An empirical review of 75 years of Population Studies
A new open access article published today in Population Studies' 75 year diamond anniversary issue empirically reviews the journal’s storied history. Analysing almost two thousand articles in the journal's 75 years of existence, LCDS Director Melinda Mills and co-author Charles Rahal take a scientometric approach to consider pertinent issues such as the structural changes correlated with editorship, the variety of different topics over time, and inequalities in authorship and authorship networks more broadly.
As described in the article, the journal Population Studies has a fascinating history which many may perhaps be unaware of. Founded by the Population Investigation Committee with David Glass as its inaugural editor in 1947, it was for some considerable time the only English-language journal devoted entirely to demography. In an alternate universe, suggestions to have named it more narrowly as the ‘British Journal of Demography’ may have been followed, but were resisted. Glass edited the journal for 31 years, with Eugine ‘Grebby’ Grebenik joining as a co-editor in 1954 (where Grebby himself served – quite remarkably – until 1992). Following this, John Simmons held the editorship for a remarkable twenty years before abdicating to our very own LCDS John Ermisch.
The authors created various datasets from numerous APIs to systematically analyse the seminal work published in the journal over the past 75 years. For example, the variance in article pages decreased substantially over the four editorial regimes (from 8.5 under Glass and Grebenik, to 3.4 under Ermisch’s current leadership), the average number of co-authors per paper skyrocketed, and the percentage of articles which were solo authored fell from nearly 95% in the 1950s, to under 20% in 2020.
However, some of the most substantial contributions of the paper surround an analysis of the topics of research published within Population Studies. In the earlier years of the journal, the focus was largely on ‘macro’ style factors, with migration being a key early topic in the late 1940s, and re-emerging in recent years and as part of Robert Woods’ involvement as an editor. Indeed, the paper showed patterns consistent with the interests not just of who was the editor-in-chief, but of the editorial board more broadly.
Using unsupervised machine learning algorithms they were able to identify the areas of research which were most frequently published upon; fertility decline, mortality and lifetables, child mortality, family demography, and international demographic comparisons. Using the tools of natural language processing, they developed functionality to identify whether papers discussed men, women or children separately or together, and what this variation looked like across topics of study.
Women were much more frequently analyzed in terms of fertility and family, and children more frequently in terms of mortality. For men it was war, migration and ‘Other’ broader topics.
As part of this work, seminal articles endogenously rise to the top of our visualisations, including Bumpas and Lu’s article published in 2000 on ‘Trends in Cohabitation’, and Preston’s seminal ‘curve’ article in 1975 which links mortality and economic development.
As shown in the figures above, and mirroring Mills & Rahal's earlier work which analyzed the Scientometrics of Genome Wide Association Studies they again considered the inequalities in academic authorship. They first showed that while trends in gendered authorship had increased towards equality (from 25.2% female authorship in the first half of the journal’s existence, to 36.8% in the second), female authorship remained below that of males. Indeed, when ranking authors by their custom ‘PopStudies H-Index’, they found no women near the top of the table; it was dominated by the perhaps familiar sounding names of male demographers such as John ‘Jack’ C. Caldwell, Samuel H. Preston, and Michael Murphy. International trends in global authorship patterns increased substantially from the journal’s beginnings out of the LSE, but authors affiliated with the LSE still contributed a substantial proportion of the article’s content over time (at 4.66%). This aligns with the broader continental dominance of European and North American based authorship. They also revealed a tight, dense core of the co-authorship network which is currently rapidly expanding due to the emergent pattern of large and substantial pieces of collaborative research.
Reflecting on the work, author Charles Rahal commented; “It has been a great honour to be able to contribute to this Special Issue in honour of the journal’s 75th anniversary, and indeed ourselves learn more about it’s fascinating history with help from Anne Shepard at the PIC. While key patterns in topics and authorship have dynamically changed with time in line with our prior expectations, this neutral scientometric review quantifies important narratives about what demographers do, who does it, and the networks which they form to do it within."
Author Melinda Mills added "We can’t wait to expand this line of work beyond just one journal to the broader field of demography and the academic literature more generally."
Also featured in this Special Issue is work by LCDS member Ridhi Kashyap Has demography witnessed a data revolution? Promises and pitfalls of a changing data ecosystem? available here.
The article is available via Open Access and can be found here.