This post continues where "Is Society Coming Apart? Part I" ended.

The second half of Coming Apart deals with the new Lower Class, which Murray defines as people in the lower 30% of an index of cognition that reflects a person’s education and the cognitive demands of a person’s occupation. As a literary device, Murray contrasts “Belmont,” an upper middle class district containing people in the top 20% of Murray’s index, and “Fishtown,” where the people are in the lower 30% of the index. Belmont and Fishtown are modeled on actual districts in Philadelphia. Murray homogenizes them by classifiying households as residents of Belmont if at least one member of the household has a college degree and if the head of the household is employed in a professional or management position, and as a member of Fishtown if the highest level of education in the household is high school graduate or below, and if the head of the household is in a blue collar, lower white collar (e.g. receptionist), or service occupation (e.g. hair stylist). He then looks at how each group fared from 1960 to 2010, using as criteria various aspects of social capital, such as the fraction of two parent families, health indices, indices of social networking (e.g. churches, service clubs), and criminal records.

Murray found surprisingly little social change in Belmont over 50 years. There was been some decline in marriage (it is not clear whether he regards civil union declarations as marriage), some increase in single parent families, and, in spite of what I think many academics might expect, a startling array of social networks, including various civic organizations.

Fishtown is far different. Here Murray finds sharp declines in two parent, stable families, a deterioration of social networks, and (most ominous in his analyses) an increase in the number of men 30-49 years old who are neither employed nor looking for work. In British terms, many of these people would be on the dole and resigned to it. Murray points out that at least until recently American society expected adult males to either have a job or be actively seeking one.

Why has Fishtown deteriorated? Two causes have been suggested. The first is psychological. Suppose Goddfredson (1997) was right, life is an intelligence test. It may be that the early 21st century version of the test is just a bit too hard for people in the lower 30% of the distribution. While Murray does not use these terms, the explanation seems consistent with most of his thinking. However there is also a sociological explanation.

The residents of Murray’s statistical Fishtown were defined by a combination of educational attainment and occupation. Over the past 50 years there has been an increase in the frequency of college graduates and an increase in the percentage of people in the workforce who hold mid-level managerial and skilled technician jobs. There has been a decrease in the frequency of blue collar work, especially in manufacturing. Murray points out that in the 1960s a larger percentage of the population would have been categorized as “Fishtown” than is the case today. One could argue that in the 1960s “Fishtown” and similar communities were held together by the leadership of the most capable people amongst the blue-collar industrial and skilled service population groups. As the economy changed the more capable members of the Fishtown community upgraded their skills and migrated away from Fishtown, not to Belmont, but to the middle-class communities in between. This left the Fishtown society without its leaders, and the deterioration that Murray documents followed.

There is an interesting parallel between Murray’s sociological explanation, which is based exclusively on data for White Americans, and explanations that have been offered for the persistent prevalence of social problems in poor African-American communities. W.J. Wilson (1997), in particular, has been an articulate proponent of the view that the end of segregation and legal discrimination in the 1950s primarily benefitted the better educated, more cognitively skilled African-American upper classes, who then moved from the inner-city areas.

The psychological and sociological explanations are not mutually exclusive. The cognitive skills required by the modern workplace have changed, and there has been a selective outward movement from the city to the suburbs. Both explanations emphasize the importance of having a good score on the intelligent test of life.

I will now shift from describing Murray’s work to critiquing it.

Has Murray presented a fair picture of the data?. His data sources, census data and widely respected survey reports, are certainly correct. As was the case in The Bell Curve, Murray presents his analyses by using simple graphics that deal with only two or three variables at once. Statisticians who criticized The Bell Curve offered more comprehensive multivariate models that made a case for minor qualifications to the story presented by the simple graphs. See, for instance, the papers in the Devlin et al. (1997). A similar multivariate approach would probably provide minor qualifications of the analyses presented in Coming Apart. However Coming Apart is not addressed to the members of the American Statistical Association, it is addressed to those Overeducated Intellectual Snobs who Murray believes influence decision making. The models used to question The Bell Curve were presented in multi-page tables. I suspect that the analyses needed to qualify Coming Apart would require similar reporting. The tables might be closer to THE WHOLE TRUTH, but the graphs communicate ideas better.

I believe Murray is right about the isolation of the New Upper Class. The gated community was virtually unknown fifty years ago. (I remember my surprise when I encountered one in 1967. I promptly bluffed my way past the watchman.) Although Murray touches on this only tangentially, the proliferation of specialized cable-television channels allows a segregation of sources of information that may be even more isolating than the physical isolation afforded by the suburbs.

The discussion of Belmont and Fishtown present a contrast between the life styles of people in the top 20% or bottom 30% of Murray’s index of cognitive skills. What about the middle 50% of the population? By focusing on extreme groups Murray gives the impression of a social, economic, and possibly cognitive chasm between different segments of American society. Does the (unstudied) middle class provide a smooth path between Fishtown and Belmont, or are there washed out bridges and potholes along the way? And if so, where are they? The answer to this question is important for understanding inter-generation mobility across social classes. Moves from Fishtown to Belmont, or the other way, may be rare. How rare is the two generation move, from Fishtown to Midville to Belmont (or the other way).

The talking heads on Sunday morning TV could profitably discuss for weeks the meaning of Murray’s results for society at large. I will take a narrower view. What are the research questions that Murray raises for psychologists, and particularly for psychologists interested in intelligence?

There is a methodological issue. In the United States relatively few surveys include explicit cognitive test scores. They often contain information that is known to be statistically related to intelligence test scores. Therefore it is possible to construct inferred cognitive scores. In Murray’s case an estimate of intelligence was constructed by combining the known positive relation between educational status and cognitive score with an index of the cognitive complexity of a person’s occupation developed by Tara Madhyastha and myself (Hunt & Madhyastha, 2012).

I think that this approach is increasingly likely to be useful. It is particularly likely to be used by industrial researchers who have access to the very large commercial databases that are now being compiled. For example, a researcher who had access to a credit card company’s records would not find an intelligence test score there, but, given an adequate program of research, could make a pretty good estimate of a person’s intelligence. According to Google’s privacy statement (as of April, 2012) the company records every search that a user makes. There is a lot of information about intelligence there! What is needed is both the development and widespread dissemination of knowledge about how to use such indirect indicators. The social sciences may have to revisit the curriculum on multivariate analysis yet again.

Murray’s results show that we should amplify on the idea that life is an intelligence test. To what extent can the success of Belmont and the failures of Fishtown be traced to inabilities to manipulate certain key aspects of life, such as financial management? How does individual intelligence relate to understanding of information in the media, including consideration of likely biases in the source of the information? Where do people at different socioeconomic levels, and with different levels of intelligence, get their information about health care? Here I am not calling for a series of studies the will produce yet more correlations between test scores and this or that aspect of daily life. We need studies of the process by which information is acquired and used, as influenced by both intelligence and socioeconomic status.

Conservatives and Libertarians may see the isolation of the New Upper Class as a reason to call for decentralizing of decision making. However I do not think that this is going to happen. We live in a highly interconnected society. Politics in the Mid-East can drive the price of fuel up or down, which in turn has implications for transport costs, thus influencing the price of Costa Rican bananas in Michigan in January. Central control is always the most efficient way of controlling a system of interconnected, providing that there is no cost for computation at the center and no loss of information transmission between the central decision maker and the peripheral actors. These conditions are never satisfied. However the development of ever more sophisticated computing and communication technology does increase the efficiency of centralized decision making. I believe that trend will continue, so it would be a good idea to investigate ways of mitigating the isolation of the decision makers. Here are a few of the issues.

1) What are the relative roles of personal intelligence and social context upon decision making? Imagine that we gave a non-verbal, knowledge reduced intelligence test, such as a progressive matrix test, to every resident of Belmont and Fishtown. We then study their information gathering and decision making processes in everyday life. What would be the best predictor of behavior; community identity or personal intelligence? In such an experiment it would be important to go beyond trends and averages, to investigate variability. The behavior of the high scorers in Fishtown and the low scorers in Belmont would be of particular interest.

2) With a few exceptions, decision makers do not start their careers in important positions, they work their way into them. How should early careers be planned so that decision makers will retain some feeling for people outside their own social/professional circle? To illustrate, the worldwide coffee shop chain, Starbucks, has many of their professional level employees begin their careers by training as coffee servers (“baristas”). Does first-hand experience at the ground level improve later executive performance? At higher levels, to what extent is it feasible to have CEOs move from one industry to another? The United States seems to be something of an extreme in the extent to which appointees outside of the government are brought in to fill executive positions in government departments (e.g. Undersecretaries, ambassadors) that, in European countries, would be filled by members of the professional civil service. A case can be made for “injecting new blood” so that government functionaries are not isolated from the rest of society. A case can also be made for experience at a low level as a perquisite for holding high level government positions. How should these demands be balanced?

3) Murray and Wilson, for the African-American community, argue that one of the reasons for the deteriorating position of low SES communities is that as the post-industrial economy opened up new opportunities and closed others people whose cognitive and social backgrounds had prepared them to take advantage of the changes simply did so. They moved out of Fishtown, depriving it of its natural community leaders. This explanation raises a host of questions about the role of leaders in social networks, and the traits those leaders require. Barack Obama was once involved in establishing community networks in Chicago. So was Al Capone, though not at the same time nor for the same type of network. The differences between these leaders are pretty clear. What are the commonalities?

There are two very different ways of looking at the issues raised in Coming Apart. From the perspective of social sciences, Murray has given them a host of questions. They will not be answered by research in any one discipline. Can we assemble the necessary interdisciplinary teams?

But from the perspective of a decision maker the issue is more immediate. Are these differences problems to be ameliorated? And if so, how? In Coming Apart Murray is silent on corrections, though he does discuss some in his March 7 New York Times piece. I will leave their discussion for another time.



College Board (2006) Trends in College Pricing. New York: College Board Trends in Higher Education Series.

Devlin, B., Fienberg, S.E., Resnick, D.P., Roeder, K. (eds.) (1997) Intelligence, genes, and success: Scientists respond to The Bell Curve. New York: Springer-Verlag,

Friedman, T.L. (2005) The world is flat: A brief history of the 21st century. New York: Farrar, Strous, & Giroux

Friedman, T.L. (2006) The world is flat: A brief history of the 21st century 2nd edition.. New York: Farrar, Strous, & Giroux.

Friedman, T.L. (2007) The world is flat: A brief history of the 21st century 2nd edition (Further updated and expanded: Release 3.0).. New York: Farrar, Strous, & Giroux.

Fukuyama, F. (2011) The Origins of Political Order: From Prehuman Times to the French Revolution. New York: Farrar, Straus & Giroux.

Gottfredson, L.S. (1997) Why g matters: The complexity of everyday life. Intelligence. 24(1): 79-132.

Herrnstein, R.J. (1973) I.Q. in the Meritocracy. Boston: Little, Brown.

Hunt, E (2011). Human Intelligence. New York: Cambridge U. Press.

Herrnstein, R.J. , Murray, C. (1994) The Bell Curve. Intelligence and Class Structure in American Life. New York: The Free Press.

Hunt, E. (1995) Will we be smart enough? A cognitive analysis of the coming workforce. New York: Russell Sage

Hunt, E. (2012) What makes nations intelligent? Perspectives on Psychological Science. (to be published May 2012).

Hunt, E., Madhyastha, T.M. (2012) Cognitive Demands of the Workplace. J. of Neuroscience, Psychology, and Economics. 5 (1) 18-37.

Jensen, A. R. (1969) How much can we boost IQ and scholastic achievement? Harvard Educational Review, 39(1), 1-123.

Lynd, Robert S. and Helen M. Lynd. Middletown: A Study in Contemporary American Culture. New York: Harcourt, Brace, and Company, 1929

Murray, C. (200) Human Accomplishment: The Pursuit of Excellence in the Arts and Sciences, 800 BC to 1950. New York: Harper Collin

Murray, C. (2012) Coming Apart: The state of White America 1960-2010. New York: Crown Forum.

Reich, R. (1991) The work of nations: Preparing ourselves for 21st century capitalism. New York: Knopf.

Zuboff, S. (1988) In the Age of the Smart Machine: The Future of Work and Power. New York: Basic Books.

About the Author

Earl Hunt, Ph.D.

Earl Hunt, Ph.D., is Professor Emeritus in psychology at the University of Washington.

You are reading

Exploring Intelligence

If Everyone Gets Smarter

This is a speculation about what might happen if everyone became smarter.

Why we test determines what we test

Before selecting a test, consider why you are testing at all.

Improving Intelligence

How can we improve intelligence and how do we know if a method works?