Is Political Science Meant for Every Tom, Dick, or Harriet? The Role of First Names and Middle Initials as Predictors of Academic Success

Abstract

We take advantage of the data set compiled by Masuoka et al. (2007, PS: Political Science and Politics 40 (1): 361–66) on lifetime citation counts of political science faculty at PhD granting institutions in the United States to look for “lucky names,” that is, names parents can give babies that predispose them toward scholarly success in political science. Seeking to test an hypothesis offered by Wuffle (1972, PS: Political Science and Politics (Summer): 290), we also briefly look at the importance of middle initials for citation success in political science.

Why are some successful and others not? Many scholars look to biological features that are given cultural meaning, for example, gender (Guinier, Fine, and Balin 1997; Monroe et al. 2008), good looks (Rosenberg et al. 1986), height (Hensley 1993; McCann 2001; Sorokowski 2010; ¹ Wilson, 1968; Young and French 1996), or race (Arrow 1998). Here, our interest will be in the determinants of citation counts, which have been found to be an important determinant of academic salaries in political science (Grofman 2009). ² We use the large data set compiled by Masuoka et al. (2007a, b) on lifetime citation counts of political science faculty at PhD-granting institutions in the United States ca. 2005. Unfortunately, variables such as good looks, height, and race are not found in the Masuoka et al. (2007a, b) data set, although we can approximate gender by using gender coding of first names. Moreover, as everyone knows, the most important predictor of success in any area is “luck,” but we found no plausible way to properly operationalize this variable.

In the light of these methodological limitations, we take our inspiration for choice of the key independent variable to predict citation success in political science from the art of naming, or “Nameology,” as this branch of scientific astrology is properly named. ³ With the notable exception of Cash's (1969) definitive musical monologue on the importance of baby names for the formation of gender identity, the science of naming was remarkably neglected in the social sciences until Steven D. Levitt and Stephen J. Dubner, in their magisterial work, Freakonomics (2005), demonstrated the insights that could be gleaned from studying the evolution of racial and class preferences for baby names as a Veblenian race to distance oneself from the Jones's. ⁴

The direct inspiration for this article, however, was the intention of the first-named author's long-time officemate to teach a graduate course whose working title was “The Importance of Being Gary,” using the seminal contributions to the discipline of Gary Cox, Gary Jacobson, and Gary King as a means of giving students an appreciation of first-rate political science methodology and research design. The potential for such a course suggested that mothers/fathers who wished their (male) babies to grow up to be political scientists should consider giving them this auspicious first name. But, perhaps there are other Garys who do not publish as often or as well, but one tends only to remember the Garys who do? ⁵ It was that question that led to this article. But, in the interests of comparative research, as the article has evolved, we go well beyond “Gary” studies.

In the remainder of this article, drawing on the Masuoka et al. (2007a, b) data set of lifetime citation counts of political science faculty at PhD-granting institutions in the United States ca. 2005, we examine citations according to first name. ⁶ There are 3,743 names in this data set. First, we compare all names with at least 20 instances, ⁷ of which there are 29, to see if there are some that appear statistically distinct in their mean per year citation counts from the overall average of 6.7 citations per year. ⁸ In particular, we check to see if “Gary” has an unusually high mean citation count. Then, we look to see if those whose first names match those of recent past presidents have a distinctive citation profile. Also we check whether sets of first names with Christian religious significance, such as the four authors of the Gospels (“Matthew,” “Mark,” “Luke,” and “John”), or famous names from the Old Testament (“Abraham,” “Isaac,” and “Jacob”), have an unusually high citation count relative to the entire data set. ⁹ Next, to check for gender effects, we take advantage of a natural experiment involving paired comparisons of matched pairs involving similar first names (e.g., Paul and Paula), ¹⁰ and we also compare some famous mixed gender pairings (e.g., Jack and Jill). Finally, to deal with potential confounding effects, we check to see whether those with middle initials or middle names have any citation advantage, and we look to see if those with first letters of their first name toward the end of the alphabet appear penalized relative to those with letters closer to the front.

DATA ANALYSIS

We first look to see if the “Gary” phenomenon is statistically significant. It is not. Although the mean yearly citation count of the 23 political scientists in the Masuoka et al. (2007a, b) data set with first name “Gary” or “Garry” is 14.6, more than twice as high as the 6.7 mean for the data set as a whole, when we use the t-test option with unequal variances, the p value misses statistical significance at conventional levels, (p = .08). The only two first names with more than 20 instances that have statistically significantly more than average yearly citations are “Kenneth/Ken,” with a mean of 11.9 (p = .04) and “Robert/Bob” with a mean of 10.1 (p = .047). But, of course, because we have 29 possibilities to obtain p values as low as .04, this is not convincing evidence that any common names are “lucky names” vis-à-vis political science. ¹¹ Although there are certainly names that are, at least in the United States, unusual and/or only one of which is found in our data set, an informal review revealed no clear link between uniqueness of name and academic citation success. For every first name such as “Arend” (Lijphart) or “Elinor” (Ostrom) or “Aristide” (Zolberg) or “Atul” (Kohli) associated with high name recognition in political science, there is a singleton first name in the Masuoka et al. (2007a, b) data set that has close to zero annual citations.

Turning to first names of recent US presidents, we find no advantage to having a president's name. Looking at the combined mean yearly citation count for “George,” “Bill/William,” “Ronald,” “Jimmy/James,” and “Dick/Richard” we find a mean citation count of 6.9 (p = .40) However, turning to names with great significance in Christianity or Judaism ¹² (the authors of the Gospels: “Matthew,” “Mark,” “Luke,” and “John”; the three (Catholic) archangels: “Gabriel,” “Michael,” and “Raphael”; iconographic figures such as “Peter,” “Paul,” and “Mary”; and the biblical patriarchs (“Abraham,” “Isaac,” and “Jacob”) ¹³ , we find that, in toto, that, as the Church Lady might have said, there is something “special” about the mean citation count of political scientists with one of those first names. As a whole, the group has a mean yearly citation count of 8.0 (p = .02). ¹⁴

Turning to gender differences, despite the title of our article (chosen primarily for its euphony), and although there are five “Harry's” and even more “Harold's,” there is no one in the Masuoka et al. (2007a, b) data set named “Harriet.” Thus, a “Harry” versus “Harriet” comparison could not be made. Looking instead at paired comparisons of other names with both a male and a female version: “Frances/Francesca” and “Frank/Francis/Francisco,” “John” and “Joan/Joanne,” “Michael/Michel” and “Michelle/Michele,” “Patricia and Patrick,” and “Paul” and “Paula/Paulette/Pauline,” ¹⁵ we find that, as a whole, the male names in the set average more yearly citations than their female counterparts (8.3 vs. 5.0), although the results are not quite statistically significant (p = .07). Similarly, the mean citations for Bob (Robert) and Ted (Theodore) are higher than those for Carol (Carole/Caroline) and Alice (10.2 vs. 6.9), but the differences are again not statistically significant (p = .19), because of the very high variances involved. However, we would note that “Jack” has a higher yearly citation count than “Jill” (14.9 vs. 4.6: p = .02), and that “Ken/Kenneth” has a higher citation count than “Barbie/Barbara” (11.9 vs. 5.0: p = .02). ¹⁶ Moreover, if we compare the male and the female halves of each of these nine pairs of names in toto, we do get differences (9.3. vs. 5.7) that are statistically significant (p = .02).

Several researchers have suggested that middle initials or middle names lend gravitas that may be conducive to greater success. Wuffle (1972), in work that has been little noted nor long remembered, provided a useful early contribution to Nameology by demonstrating that those with middle initials fared better in elections to the APSA Council in 1970 than those without—even after controlling for endorsements and gender. Although this long neglected work is what provided inspiration for our current investigation of the importance of middle initials for citation success, there has been much more visible recent work, dealing with vote share in US presidential elections. Wallis (2006), for example, observes that, although only three of our first 17 presidents carried middle names (John Quincy Adams, William Henry Harrison, and James Knox Polk), most modern presidents sported middle names or initials, and Anne Bernays (quoted in Wallis, 2006) has argued that today, “a name without a middle name or middle initial sounds unfinished or unsubstantial, unpresidential.” ¹⁷

To get a handle on the importance of middle initials and names, while trying to control for potential confounds, we looked at the set of those named either Tom/Thomas or Dick/Richard or Harry/Harold in the Masuoka et al. (2007a, b) data set and compared the mean yearly citation counts of members of that set who have initials or middle names with those who have neither initials nor a middle name. ¹⁸ Unfortunately for Wuffle's (1972) hypothesis, we found no real difference (5.6 vs. 6.2, p = .36). ¹⁹

Finally, we examine the letters with which first names begin. It is well known than those with last names toward the end of the alphabet may suffer from lack of recognition. Here, we find a similar result for first names: whereas no single letter seems to stand out as especially lucky, political scientists whose first names begin with the letters U through Z (n =200) do not fare that well in the citations sweepstakes. Those with names beginning A through T have a mean yearly citation count of 6.9, whereas those whose names begin with the last six letters of the alphabet average only 5.1 yearly cites (p = .001). ²⁰

CONCLUSIONS

Our inspiration for this essay was the apparent citation success associated with the name “Gary.” But, like all too many political scientists, we failed to adequately take into account interaction effects, such as those associated with the interdependence of first name and last name. ²¹ On the one hand, we now see the importance of not just being named “Gary” but also being named either “Cox,” “Jacobson,” or “King.” ²² On the other hand, using coarse inexact pairwise matching techniques, we did find that, for large enough samples, men with certain first names do seem to be more cited than women with very similar, or frequently associated, first names. Thus, we see our work as definitely having further advanced the science of Nameology.