Looking for Their Dick Vermeil: How Prior Performance Influences the Selection Decision

This question has been addressed briefly in the literature on non-profits. In the non-profit literature they do not seem to be dealing with insider and outsider empirically, but there is case study evidence supporting founders of organizations generally choose their boards and are able to choose their successors (Santora and Sarros 2001). This lends to the idea that founding individuals want to insure their legacy and do so by picking people who have similar beliefs. There is no clear evidence of whether or not these individuals will be insiders or outsiders. Founding individuals have a unique advantage in the non-profit and private sectors. These founding members are often able to pick their successors and therefore have a disruptive free succession (Santora and Sarros 2001). The closest linkage to this is in the public sector could be the interim hiring of a deputy manager when a current manager is planning for retirement.

There are a few cases where school superintendents have come from outside public education, primarily in large urban school districts. John C. Fryer, Jr., the current superintendent of Duval County Florida, is a former Air Force General. Alan D. Bersin, superintendent in San Diego, is a former U.S. Attorney. Former Colorado Governor Roy Romer served as the superintendent in Los Angeles from 2000–2006. While there are several other cases, the number is not large. These of course would be cases where the districts went way outside the organization when selecting the successor. We do not identify these cases in this analysis, however, they would be an interesting avenue for future research.

The TAAS was replaced by the Texas Assessment of Knowledge and Skills (TAKS) in 2003.

The independent variables in both models are measured in the year prior to the hiring of a new superintendent.

We start in 1993 to insure that we have the performance data for the previous years in each district where a superintendent was replaced.

We were concerned that the 490 cases of hiring in our data could be related to performance in the school district. This would mean that low performing districts would be more likely to make it into our data and create selection bias. To test this we ran a model for all the school districts to see if performance was related to a hiring decision. We found that average performance was negatively and insignificantly correlated (−0.22) with the decision to hire a new superintendent.

To check for multicollinearity in the model we examined the bivariate correlations and the variance inflation factors and no problems were detected. We were particularly concerned with the correlation between the all pass rate and class attendance. The variables are correlated at only 0.25 and attendance has a VIF of only 1.59.