This research explores the circumstances under which tenure protects incompetent faculty.Footnote 1 Incompetent faculty are those who fail to meet the teaching, research, and service expectations at their institution. Systematic research by Katz (Reference Katz1973), Kawar (Reference Kawar1983), and Park and Riggs (Reference Park and Riggs1993) indicates that these were the criteria used for assessing faculty performance at every institution investigated, and an examination of American Association of University Professors (AAUP) documents reveals that these have been the accepted standards for evaluating faculty for nearly a century (AAUP 1915; 2009).Footnote 2 As many readers know, the possible connection between tenure and incompetence is among the most contentious issues in higher education. Controversies associated with this subject have prompted many institutions to tighten the procedures and standards they use in making tenure decisions, to implement posttenure reviews, and to consider time limits on grants of tenure (see Goodman Reference Goodman1994; Kelley Reference Kelley2000; Turner Reference Turner1997; Whicker Reference Whicker1997; Wood and Des Jarlais Reference Wood and Des Jarlais2006).Footnote 3
Although changing the system to try to prevent tenure from shielding nonperforming faculty could affect faculty careers and the freedom and creativity that are at the heart of academia, very little research has focused on the relationship between tenure and incompetence. Instead, the scholarly discussion of the problem has consisted of speculation regarding why some tenured faculty may not perform as expected. For example, in an early examination of the subject, Machlup (Reference Machlup1964, 115–16) argued that the explanation may reside in an improperly conducted tenure review process, the failure of institutional leaders to maintain a challenging academic environment, the personality traits of individual faculty, or characteristics specific to certain types of institutions. More recent discussions also point to these possibilities, with Lewis (Reference Lewis1980, 87), Taylor (Reference Taylor2010, 204–09), and Wood and Des Jarlais (Reference Wood and Des Jarlais2006, 563) describing faculty personalities and age as potential explanations; Whicker (Reference Whicker1997, 22–23) asserting that institutional leadership plays a key role; and Brown and Kurland (Reference Brown and Kurland1990, 331) and Pilant and Ellison (Reference Pilant and Ellison1997, 17) maintaining that tenure decision procedures and standards account for the problem.
The research to date has done little to clarify the matter, for it has dealt with such matters as how posttenure reviews are structured (Goodman Reference Goodman1994; Wood and Des Jarlais Reference Wood and Des Jarlais2006); the use of external review letters in tenure cases (Schlozman Reference Schlozman1998); whether the materials used in tenure reviews can predict successful academic careers (Lewis Reference Lewis1980); how well peer reviews, student evaluations, and portfolios measure teaching (Boyer Reference Boyer1990; Algozzine et al. Reference Algozzine, Beattie, Bray, Flowers, Gretes, Howley, Mohanty and Spooner2004; Kelly-Woessner and Woessner Reference Kelly-Woessner and Woessner2006; Kohut, Burnap, and Yon Reference Kohut, Burnap and Yon2007; Langbein Reference Langbein1994; Yon, Burnap, and Kohut Reference Yon, Burnap and Kohut2002); the standards differing types of departments use in tenure decisions (Rothgeb and Burger Reference Rothgeb and Burger2009); and what variables are associated with the denial of tenure and with promotions in political science departments (Hesli, Lee, and Mitchell Reference Hesli, Lee and Mitchell2012; Marshall and Rothgeb Reference Marshall and Rothgeb2011). This absence of research regarding when tenure protects the incompetent leaves a serious gap in our understanding of an important professional issue. It also prompts one to wonder how effective any tenure reforms can be if they are based more on conjecture than on empirical analysis.
This essay addresses this need for research by systematically investigating how an institution's characteristics and the procedures and standards used to assess tenure applications are related to when tenure shields incompetent faculty.Footnote 4 The following text describe how the data were obtained and analyzed.
THE DATA
This study uses data from a mail survey sent in February 2010 to 1,248 chairs of political science departments in the United States that offer bachelors, masters, and doctoral degrees.Footnote 5 The APSA provided the names and addresses.Footnote 6 Faulty addresses led to 36 returns for an effective population of 1,212. Responses were received from 361 chairs (58 doctoral departments, 77 offering a masters degree, and 226 with a bachelors program), yielding a response rate of 29.8 %.Footnote 7
Information for the dependent variable, tenure protects incompetent faculty, was acquired by asking the chairs to respond yes (coded 1) or no (coded 0) to the following statement: “At my college/university, tenure has shielded incompetent faculty from dismissal.” Among the respondents, 61.5% answered yes, while 38.5% said no.Footnote 8 This procedure yielded a measurement indicating whether the problem of tenure protecting incompetence existed at the responding institution. As Euben (Reference Euben2002) illustrates, more detailed data pertaining to the magnitude of the incompetence problem from campus to campus requires the examination of data that often are afforded legal protection and/or that institutions hesitate to release. Acquiring such information, when it is available, probably is best pursued by means of campus-specific case studies. Still, the current measure should provide a useful starting point for the long-overdue analysis of an important tenure-related issue.Footnote 9
The first of the independent variables pertained to institutional characteristics. These included total enrollment (coded 1 for fewer than 5,000 students, 2 for 5,000–10,000, 3 for 10,001–20,000, and 4 for more than 20,000), location (rural = 1, other = 0), whether the institution was public (coded 1) or private (coded 0), and unionized (coded 1) or nonunionized (coded 0). The public/private and union/nonunion variables were included to assess assertions from the ongoing national debate regarding how these variables affect education. Privatization frequently is depicted as forcing institutions to use resources carefully because they receive less government support, making them more careful about granting tenure and less likely to have nonperforming tenured faculty on their staffs.Footnote 10 For their part, unions sometimes are described as protecting all employees regardless of performance, thus increasing the probability that unionized campuses will have incompetent tenured faculty members.Footnote 11
As noted above, additional independent variables relating to the procedures and standards used during tenure reviews were analyzed. The procedural items comprised whether the chairs reported that collegiality is an important factor in tenure decisions at their college/university (coded yes = 1, no = 0), whether the institution required external review letters in tenure cases (yes = 1, no = 0), whether the department had set standards to guide tenure votes (yes = 1, no = 0), and whether a positive departmental tenure decision has been reversed by higher authorities (yes = 1, no = 0). The letters and standards variables were included to assess how much an institution attempted to reduce the roles of personality and institutional social connections in tenure decisions and to ensure that tenure candidates' records conform to discipline-wide views of excellence. The collegiality variable tends to tap the opposite impulse, reflecting the willingness to include in tenure decisions someone's ability to work as part of a professional team. Finally, reversal by higher authorities indicates how actively an institution's top administrators participate in tenure decision making.Footnote 12
The standards variables pertained to research, teaching, and service. Six research variables were analyzed. Three related to publishing requirements to receive tenure: (1) the number of articles or the equivalent tenure candidates were expected to publish (no set requirement = 0, one article = 1, two or more = 2, one per year = 3, two or more per year = 4), (2) the number of articles tenure candidates were expected to publish in the most prestigious journals in their specific field (no set requirement = 0, one = 1, two or more = 2, one per year = 3, two or more per year = 4), and (3) the number of books tenure candidates were expected to publish (no set requirement = 0, one = 1, two or more = 2). In addition, chairs were asked whether single-authored publications were essential for tenure (yes = 1, no = 0), whether teaching publications were equal to substantive publications (yes = 1, no = 0), and whether research is the most important factor in the institution's tenure decisions (yes = 1, no = 0).Footnote 13 Marshall and Rothgeb (Reference Marshall and Rothgeb2011, 576) and Rothgeb and Burger (Reference Rothgeb and Burger2009, 517) used the first five of these variables to assess research requirements in tenure cases. The sixth variable was included to control for the emphasis placed on research in tenure decisions.
The teaching variables included several of the items Rothgeb and Burger (Reference Rothgeb and Burger2009, 516) report are used to assess teaching during tenure reviews. These were (1) student evaluations (yes, teaching evaluations are required = 1, no = 0), (2) portfolios (yes = 1, no = 0), (3) peer reviews by a faculty member or administrator (yes = 1, no = 0), (4) whether tenure candidates are expected to create new courses (yes = 1, no = 0), and (5) whether the candidate teaches courses required by the department or the college/university (yes = 1, no = 0). In addition, the chairs were asked to state whether teaching was the most important factor in tenure decisions at their institution (yes = 1, no = 0). Student evaluations, portfolios, and peer reviews tap the quality of an applicant's teaching, whereas teaching required courses and setting up new courses pertain to the candidate's curricular contributions. The teaching importance variable controlled for the emphasis placed on teaching during tenure reviews.
The last group of variables was for service. These included (1) whether tenure applicants were required to serve on at least one departmental committee (yes = 1, no = 0), (2) whether candidates were required to work with student clubs (yes = 1, no = 0), (3) whether a strong commitment to advising was required (yes = 1, no = 0), and (4) whether community service was expected (yes = 1, no = 0).Footnote 14
Logit regression analysis was used to assess how each cluster of independent variables affected incompetence. Logit is a regression technique that assesses how an independent variable affects a dichotomous dependent variable while controlling for the remaining variables in the analysis.Footnote 15 The model used to examine the institutional variables provides an example of the basic equation:
$$
{\rm Incompetence} = {\rm a} + {\rm b}_{\rm 1} \,{\rm Public} + {\rm b}_{\rm 2} \,{\rm Location} + {\rm b}_{\rm 3} \,{\rm Unionized} + {\rm b}_{\rm 4} \,{\rm Enrollment} + {\rm e}
$$
The same type of equation was used to evaluate each of the other variable clusters.
Before turning to the results, note that because colleges and universities have differing missions, a series of parallel analyses were conducted that controlled for such variables as the emphasis an institution places on research and teaching when making tenure decisions, whether attempts have been made to limit tenure at the institution, and the size of and the highest degree offered by the responding department. To save space, these additional results are not reported here, but in no case did these controls affect the results that are reported.Footnote 16
THE RESULTS
Table 1 has the findings for institutional characteristics. As can be seen, none of the variables in this cluster affects the dependent variable, incompetence.Footnote 17 This casts doubt on the claims some make regarding the effects of unionization and privatization on higher education.
Table 1 Institutional Characteristics and Incompetence
* p < .05,
** p < .01,
*** p < .005
The procedural results are in table 2. Two variables from this group affect incompetence, for collegiality is positively related, whereas reversal has a negative effect.Footnote 18 Indeed, these findings suggest that when collegiality is an important consideration in a tenure case, the probability of incompetent faculty increases by .12, and that when higher authorities have reversed a positive department recommendation, the probability of incompetence decreases by .30. Neither of the remaining variables in this cluster affects incompetence, which is especially notable as far as outside letters are concerned because many institutions emphasize this evaluative device (see Schlozman Reference Schlozman1998).
Table 2 Tenure Evaluation Procedures and Incompetence
* p < .05,
** p < .01,
*** p < .005
Table 3 has the research findings. Two variables stand out: the number of expected articles or chapters display a positive relationship, and the number of articles in prestigious journals reveal a negative association. The remaining variables were not related to incompetence.Footnote 19 The positive relationship for the number of articles variable implies that a one-unit increase in the number of articles demanded from tenure candidates (that is, from one to two or more, or from two or more to one per year) increases the probability of incompetence by about .08, and the negative effect for the prestigious article variable suggests that a one-unit increase in the value of this variable decreases the probability of incompetence by .16. In other words, chairs from departments that focus on the quantity of publications appear more likely to report that tenure shields incompetence, whereas chairs of departments that emphasize quality journal articles have a greater tendency to state that tenure does not protect the incompetent.
Table 3 Research and Publishing and Incompetence
* p < .05,
** p < .01,
*** p < .005
The results for teaching and service are shown in tables 4 and 5. As table 4 shows, neither the teaching quality nor the curricular contribution variables are related to incompetence.Footnote 20 Table 5 reveals that the service variables also have no apparent association with incompetence.Footnote 21
Table 4 Teaching and Incompetence
* p < .05,
** p < .01,
*** p < .005
Table 5 Service and Incompetence
* p < .05,
** p < .01,
*** p < .005
DISCUSSION
Before discussing conclusions, some caveats are in order. One is that the results here reflect relationships, not causal direction. For example, the absence of an association between outside letters and incompetence may be because some institutions adopt letters only after they recognize that they confront the problem of incompetence, while others adopt letters earlier. Another caveat pertains to the need for caution with the research results because in-depth evaluations of their application in differing teaching and research contexts are needed.Footnote 22 One can, however, still glean useful insights from the current findings.
One insight comes from the institutional results indicating that the problem of tenure as a shield for the incompetent affects colleges and universities of all types. Whether large, medium, or small, unionized or nonunionized, rural or urban, or public or private, department chairs report that their colleges and universities confront problems stemming from tenure as a protector of the incompetent. These findings also show that much of the discussion in the media about the effects of unionization and privatization is misplaced, at least as far as higher education is concerned.Footnote 23
A second insight is that some procedures do weed out the incompetent, whereas others do not. Reducing the role of collegiality in tenure decisions and demanding that higher level administrators carefully monitor tenure recommendations may help limit incompetence. Requiring evaluation letters from external reviewers and setting standards to guide tenure votes, however, appear less useful. Indeed, when one considers the many logistical and other problems Schlozman (Reference Schlozman1998) states accompany letters and the absence of an association between letters and the denial of tenure that Marshall and Rothgeb (Reference Marshall and Rothgeb2011, 574) report, one wonders whether this form of evaluation is worth the effort.
A third insight comes from the research results: quality matters. The findings herein suggest that colleges or universities that require candidates to publish in the top journals in their research field to receive tenure are less likely to regret tenure decisions over time. When the focus is on quantity, however, the chances seem to grow that future faculty performance may not match expectations. Juxtaposing this with Marshall and Rothgeb's (Reference Marshall and Rothgeb2011, 574) finding that the number of articles, not their location in prestigious journals, is a key variable in tenure decisions implies that departments and institutions that emphasize research and publishing when granting tenure may need to reconsider the way they evaluate applicants' records.
To close, this research is an initial exploration of a complex and controversial subject. To better understand the possible connection between tenure and incompetence, further inquiry is required. For instance, case studies and/or in-depth comparative analysis should seek the data needed to develop measures of the magnitude of the problem from campus to campus and to study the effects of such variables as personality, age, and institutional leadership. In addition, future analysis should examine the role of collegiality, how and why deans, provosts, and other officials and committees reverse positive departmental recommendations, how the quality and quantity of publications interact during tenure decisions, and how teaching and service expectations relate to the evaluation of faculty competence at differing institutions.
