An article by Larry E. Suter and Herman P. Miller titled “Income Differences Between Men and Career Women” was published in the American Journal of Sociology in January 1973. This article, which used data from the NLS Mature Women, was the first peer-reviewed article based on NLS data. [Read the abstract.]
Larry Suter and Herman Miller both worked at the U.S. Census Bureau when the research culminating in the 1973 AJS article was carried out:
In Spring 2015, Larry was kind enough to share the following recollections of his time as an early user of the NLS. In reading his account, bear in mind that researchers at the Center for Human Resource Research at Ohio State University were awarded the initial contract to conduct the NLS; they sub-contracted with the Census Bureau to handle respondent interviews. Larry alludes to the fact that NLS data were difficult to access in the early years, to the point that researchers at Ohio State relied on researchers at the Census Bureau for analysis of the data. Readers who use NLS data via “instant” downloads enabled by NLS Investigator should give thanks for modern technology!
Larry Suter’s remarks:
Thank you for telling me that this was the first published article on the NLS. I guess I am only slightly surprised that our article was an early publication because one of the reasons I was given access to the data was to assist the researchers at Ohio State in obtaining publications and I was close to the source of data collection at the Census Bureau. I later wrote other papers from other cohorts of the NLS and have been on the mailing list of their reports since then.
I began working full time at the Population Division of the Census Bureau in fall 1969. Herman Miller was the division director and I was a statistician in the Education and Social Stratification branch just out of graduate school. In 1972 I was chief of the Branch. Herman was an activist economist who was engaged in using the data from the Census Bureau to answer significant social and economic questions. He had just published Rich Man Poor Man, which was a best seller in 1967. In 1972, the “Parnes” longitudinal surveys of young men and women and middle age women were underway at the Census Bureau. Because I’d had modern training in the use of computers for statistical analysis of demographic and social statistics, Herman asked me to look into the use of the NLS. As an economist, he had noticed that discussions of women’s income had failed to account for their lifelong work history and we noted that the NLS for women ages 30 to 45 had obtained thorough lifetime work histories so it might be possible to account for the extent to which long work lives were responsible for lower incomes for women.
In my graduate education, I had used multiple regression and path models to compare Hispanic population income with white incomes. That experience was relevant for solving the problem of income differences between men and women. But the NLS women’s survey did not have estimates for men. I was discussing the dilemma with Robert Hauser of the University of Wisconsin who suggested a method that didn’t require that both populations be in the same equation. I borrowed that technique and found that after accounting for women’s educational level, marital status, occupational status, and lifetime work experience that they would still receive lower income than men with the same education and life experiences. I presented the paper at the American Sociological Association meetings in 1972 and one of the discussants (Charles Tilley) suggested that I submit the paper to the American Journal of Sociology. I did, and I included Herman as a co-author since the idea of tackling the problem was originally his.
In the conclusion of the paper I used the word “discrimination” in describing the reasons for the lower income of women since it seemed to me that all the most important conditions that account for income differences between individuals had been accounted for. Miller did not object to the use of that word. However, the politics of Washington had changed after the election of Richard Nixon and Miller was a subject of suspicion by the administration because he had published news stories about the economy that did not agree with the administration’s official position. One of the Census Bureau’s political appointees was assigned to review papers before they were released. That person took me to lunch and asked that I remove the word discrimination before submitting it for publication. However, I was feeling my graduate student oats and did not believe that he had presented a sufficient case for removing the word and I submitted the article as it was. The reviewers of the article recommended it for publication in an issue of AJS that contained a number of other articles about the status of women and it was turned into a special edition that was widely circulated. I was never overtly punished for my insouciance, but it was clear that Miller and I were not affected by efforts to report only positive views of the economic conditions in the early 1970s.
I was not the only person to analyze the lifetime experience of women with that data set however. Other economists later used a different model and found similar results.
My research on women’s work experience led to a number of other papers from the NLS that were written within the Census Bureau but never published. Because of my experience, I was interviewed by Gale Sheehy for her book Passages; you will find a number of references to our interview in the footnotes. Later, I was asked to be an expert witness in a federal court case that was determined by estimates of women’s lifetime work experiences. So the one paper on women’s cohort in 1973 had many spin off effects.
… to the thousands of users of NLS data whose research produced roughly 9,000 journal articles, books, monographs, and dissertations during the first 50 years of the NLS. Without a dedicated community of NLS users, there would be no need for the NLS!