CPRE : Consortium for Policy Research in Education

Choosing the school to lead: Exploring distributional preferences of school principals

When applying to school leadership positions principals highlight the skills and knowledge that represent the intersection of their own abilities and the attributes they believe are most valued by employers. This perspective suggests that we can learn about labor supply and demand by examining how applicants represent themselves on the job market. Personal statements are particularly salient as they represent a unique insight into applicants’ philosophy, views, and beliefs.

Correlated topic modeling (CTM) is an automated text mining technique that uses machine learning algorithms to identify the most salient topics across a large corpus of text. We use CTM and leverage our access to the University of Wisconsin–Madison’s super-computing cluster to analyze the personal statements of 2,497 leadership applicants from over 300 of Wisconsin’ 426 districts, constructing a typology of school leaders. Our preliminary results identified seven leadership types, several of which align well with established theories of educational administration (e.g., learning-centered, transformational, relational).

We use the leadership typology generated from the CTM together with factors such as teaching experience, leadership experience, race, and gender to answer the following question:

To what extent is the non-random distribution of school leaders attributable to the preferences of labor supply (i.e., principals) as opposed to labor demand (i.e., schools)?

This study will provide the first description of a large population of school leaders active within the labor market, resulting in novel findings that will directly inform future policy and practice, as to date no literature has captured the total number of topics, the proportion of the applicant pool that writes to specific topics, and how leadership type corresponds to the possible positions they are hired into. As an example, it may be that schools in historically underserved communities are searching for principals who espouse social justice and distributed leadership theories; yet, only a small number of applicants may write to these topics, versus others that write about more general issues such as “all students should succeed.” Knowing this difference could help preparation programs focus instruction for aspiring principals, schools to write more focused position postings, and help leaders who aspire to these positions find the right match to help them and their students succeed.

We see this work as speaking directly to concerns of staffing leadership vacancies, as our analysis will permit us to examine which types of principals apply to which types of schools. We will also be able to examine the types of principals preferred by various school types, conditional on the applicants within the hiring pool.

This work is part of a larger project to better understand the ways in which educators select and are selected by the schools in which they work. Through a deeper understanding of this matchmaking process we will develop policy options to improve the efficiency of the labor market and the equity of the sorting process.

Several other studies are underway, each enhancing insights into educator labor markets by leveraging information provided the application phase of the job search process:

Do different preparation programs yield a competitive advantage? In this study we leverage districts’ revealed preferences for candidates from certain certification programs (conditional on other candidates in the applicant pool) to ascertain the relative effectiveness of educator preparation programs. This research allows us to see how preparation is valued in the labor market. Should we find notable variation in preferences by program, then further inquiry into the structural components of programs is warranted and our findings will provide further direction to engage this line of research.

Is there evidence of discrimination in the education labor market? Gender bias has been suspected of playing a notable role in the school leadership labor market and the role of ethnicity is unclear. While recent research has shown that the majority of school leadership positions are now held by women in comparison with trends prior to 2005, in the absence of application data it is uncertain whether women are less likely be offered positions or less likely to apply. Using our large application database we can analyze hiring patterns conditional on vacancy application while controlling for confounding factors (e.g., experience, education) to identify the extent of potentially discriminatory behavior.

The role of timing in teacher hiring: How early to start? How long to wait? The timing of position postings is a critical aspect of the hiring process, yet research regarding the impact of timing has been limited. Broadly, research has borne out what conventional wisdom would assume to be true - the early bird gets the worm. Timing however is a nuanced game where institutions must constantly weigh the trade-offs between initiating the applicant review process with existing candidates (which may be less than optimal) or extending the job-posting window to attract more candidates (during which time they risk losing desirable candidates within their applicant pool to other institutions). While each institution may have their own intuition and restraints regarding the timing and duration of job postings, a systematic inquiry across all institutions may reveal promising strategies for districts to refine their approach to this challenging problem. 

Project Team: Peter Goff (University of Wisconsin-Madison) and Alex Bowers (Columbia University)

More info? Peter Goff