosmopolitan than stayers and being more at ease with others they maybe less misanthropic.

Finally, we expect a number of other variables to be associated with misanthropy, but we are much less sure of the casual ordering. Misanthropy should be lower among those who belong to many voluntary associations (Brehm and Rahn, 1995), people who do not drink to excess, those who are happy, and, among the married, those in happy marriages.

Bivariate Tests of Hypotheses

Table 8 shows the bivariate tests of these expectations. Overall, there are significant relationships in the predicted direction in all but three cases.

First, higher income and education are both associated with less misanthropic evaluations.

Second, both intergenerational downward occupational mobility and a recent worsening in financial situation are related to negative assessments.

Third, with one exception recent negative life events (including criminal victimization and violence, health problems, unemployment, and traumas in general) increase misanthropy. However, counter to this pattern and to expectations, deaths in the family are associated with less misanthropy.[12]

Fourth, having come from a broken home, having been divorced oneself, or not being married are related to negative judgments.

Fifth, those living in a large central city are more misanthropic.

Sixth, in general minorities have more pessimistic outlooks than do majority groups. Elaborating on the groups used in Table 8, ethnoracial differences were examined for more than 40 groups. This showed that misanthropy was lowest among early European immigrant groups (British) and Scandinavians and higher for more recent European immigrant groups and non-European groups (Africans, Asians, Amerindians, and Hispanics). To simplify the pattern and come up with enough observations for reasonably stable estimates seven groups were distinguished in the final analysis (1, early immigrants from Scandinavia and Great Britain; 2, middle immigrants from France, French Canada, Belgium, Ireland, Germany, Austria, Holland; 3, late immigrants from Eastern and Southern

TABLE 8. Bivariate Correlates of Misanthropy

Most misanthropic			probability	Fita	category
A. Socioeconomic Status
Household income (REALINC)	2.191/.000	LC	Less than $10,000	(19823)
Education (EDUC)	2.233/.000	LC	LT high schoolb	(20871)
B. Social mobility
Intergenerationalc	2.060/.000	L	Downwardly mobile	(19906)
Recent financial changes	.114/.000	L	Got worse	(20766)
(FINALTER)
C. Life events
Robbed/burglarizedd	.100/.000	L	2 Victimizations	(9927)
Hit/threatenede	.137/.000	LC	2 Occurrences	(15737)
Hospitalized/disabled	.058/.000	L	2 Occurrences	(13307)
(HOSDIS5) f
Health (HEALTH)	.144/.000	LC	Poor health	(13720)
No job (UNEMP5) g	.144/.000	L	2 Occurrences	(13383)
Family deaths (DEATH5)h	2.056/.000	LC	None	(13370)
Traumas in last year	.088/.000	L	4 Occurrences	(12892)
(TRAUMA1)
Traumas last 5 years	.082/.000	LC	4 Occurrences	(12892)
(TRAUMA5)
D. Family status/divorce
Marital statusi	.110/.000	S	Separated	(20912)
Parents divorced j	2.102/.000	L	Parents divorced	(20903)
Ever divorcedk	.079/.000	L	Been divorced	(20913)
Recent divorce (DIVORCE5)l	.059/.000	LC	2 Divorces	(13208)
E. Community typem	2.090/.000	LC	Big cen. cities	(20913)
F. Minorities
Race (RACE)n	2.225/.000	L	Black	(20913)
Jewish (RELIG)o	.025/.000	L	Non-Jewish	(20862)
Contacts (blacks only)
Same neighborhood	.067/.014	L	Segregated	(1633)
(RACLIVE)
Same church (RACCHURH)	.081/.003	L	Segregated	(1483)
Dinner guest (RACHOME)	.082/.029	L	No nonblack guests	(869)
Segregatedp	.128/.000	L	No nonblack contact	(749)
Immigrant statusq	.028/.000	LC	Immigrants	(13974)
G. Religion
Has religion (RELIG)r	.032/.000	L	None	(20862)
Attends church (ATTEND)	2.101/.000	LC	Never	(20756)
Theology (FUND)	2.139/.000	LC	Fundamentalist	(20315)
H. Age (AGE)	2.130/.000	LC	Younger adults	(20855)
I. Gender (SEX)	2.031/.000	L	Men	(20913)
J. Geographic location
Region (REGION)	.125/.000	S	South	(20913)
Mobility (MOBILE16)	2.069/.000	LC	Same city	(20624)
K. Work situation
Supervisions	2.093/.000	L	Bottom	(6979)

TABLE 8—Continued

r/ probability	Fita	Most misanthropic category
L. Other
Group membership	2.182/.000	LC	No memberships	(14889)
(MEMNUM)
Drinkingt	2.026/.000	LC	Drinks to excess	(13362)
General happiness (HAPPY)	.182/.000	LC	Not too happy	(20765)
Marital happiness (HAPMAR)u	.095/.000	LC	Not too happy	(12618)

^a One way analysis of variance. L, linear—a statistically significant difference between groups and no statistically significant deviation from linearity. Note that all dichotomies that statistically differ are necessarily linear. LC, linear component—a statistically significant linear component and a statistically significant deviation from linearity. S, a statistically significant difference between groups, but a nominal variable for which the linearity test is not appropriate.

^b EDUC was used for the Pearson’s r and DEGREE for the breakdown.

^c Respondent’s occupational prestige minus father’s occupational prestige.

^d Robbed (ROBBRY) and/or burglarized (BURGLR) in the last year.

^e Hit (HIT) or shot at/threatened with a gun (GUN) as an adult.

^f Times hospitalized/disabled during last five years.

^g Times unemployed during last five years.

^h Recoded into 0, 1, 21 deaths of relatives during last five years.

ⁱ For the Pearson’s r marital status (MARITAL) was recoded married vs never married.

^j Coded as family intact 5 1, not intact, but not divorced 5 2; and not intact because of divorce 5 3 (FAMILY16 and FAMDIF16).

^k Either currently or ever divorced (MARITAL and DIVORCE).

^l Recoded no divorce in last five years, one divorce, two divorces.

^m Combination of SRCBELT and XNORCSIZ 1 5 exurbia, 2 5 small town, suburbs, etc., 3 5 central cities, 4 5 top 12 central cities.

ⁿ Recoded black vs not black.

^o Recoded Jewish vs not Jewish.

^p Scale of RACLIVE, RACCHURH, and RACHOME ranging from 0 contacts to 3 contacts.

^q Scale of immigrant generation (BORN, PARBORN, and GRANBORN) ranging from 0 (first generation—born outside country) to 3 (all four grandparents born in country).

^r Recoded has religion vs no religion.

^s Supervision over others—supervision by others. 2-supervises those who supervise others and no one supervises respondent; 22-supervises no one, is supervised by someone who is supervised by someone else (WKSUP WKSUPS WKSUB WKSUBS).

^t Doesn’t drink-1; drinks, but not to excess-2; sometimes drinks to excess-3 (DRINK and DRUNK).

^u Words in capitals are GSS mnemonics (Davis and Smith, 1994).

Europe, Jewish, and miscellaneous; 4, Hispanics from Spanish-speaking countries; 5, Amerindians; 6, Asians; and 7, blacks (Table 9).

This breakdown shows that the difference between blacks and whites is especially pronounced. Blacks are much more misanthropic than whites: 51.2% of whites and 80.9% of blacks consider people untrustworthy (125.4 percentage points), 31.5% of whites and 60.6% of blacks judge people as unfair (125.0), and 40.9% of whites and 62.7% of blacks think people are unhelpful (121.8%). These single-item differences convert into a mean difference of 11.58 on the misanthropy scale (5.65 for whites vs 7.23 for blacks).

Table 9. Ethnoracial Differences in Misanthropy

A. Ethnoracial groupsa
b	Probability
Blacks	.308	.000
Late immigrants	.142	.000
Hispanics	.125	.000
Amerindians	.086	.000
Middle immigrants	.069	.000
Asians	.035	.000
r2	.078	(21243)
B. Race (black/nonblack)	b
b	Probability
Race	.237	.000
r2	.056	(21243)

^a Early Immigrants, from Scandinavia and Great Britain (omitted category); Middle Immigrants, France, French Canada, Belgium, Ireland, Germany, Austria, Holland; Late Immigrants, from Eastern and Southern Europe, Jewish, Misc.; Hispanics, from Spanish-speaking country on ETHNIC; Amerindians, American Indian on ETHNIC; Asians, from Asian country on ETHNIC; Blacks, black on RACE.

^b Race: nonblack, 0 vs black, 1 on RACE.

The racial difference is probably a joint reflection of the position of blacks as a social and numerical minority. Two of the misanthropy items are framed in terms of whether the majority of people are trustworthy or fair. In America of course the majority of people are white. So from a strictly demographic point of view the questions essentially ask blacks whether most whites are trustworthy or fair. For the trust dimension we know how blacks explicitly evaluate the trustworthiness of whites. In the 1982 GSS blacks were asked “Do you feel you can trust most white people, some white people, or no white people?” Only 7.3% thought most whites could be trusted, 81.4% that some whites could be trusted, and 11.3% that no whites could be trusted. This is even more negative than the judgment about people in general since in 1980–1983 17.5% of blacks thought most people could be trusted, 4.7% said it depends, and 77.8% thought most people were untrustworthy.[13] The connection between race and trust is also shown by the fact that blacks have more trust in local governments where a black is the mayor than they do in governments with white mayors (Bobo and Gilliam, 1990; Howell and Fagan, 1988). Thus, the lower ratings of people by blacks is consistent with and probably largely a reflection of how blacks view whites.[14]

Among blacks those who are the most isolated (i.e., having less contact with nonblacks) are the most negative. As expected, among nonblacks there is no association between interracial contacts and evaluations of human nature (Pearson’s r 5 2.001/NS).

Second, the relationship between immigration status and misanthropy is complex. Negative perspectives decline from immigrants (6.15) to the third generation (5.4) (native born of native born parents), but then rise again for those in the 41 generation (grandparents 1 native born) (6.0). As a result, the overall relationship is weakly positive (r 5 .036/.000) with higher immigrant generation associating with more misanthropy, opposite the hypothesis. This is probably due to the concentrations of blacks and Southerners in the latter category.

Finally, the lower misanthropic level among Jews is also counter to our hypothesis. Some earlier research (Schoenfeld, 1978) reported that Jews were more misanthropic than non-Jews and hypothesized that this was due to the negative impact of the Holocaust on the world view of Jews. However, the 1972–1994 GSSs find that Jews have the least negative view of people of adherents of major religions (Jews 5 5.45; Catholics 5 5.73; Protestants 5 5.88; No Religion 5 6.09; and Other Religion 5 6.11).[15]

Seventh, having no religion is weakly (but significantly) related to more misanthropy as is attending church less frequently. However, being a member ofa Fundamentalist denomination is associated with more negative views.

Eighth, misanthropy declines with age.

Ninth, men are slightly more misanthropic than women.

Tenth, misanthropy is higher in the South than in other regions.

Eleventh, geographic mobility is associated with lower misanthropy.[16]

Finally, nonmembership in voluntary associations, personal and marital unhappiness, and excessive drinking are related to pessimistic judgments.[17]

In brief, most of the hypothesized relations did appear, but the preliminary analysis indicates that trauma and minority status have more nuanced patterns than covered by the initial hypotheses and the competing geographic mobility hypotheses receive mixed and fairly weak support.

Multivariate Model

In developing a multivariate model we first dropped 8 of the 32 variables believed to be independent predictors in Table 10. First, the three individual measures of segregation were dropped in favor of the summary scale since all three items revealed the same pattern and the additive scale showed a stronger relationship. Second, the general measures of trauma during the last year and last five years (TRAUMA1 and TRAUMA5) were deleted because they were summary scales of the unemployment, hospitalization, divorce, and death items and therefore duplicated these items. The individual items were used rather than the summary scales since the individual items did not consistently relate to misanthropy (i.e., more deaths did not associate with more misanthropy). Next, the ever divorce item was used instead of the recently divorced item since they overlapped and ever divorced was the stronger variable of the two. Then, the supervision item was dropped since it only applied to those currently employed and in a separate regression analysis was not an independent predictor with basic demographics controlled for. Finally, being Jewish was left out since it had the weakest relationship with misanthropy, was opposite the hypothesized direction, and was a very skewed variable.

Second, we ran the regression models in stages. Initially we only utilized those variables included in all surveys (n’s of 19,0001 in Table 10). This included 16 of the 24 variables, but since we changed race and ethnicity into a series of dummy variables 21 variables appear in the equation. With listwise deletion of missing values there are 18,138 cases in this model. We then tested numerous further variables adding in various combinations of the remaining seven variables. The models produced virtually identical results and we eventually opted for one including all of these additional variables.

The full multivariate models (Table 10, Models 2 and 3) indicate that misanthropy is greater among those who (1) are less educated and have lower incomes,[18] (2) have had their finances worsen lately (but not related to intergenerational mobility), (3) have been criminally victimized, in poor health (but not related to recent hospitalizations, unemployments, or bereavements), (4) unrelated to a personal or parental history of divorce or to current marital status, (5) unrelated to urbanness, (6) are racial or ethnic minorities and recent immigrants (and especially among blacks), (7) rarely attend church and are fundamentalist (but unrelated to having a religion), (8) are younger/from more recent cohorts, (9) unrelated to gender, (10) live in the South, and (11) unrelated to geographic mobility.

Table 10. Multiple Regression Models of Misanthropy (Standardized Coefficient/Probability)

Variables (top category)	Model1	Model2	Model3
Household income (high)	2.052/.000	2.049/.001	2.046/.001
Education (more years)	2.197/.000	2.168/.000	2.170/.000
Intergenerational mobility (up)	2.013/.080	2.018/.190	2.015/.242
Recent financial changes (down)	.072/.000	.050/.000	.051/.000
Marital status (never married)	.013/.098	2.005/.738	2.009/.545
Parents divorced (divorced)	2.020/.004	2.018/.168	2.021/.096
Ever divorced (been divorced)	.038/.000	.001/.928	2.001/.940
Community type (big cities)	.032/.000	.016/.239	.021/.110
Race/ethnicitya

Black	.193/.000	.195/.000	.190/.000
Hispanic	.076/.000	.064/.000	.074/.000
Asian	.026/.000	.003/.820	.008/.501
Amerindian	.038/.000	.037/.008	.037/.006
Late Immigrants	.089/.000	.071/.000	.080/.000
Middle Immigrants	.044/.000	.023/.179	.025/.131
Has religion (no religion)	.025/.001	.005/.727	2.001/.950
Attends church (weekly1)	2.078/.000	2.079/.000	2.079/.000
Theology (liberal)	2.060/.000	2.039/.009	2.041/.005
Age (older)	2.163/.000	2.189/.000	2.189/.000
Gender (female)	2.040/.000	2.016/.232	2.014/.272
Region (South)	.068/.000	.078/.000	.076/.000
Mobility (changed states)	2.014/.050	2.018/.177	.014/.254
Robbed/burglarized (both)	—	.033/.011	.081/.000
Hit/threatened (both)	—	.080/.000	.041/.001
Hospitalized/disabled (both)	—	.003/.832	2.003/.796
Health status (poor)	—	.090/.000	.087/.000
Unemployment (21 times)	—	.011/.436	.011/.398
Family deaths (21)	—	2.022/.097	2.022/.097
Immigrants (41 generation)	—	2.029/.049	—
r2	.172	.190	.191
N	18,138	5,124	5,444

^a Early arriving ethnicities are the omitted category. See Table 9.

Finally, we conducted separate regressions for blacks and nonblacks. As before we ran the model first for the always occurring variables and then added in occasionally occurring variables. These models differ from the overall models by (a) dropping the race/ethnic variable for blacks[19] and the black category on the race/ethnic variable for nonblacks, (b) dropping immigration generation (since it is very skewed among blacks), and (c) introducing the black/white contact variable.

The nonblack models (Table 11A) are very similar to the overall models (Table 10). As expected from the earlier bivariate analysis the segregation variable has no impact on misanthropy among nonblacks. The black models (Table 11B) differ from the overall and nonblack models in two main ways. First, most relationships are weaker and the r² is only a little more than one-third as large. Second, contact with whites is related to less misanthropy and is among only three or four variables significantly related to interpersonal pessimism (education, age, segregation, and (maybe) parental break-up). This further suggests that black misanthropy may largely be shaped by racial factors.

Conclusion

Misanthropy is shaped by socioeconomic and minority status, noneconomic life events, religion, and age-cohort. First, it is higher among the less educated, those with lower incomes, and those with recent financial reversals (but among those with downward intergenerational mobility). The especially strong education effect compared to the smaller income and financial situation effects suggests schooling has an impact beyond material and social standing. A college education may cultivate a more benign view of the world and of humanity.

Second, it increases among subgroups toward the social periphery: racial and ethnic minorities (especially blacks), Southerners, and (perhaps) immigrants. Among blacks it is greater among those with limited contacts with whites. The more subordinate a group is and more isolated members of the group are, the greater the misanthropy.

Third, it is greater among those with recent or on-going, noneconomic problems—victims of crime and violence and those in poor health.

Fourth, it is higher among those who do not attend church and among fundamentalists. These effects are counter to each other since fundamentalists tend to be more frequent church attenders than nonfundamentalists. Church attendance probably diminishes misanthropy both because attendees tend to be people with faith (in God, their church, and, perhaps by extension, their fellow citizens) and because of the positive interpersonal ties that congregations further. Fundamentalists however are taught a theology that both stresses the sinful nature of people and tends to divide people into the few devout and the many fallen away.

Finally, misanthropy is greater among younger adults/members of more recent cohorts. The relationship between age/cohort and misanthropy may help to explain the trends over time. At least over the past two decades the evidence points to some decline. The MTF series start to tip downward in the early to mid-1980s and the GSS scale shows a modest, cumulative decline after the 1993/94 points are added to the 1972–1991 time series. Other series also mostly point in a negative direction, but the pattern and magnitude of the change is complicated and not very clear. If the MTF youths are undergoing an Mannheimian socialization process and if some substantial part of the current age differential among adults represents a cohort effect, then misanthropy should grow in the near-term future due to cohort replacement. Davis’ work (1995) suggests that this is at least partially the case with cohort replacement accounting for just over 40% of the total change in trust from the 1970s to the 1990s.

Table 11. Multiple Regression Models for Misanthropy by Race (Standardized Coefficient/Probability)

A. Nonblacks
Variables	Model1	Model2	Model3
Household income	2.050/.000	2.059/.000	2.050/.002
Education	2.207/.000	2.195/.000	2.181/.000
Intergenerational mobility	2.014/.075	2.032/.016	2.019/.194
Recent financial changes	.075/.000	.062/.000	.056/.000
Marital status	.008/.368	2.029/.184	2.013/.407
Parents divorced	2.023/.002	2.020/.131	2.011/.418
Ever divorced	.043/.000	.028/.039	.007/.652
Community type	.035/.000	.029/.024	.023/.104

Race/ethnicity

Hispanic	.078/.000	.075/.000	.069/.000
Asian	.028/.000	.010/.428	.008/.542
Amerindian	.037/.000	.046/.001	.032/.031
Late immigrants	.090/.000	.092/.000	.077/.000
Middle immigrants	.045/.000	.027/.104	.026/.155
Has religion	.024/.004	.005/.737	2.009/.562
Attends church	2.082/.000	2.084/.000	2.082/.000
Theology	2.062/.000	2.050/.001	2.036/.021
Age	2.170/.000	2.202/.000	2.196/.000
Gender	2.045/.000	2.042/.001	2.016/.288
Region	.073/.000	.072/.000	.082/.000
Mobility	2.015/.056	2.012/.381	2.010/.487
Robbed/burglarized	—	—	.035/.012
Hit/threatened	—	—	.086/.000
Hospitalized/disabled	—	—	2.004/.775
Health status	—	—	.096/.000
Unemployment	—	—	.013/.389
Family deaths	—	—	2.027/.061
Segregated	—	2.011/.418	2.004/.760
r2	.136	.138	.153
N	16,118	5,384	4,320
B. Blacks

Variables	Model1	Model2	Model3
Household income	2.063/.013	2.048/.280	2.044/.375
Education	2.143/.000	2.168/.000	2.148/.003
Intergenerational mobility	2.000/.985	2.003/.948	2.012/.785
Recent financial changes	.069/.003	.000/.996	2.019/.651
Marital status	.061/.028	.068/.185	2.088/.109
Parents divorced	.005/.822	2.078/.043	2.062/.134
Ever divorced	.006/.814	.043/.323	.007/.889
Community type	2.019/.430	2.003/.948	2.037/.426
Has religion	2.063/.013	.047/.301	.045/.359

TABLE 11—Continued

B. Blacks—Continued
Variables	Model1	Model2	Model3
Attends church	2.059/.020	2.068/.113	2.043/.352
Theology	2.023/.386	2.004/.938	.013/.794
Age	2.116/.000	2.125/.015	2.130/.030
Gender	2.030/.187	2.012/.755	2.005/.915
Region	.012/.649	.015/.756	2.008/.870
Mobility	2.008/.740	2.029/.477	2.029/.499
Robbed/burglarized	—	—	.056/.183
Hit/threatened	—	—	.086/.058
Hospitalized/disabled	—	—	.056/.226
Health status	—	—	.003/.955
Unemployment	—	—	.010/.830
Family deaths	—	—	.023/.602
Segregated	—	.103/.014	.105/.018
r2	.052	.053	.055
N	1,999	667	588

It is also instructive what does not explain misanthropy. First, urbanness has no direct impact on negativism and the one distinctly misanthropic region, the South, is not highly urbanized. Thus, the notion that closely-knit, small-town communities engender faith in people is not supported. Second, having gone through a divorce as a child or as a spouse does not increase misanthropy. Whatever the long-term impacts of divorce maybe, misanthropy does not appear to be one of them. This is particularly noteworthy since it had been hypothesized that the “intimate betrayal” of parental or spousal divorce would have a strong impact on faith in people. Third, never having been married does not lead to more pessimism. Thus, while isolation has an impact in other circumstances (e.g., minority status), being without a spouse does not exert a similar effect.[20] Fourth, geographic mobility has no clear relationship. Its impact may pull in opposite directions as the competing hypotheses had suggested. Fifth, gender is unrelated. Finally, while some traumas (e.g., victimization) are associated with more misanthropy, others (e.g., deaths, unemployments, and hospitalizations) are not. Combined with the null divorce findings, this indicates that some, but not all, negative life events influence misanthropy.

References

[1] For some alternative conceptualizations or refinements of misanthropy and interpersonal trust see Lagace and Rhodes, 1988; Lewis and Weigert, 1985a & 1985b; Robinson, Shaver, and Wrightsman, 1991; Yamagishi and Yamagishi, 1994.

[2] Factor analysis available from author.

[3] We discuss both trends in the individual components of the scale as well of the scale itself because (a) there is considerable data that relates to the individual items rather than the scale and (b) there are differences in trends across the components.

[4] Looking at people 18–24 on the GSS does not indicate that Don’t Know responses are higher among young adults. Thus age, is not likely to be a major factor in explaining the differences.

[5] Putnam compared the 1960 point in Table 2B with the 1993 GSS point in Table 4A.

[6] Despite the context effects on two of the three items on the GSS, it is possible to combine them together into a relatively unbiased, across-time measure. First, the trust and helpful context effects are largely off-setting with trust high in surveys (and years) that helpful is low and vise versa. Second, because of this, there is no statistically significant context effect for the scale (means in 1988–1994 are 5.91 for context1 and 6.01 for context2; probablity 5 .056).

[7] That is, the correlation for 1972–1991 is only .010 and probability 5 .191. Only with the addition of the 1993 data point does the relationship become statistically significant.

[8] See also Yamagishi’s (1995) analysis of the joint trends of trust and helpfulness.

[9] The social contact literature (e.g., Powers and Ellison, 1995; Sigelman and Welch, 1993) suggests that interracial contacts that are cooperative, voluntary, between people of equal status, and endorsed by authorities lead to better, intergroup relations. Presumably this involves increasing intergroup trust which for the minority group should in particular promote judgments that people in general are trustworthy.

[10] But Robinson, Shaver, and Wrightsman (1991) find a curvilinear association with misanthropy highest among the young and the old.

[11] Lagace and Gassenheimer (1989) found no gender differences on global trust, but found that men were more suspicious than women. In a student sample Terrell and Barrett (1991) found that women were less trusting than men were.

[12] An analysis of the negative life events scale used on the 1991 GSS (Smith, 1992) also showed these to significantly relate to misathrophy.

[13] This negative assessment on interracial trust is also shared by the general population. In response to the question “Do you think most blacks trust whites, or do you think most blacks don’t trust whites?” 17% said Trust, 67% Don’t Trust, and 17% Don’t Know and Other (New York Times poll, 2/1994, n 5 1193).

[14] On black mistrust of whites in general see Biafora, Taylor, Warheit, Zimmerman, and Vega, 1993; Jeanquart-Barone, 1993; Kuran, 1993; and Taylor, Biafora, and Warheit, 1994.

[15] We searched for the proposed Holocaust connection by looking at misanthropy by immigrant status, birth cohort, and period, but found only minimal support. Details available from the author.

[16] However, three other measures of geographic mobility asked in single years showed a mixed pattern. In the 1987 GSS having recently changed residences was associated with more misanthropy. In the 1986 and 1987 GSSs two measures of how long one had lived in the local community were not significantly related to misanthropy.

[17] The association with alcohol consumption is nonlinear. Misanthropy is highest among nondrinkers (6.0) and heavy drinkers (5.9) and lowest among moderate drinkers (5.6).

[18] And in an alternative model based on those with occupations, it is also greater among those with lower prestige jobs.

[19] There may be some “ethnic” variation among blacks. Biafora, et al. (1994) found differences among blacks by immigration status and place of origin. However, Ostheimer (1985) didn’t find that misanthropy differed by how blacks expressed their ethnic background.

[20] In 1986, 1987, and 1988 different measures of number of friends (or similar) were included on the GSS. In general, misanthropy was lower when people had more friends, but only the 1987 variable based on a listing of people with whom one had discussed important personal matters recently was significantly related when added to Model3 in Table 10.