WLS: The long-term study that almost wasn’t

It’s now one of the longest social science investigations ever. Yet, at the beginning, the Wisconsin Longitudinal Study (WLS) wasn’t meant to last.

Faced in the late 1950s with growing numbers of high school students who hoped to attend college, the State of Wisconsin commissioned the UW–Madison School of Education to conduct a one-time survey of all high school seniors. In 1957, some 30,000 students answered four pages of questions about their family backgrounds, high school courses and plans after graduation. The state then used this information to help plan its expansion of the UW System of colleges.

The study may very well have ended there, if not for a discovery on the UW–Madison campus years later that was brought to the attention of sociology professor William Sewell. "Somebody apparently came to Sewell in 1962, saying that there were a bunch of old, paper questionnaires in the basement of Bascom Hall," says UW–Madison sociologist and WLS director Robert Hauser. "I never learned who it was, but this person thought the questionnaires might be of interest to him."

When he strolled over to take a look, Sewell found the 1957 higher education surveys, which serendipitously included the names of the graduates and the names and addresses of their parents. The find delighted the pioneering sociologist. He’d been interested in the formation and consequences of youthful aspirations for a long time, but had lacked an appropriate population to study – until now. Soon, he began planning his first follow-up with the class members.

One major question remained, though: Would they respond? "Bill was a very clever guy, but he was fighting against the prevailing belief in the social sciences back then that you couldn’t do long-term follow-ups because you simply couldn’t get people to participate," says Hauser. "And, at that time, several years had already passed."

After ruminating further – especially on the habits and attitudes of his own twenty-something children – Sewell realized that the graduates, now 25-years-old, might indeed be reluctant to take part. But, he reasoned, parents are always eager to talk about their kids.

Thus, in 1964, Sewell randomly selected one-third of the participants from the original education study. To the parents of this "sample" he sent a brief survey on a postcard, asking about the graduates’ post-high school education, marital status, occupation and military service. To those who failed to respond he sent another postcard, and then another, eventually following up five times in all by mail and phone.

In return, Sewell got a response rate of 87 percent that Hauser says was "just unheard of at the time." In one fell swoop, he had both successfully collected another round of survey data from the Wisconsin high school class of ’57 and demonstrated conclusively that longitudinal research could be done.

"One of the big effects of this study was that it inspired a whole series of federal surveys following cohorts of people for long periods of time, which have been carried out more or less regularly since then," says Hauser.