Visualizing Bring-backs

Abstract

The figure plots the number of articles that have attempted to “bring” something “back in” in the social sciences by publication year and number of citations. Andrew Abbott, taking a (pessimistic) sociology of knowledge perspective, identified this tendency—beginning with Homans’s classic article “Bringing Men Back in”—as emblematic of the tendency to rediscover old ideas in sociology. The plot shows that “bring-backs” did not become a common yearly occurrence until the mid to late 1990s but are now relatively frequent. The most successful bring-backs have been relatively abstract things such as the “state” and “society” and more recently, “culture,” “knowledge,” and “values.”

The publication of Homans (1964) brought an interesting thing into sociology: the repeated attempt of people to bring something believed to be ignored “back in.” Abbott (2001), taking a (pessimistic) stance, identified this as emblematic of the tendency to rediscover old ideas in sociology, noting that about 91 things had been brought back at the time of his writing. We set out to update and expand the quest for things brought back. To generate the population of bring-backs, we used the keywords: intitle: “bringing” intitle: “back in” in Google Scholar, capturing all entries having the words bringing and back in in the title. After cleaning, removing duplicates, and consolidating different versions of the same publication, we ended up with 1,303 entries published since 1964. ¹

Figure 1 summarizes the main patterns, plotting year of publication in the horizontal axis against the total number of citations in the vertical axis. The upper histogram shows the number of articles bringing something back every year. The right-side histogram shows the univariate distribution of citations among bring-backs (predictably skewed). Two sets of points are labeled in the main plot. On the top half of the plot, we label all points that fall above the 95th percentile of total citations among articles cited at least once published between 1964 and 2005 (successful bring-backs). Toward the bottom left, we label all papers published before 1985, allowing us to see the range of “early adopters,” both successful (upper left) and less successful (lower left). We used automated text analysis to code entities being brought back in on a gradient of abstract to concrete (Bhatia and Walasek 2016). Orange labels fall toward the abstract side, and purple labels are toward the concrete side (see Supplementary Material for details).

Figure 1.

Scatter plot of citation totals for papers in Google Scholar published between 1964 and 2018 claiming to be “bringing [the] _______ back in” and having at least one citation. We used the rvest package in R (Wickham 2016) to collect Scholar results using the search operator: intitle: “bringing” intitle: “back in” (see Supplementary Material for more details on data cleaning procedures). The y-axis is the total number of cites the paper has received since publication; the x-axis is the year of publication. We scaled the y-axis to the logarithm of the total citation count. Marker size is proportional to the raw citation count for each paper (see legend on left). Labels appear next to markers for papers with a total citation count at or above the 95th percentile among papers with at least one citation published on or after 1964 but before 2005 (upper region of plot). We also label markers for all articles published before 1985 with their attempted bring-back (lower left region). Marker colors vary proportionally with the brought-back thing’s construal scores, ranging from 1 (bright orange and very abstract) to 5 (dark purple and very concrete). Supplementary Materials provide more information on the construal scores. The marginal histograms are the univariate distributions of citation counts (right) and year of publication (top). Histogram bins for the year of publication (top) are equal to one year. The density plots in the top right inset show the logged citation counts for the top “repeat bring-backs”: things that have been brought back in by more than 20 papers, colored with the construal scores of the repeated bring-back.

Two patterns are evident in the figure.

First, the uptake of bring-backs took quite a long time. Two decades after Homans (e.g., 1984), only a relatively small number of papers tried to bring something back. After 1985, we see a small uptick and another one after the early 1990s. These more or less correspond to the publication of the highly successful bring-backs of Evans, Rueschemeyer, and Skocpol (1985) and Friedland and Alford (1991), who brought the state and society back in (to historical sociology and institutional analysis, respectively). After the mid-1990s, we see a gradual climb in the number of bring-backs; by the mid-aughts, bring-backs become a regular part of the scholarly landscape, peaking in 2008. Second, bring-back successes are tilted toward the abstract side of the ledger. While there have been some successful concrete bring-backs, such as bodies, strong ties, teachers, and even a repeat bring-back of “(the) men” by a feminist sociologist (Reskin 1988), they are relatively underrepresented among the successful cases (and relatively overrepresented among early failures). In addition to the state and society, recent abstract bring-backs successes include knowledge, culture, values, context, power, politics, and history. Even the top “repeat bring-backs” (things that have been brought back in by more than 20 papers) are dominated by abstract things, with “people” being the only exception (see top-right inset).

An irony presents itself. While Homans wanted to bring (concrete) “men” [sic] back in, the things that have been most successfully brought back after his call are closer to the macro-abstractions he railed against.

Supplemental Material

bringitback_supp – Supplemental material for Visualizing Bring-backs

Supplemental material, bringitback_supp for Visualizing Bring-backs by Omar Lizardo, Dustin S. Stoltz, Marshall A. Taylor and Michael Lee Wood in Socius