Not the best fillers‑in of forms?

The contribution of education to economic development is difficult to assess. On the one hand, in contemporary settings, there is little doubt that education matters for innovation and growth — to say nothing of its direct benefits for those who receive education. On the other hand, historical evidence on education’s role is mixed. Scandinavia is a good case study: Denmark, Norway and Sweden all had relatively high rates of literacy by 1800, but were at the same time also relatively poor.

But maybe it is not the basic education level of the masses that, at least in pre-industrial times, had the greatest effect on development. Rather, we should perhaps look to what Mokyr describes as the ‘upper tail’: the relative few with a high level of useful skills. Some existing empirical work provides tentative support for this idea, but comprehensive measures of the ‘upper tail’ remain elusive. In a new working paper, Kristin Ranestad, Paul Sharp and I document a novel set of sources for charting the educated elite: biographies.

Both Denmark and Norway had a tradition of graduate yearbooks through the nineteenth and early twentieth centuries. These yearbooks were published in connection with high school reunions — typically 25 years, but even 40, 50 and 60 years after graduation. The yearbooks included biographical profiles of virtually every student who had completed the high school graduation exam, and contain information on graduates’ backgrounds, their studies and their employment history. One example is shown below.

For the most part, the information included in the biographies was sourced from graduates themselves. Where graduates had died, information was often sourced from surviving relatives. Furthermore, the editors of the various yearbooks often supplemented individuals’ responses with additional sources — historical enrolment material as well as other existing biographies and relevant accounts. While graduates could choose to be excluded from the yearbooks, remarkably few appear to have avail themselves of this option. As one yearbook editor put it to us, most graduates valued being commemorated in text.

In contemplating the use of new sources in research, it is prudent to critically appraise the material. Why was it produced? Who produced it — and where did they get their information from? Is the information genuine and representative of what it purports to cover? In short, we ask, are the biographies reliable for our research?

Credibility check

We use a framework to guide our source critical analysis, which covers four parameters (outlined in the box below). We discuss these four parameters in detail in the paper, but our headline finding is that that the biographies are indeed reliable for our purposes.

Our conclusion builds principally on three observations. First, while the style and structure of the biographies changed from year to year — and differed between Denmark and Norway — the information included is broadly consistent. Similar questions were asked of graduates over time, allowing for a high level of comparability for our analysis.

Second, response rates among cohorts were generally around 90 per cent. When we compare the number of biographies with the recorded number of graduates in the enrolment records, there is very little difference. The yearbooks give a comprehensive overview of graduate cohorts in their entirety.

Third, we observe that the editors took the job of compiling the graduate yearbooks seriously. There is much evidence of the editors repeatedly reminding graduates to send in information to enable the biographies to be written. In some cases, graduates themselves took responsibility for gathering and publishing the information. In other cases, professional historians and genealogists were contracted to produce the yearbooks. Indeed, the ongoing involvement of selected genealogists over several years suggests that they were well-regarded for the quality of the work they produced.

One might also note that the incentive for graduates to overtly lie or otherwise give misleading information was relatively limited. The primary audience for the yearbooks consisted of the graduates themselves. And in a era when those who graduated high school constituted a small elite in society, the graduates often remained professionally and socially connected. The likelihood of false or incorrect information being detected was relatively high; a form of social ostracisation would not be an unthinkable consequence.

Having established the credibility of the yearbooks as a source, we are now in the process of building a graduate database which draws on the biographical information (along with other sources). Simply put, our sources will allow us to document the lives of graduates from cradle to grave. In turn, we can more precisely measure both educational attainment and employment outcomes at the individual level, while also controlling for socioeconomic backgrounds. The richness of the material we have continues to amaze me, and the possible applications of the data should keep my colleagues and me occupied for many years to come.