Which Characteristics Do Religious School Administrators Value in Teachers? Experimental Evidence from the Global Christian School Sector

Research shows that teacher quality is the most important school-related input correlated with student success. In religious private schools, teachers do not just influence academic outcomes; they may also play a role in spiritual formation. Religious school administrators report that their faith affects their hiring decisions. However, little research has examined the ways that religious school administrators choose individuals to place in teaching positions. In this study, we use conjoint analysis to experimentally examine the hiring preferences of 170 Christian school administrators. We find that having modest (compared to below-average) academic achievement and qualifications or extensive (compared to limited) experience each increases a candidate's likelihood of being hired by 26 percentage points. Having strong (compared to modest) academics, being a graduate of a Christian (compared to secular) postsecondary institution, or being a graduate of the administrator’s K-12 school (compared to a different school) produces smaller effects. We also find some evidence of effect heterogeneity across settings inside and outside the United States.


Introduction
Teachers exert significant influence over their students.During the course of K-12 schooling, the typical family entrusts their child to a school for at least 16,000 hours.Research has shown that teacher quality is one of the most important school-related inputs for student success, both in terms of short-term outcomes such as test scores (Rivkin et al., 2005;Rockoff, 2004;Whitehurst et al., 2013) and behavior (Jackson, 2018) and long-term outcomes such as college enrollment, teenage pregnancy, and adult salaries and wealth (Chetty et al., 2014).Public school administrators may indirectly influence student outcomes by hiring effective teachers (Grissom & Bartanen, 2019a;Grissom & Bartanen, 2019b).However, research suggests that teacher hiring is not a straightforward decision, but involves weighing trade-offs, as different teacher characteristics may signal different strengths (Bruno & Strunk, 2019) and administrators take many contextual factors into consideration when deciding whether a prospective teacher is a good fit for the position (Engel & Cannata, 2015;Harris et al., 2010;Ingle et al., 2011).These tradeoffs may explain why, even when given the opportunity, administrators do not always hire the teachers who will raise their students' academic test scores the most (Bruno & Strunk, 2019).
In religious private schools, teachers not only influence academic outcomes but may also play a role in spiritual formation of students (Johnson & Lee, 2023b, 2023a;Lee et al., 2020;Revell, 2008;Sikkink, 2010Sikkink, , 2012)).Therefore, a prospective teacher's capacity to provide quality spiritual formation is likely a consideration that is important to religious private school administrators in the hiring process, as providing spiritual formation to students may be a key part of their schools' missions (Revell, 2008;Swaner et al., 2019).Because teacher hiring decisions in religious schools likely affects students academically, socioemotionally, and spiritually, it is important to understand these decisions and their implications.
Although a significant body of research has examined public school administrators' dispositions, thought processes, and preferences in terms of hiring teachers (e.g., Engel & Cannata, 2015;Giersch & Dong, 2018;Harris et al., 2010;Ingle et al., 2011), the limited, nascent literature on these topics in religious schools is largely descriptive.There is some qualitative evidence that both Catholic and Jewish private school administrators give strong consideration not only to prospective teachers' professional qualifications, but also their personal character (Tamir, 2021).On surveys, most Christian school administrators and teachers report that their school community's Christian faith has a strong influence on personnel decisions (Johnson & Lee, 2023b, 2023a;Lee et al., 2020), though research on what that finding specifically and practically means has yet to be conducted.
Although Christian schools report that their faith affects their personnel decisions, instructional practices in many Christian schools have become similar to those of public or secular private schools (Cooling et al., 2016;Smith et al., 2021;Smith & Smith, 2011).For example, even though Christian schools are not required to hire certified teachers with degrees in education-and research disputes the claim that teacher certification is an indicator of teacher quality (Chingos & Peterson, 2011;Croninger et al., 2007;Goldhaber, 2019;Harris & Sass, 2011;Pelayo & Brewer, 2010)-85 percent of Christian school teachers in the United States are certified to teach.Forty five percent of Christian school teachers are certified in their state1 , and nearly half majored in education for their undergraduate degree (Johnson & Lee, 2023b).
In this study, we use experimental data from 170 Christian school administrators of Association of Christian Schools International (ACSI) schools to better understand their preferences for teachers.All else equal, what effect do academic qualifications, schooling experiences, and work experience of teacher candidates have on their likelihood of being hired by Christian school administrators to fill teaching positions?In the following sections, we review relevant literature and establish a theoretical framework for our experiment, then describe the experiment design and data, explicate our results, discuss these results, and draw conclusions.

What Makes a Good Teacher?
The body of quantitative literature on teacher quality in the public school sector has grown since 1966, when James Coleman et al. released their "Equality of Education Opportunity" study (colloquially referred to as the "Coleman Report").Coleman et al. (1966) found that teachers are the school-related input most strongly correlated with student learning, especially for disadvantaged students.Since then, many studies have confirmed that teacher quality does indeed significantly affect students' short-term outcomes such as test scores (Rivkin et al., 2005;Rockoff, 2004;Whitehurst et al., 2013) and behavior (Jackson, 2018) and long-term outcomes such as college enrollment, teen pregnancy, and adult salaries and wealth (Chetty et al., 2014).Naturally, researchers next asked the question, "What makes a good teacher good?"Taken as a whole, the results of these studies indicate that observable teacher characteristics inconsistently and weakly explain any variation in teacher quality.Gaining experience appears to increase student gains when an individual begins teaching, although these gains seem to plateau a few years into a teacher's career (Boyd et al., 2006;Buddin & Zamarro, 2009;Pelayo & Brewer, 2010;Rivkin et al., 2005;Staiger & Rockoff, 2010).Additionally, professional knowledge-having taken specific coursework or performing well on a teacher licensure examseems to have a relationship with student outcomes in some contexts, particularly in math and science (Boyd et al., 2007;Clotfelter et al., 2006;Croninger et al., 2007;Harris & Sass, 2011;Pelayo & Brewer, 2010;Phillips, 2010;Shuls & Trivitt, 2015a, 2015b).However, the evidence does not clearly favor teachers having a degree in education, an advanced degree, or certification (Boyd et al., 2007;Croninger et al., 2007;Goldhaber, 2019;Harris & Sass, 2011;Pelayo & Brewer, 2010;Rivkin et al., 2005).

What Makes a Good Christian School Teacher?
Christian schools are not simply designed to replicate secular school outcomes in a faithbased setting.As Swaner et al. (2019) explain, "Christian schools are concerned with academic outcomes, but they are also concerned with the development of the whole student-as one who is made in God's image, created to do His good works…and called to grow as His disciple.This necessitates a focus on holistic learning that includes students' spiritual, ethical, emotional, and physical development, to name but a few.Failure to examine student outcomes in multiple domains would result in failure to capture the fullest picture of flourishing in the Christian school context."This suggests that a concept of (as well as any measurement of) Christian school teacher quality should be multifaceted, attempting to capture this "fullest picture of flourishing."While there is some overlap of goals, Christian education has a different mission, and thus different priorities and desired outcomes, relative to secular education found in public and nonreligious private schools (Maitanmi, 2019).As one of the main inputs into carrying out the unique mission of Christian schools, Christian school teachers' theological views and personal faith and practices affect their dispositions toward and perspectives regarding the content they teach (Mangahas, 2017).Christian school teachers are also, importantly, responsible for a significant portion of students' spiritual formation, which is of eternal value and consequence to them (Fyock, 2008).Since Christian school teachers not only pass on knowledge and skills from a faith-informed perspective, but also create disciples of Jesus Christ, hiring academically and spiritually qualified teachers in Christian schools is crucial to achieving the goals of Christian schooling.
Few studies have examined teacher quality in Christian schools, especially using quantitative methods.Christian schools employ a wide variety of metrics and assessments to track student progress, which can be challenging to compare.However, one researcher found that Christian school teachers having graduate degrees was predictive of the school producing a National Merit Scholarship semifinalist (Clagg, 2011).Another examined which characteristics of teachers were most predictive of being intentional in student spiritual formation.However, the characteristics she found to be predictive are not always easily identifiable on a resume or in the hiring process: "exhibiting a Christ-like attitude," "creating a classroom climate that promotes spiritual growth," and "being intentional in the spiritual disciplines" (Moore, 2014).
In the absence of literature on Christian school teacher quality and under the assumption that Christian school administrators understand what qualities of teachers are associated with desirable spiritual and academic outcomes, the premia school administrators place on specific teacher characteristics could serve as proxies for the efficacy (as pertains to academic and spiritual formation) of teachers with those characteristics.However, if administrators are not formally evaluating teachers based on their students' academic and spiritual outcomes, they may hold misinformed beliefs about which teacher characteristics are most important for students' flourishing.It is important to examine Christian school administrators' preferences to better understand their perceptions of what teacher quality in Christian schools looks like.

Do School Administrators Choose Good Teachers?
Public school administrators have demonstrated high valuation of teachers' care for students, pedagogical skill, and content knowledge, but place a lower value on teaching experience (Harris et al., 2010;Ingle et al., 2011).However, research suggests that teacher hiring is not a straightforward decision, but involves weighing trade-offs, as different teacher characteristics may signal different strengths (Bruno & Strunk, 2019) and school administrators' hiring decisions do not happen in a vacuum; geographic context, grade level, content area, and the characteristics of teachers already employed in a school may affect administrators' decisions (Engel & Cannata, 2015;Harris et al., 2010;Ingle et al., 2011).These findings may explain why, even when given the opportunity, administrators do not always hire the teachers who will raise their students' academic test scores the most (Bruno & Strunk, 2019).Boyd-Swan and Herbst (2018) performed a resume audit project in early childhood education (ECE) settings, which share some similarities to Christian schools in that some are privately operated and all are released from teacher licensure requirements.They found that a resume with a mid-range grade point average was more likely to garner an interview request relative to a lower GPA; however, a resume with a higher GPA was no more likely to garner an interview request relative to a 2.8 GPA.They also found that resumes with limited ECE experience (six months), relative to resumes with no experience, were more likely to garner an interview request, but resumes with two years of experience were no more likely to garner an interview request than a resume with no experience.

Conjoint Experiments and Teacher Hiring
A conjoint experiment is a research design in which survey respondents choose a preferred option from a set of options where each option possesses randomly assigned characteristics.Since the characteristics of each option are randomly assigned and all else is assumed to be equal, we can estimate how much each characteristic causally affects the likelihood that an option is selected.This type of experimental design has frequently been used in the fields of marketing (Chrzan & Orme, 2000;Green & Rao, 1971) and political science (Abramson et al., 2022;Hainmueller et al., 2014;Leeper et al., 2020), allowing researchers to examine each facet of a complex, multidimensional candidate, policy, product, or service individually and examine consumers' or voters' favorability toward each.For instance, in political science research, survey respondents may be presented with a set of candidates each with randomly assigned characteristics such as gender or political affiliation to estimate how each characteristic causally affects voters' preferences (Abramson et al., 2022).
In their pioneering study, Giersch and Dong (2018) applied a conjoint analysis to educational research, studying public school administrators' favorability toward teacher candidate characteristics.They found a preference for teachers with some experience (though there was no difference between five and fifteen years of experience, which is consistent with research on teacher quality) and with more education (although there was no difference between less education from a selective institution and more education from a nonselective institution).
We extend this line of scholarly inquiry by conducting a conjoint experiment to understand administrators' hiring preferences in religious schools.

Experiment Design
In the fall of 2022, the Association of Christian Schools International (ACSI) administered a survey to school administrators in member schools in Cambodia, Nigeria, the Philippines, the United Republic of Tanzania, and the United States of America.In total, 189 administrators completed the survey and 170 provided demographic information: the respondents' job title, school characteristics, sex, age, race and ethnicity, marital status, educational background, and teaching and leadership experience.As part of the survey, the administrators participated in a conjoint experiment designed to capture information about which teacher characteristics administrators value most.
Each of 170 administrators participating in the survey was assigned four sets of three fictitious teacher candidates in each group.For each set, which administrators assessed one at a time, the administrator was asked, "Consider the three applicants below.Which candidate would you prefer to hire for your school as a teacher?"The candidates in each set randomly varied across the following four characteristics (see Table 1): academic achievement and qualifications, whether they were a graduate of the administrator's school, whether they had graduated from a Christian or secular postsecondary institution, and whether they had limited or extensive experience.By the end of the experiment, school administrators had chosen to hire four fictitious teachers, one from each of the four sets.characteristics that would plausibly appear on a resume or teaching application and be related, in perception or in actuality, to a teacher's ability to provide quality academic and spiritual formation.Research supports the idea that academic achievement and qualifications may be related to teacher quality under some conditions (Boyd et al., 2007;Croninger et al., 2007;Goldhaber, 2019;Harris & Sass, 2011;Pelayo & Brewer, 2010;Rivkin et al., 2005), as well as teacher experience in the early years (Boyd et al., 2007;Clotfelter et al., 2006;Croninger et al., 2007;Harris & Sass, 2011;Pelayo & Brewer, 2010;Phillips, 2010;Shuls & Trivitt, 2015a, 2015b).Therefore, we hypothesize that a teacher with above-average academic achievement and qualifications would be a higher quality teacher than one with below-average academics or even modest academics, and that a teacher with extensive experience would be a higher quality teacher than one with limited experience.
Since Christian schools also value teachers' spiritual leadership and spiritual formation capabilities, we assume that having received an education in a Christian setting (i.e., a Christian K-12 school and/or a Christian postsecondary institution), could serve as a proxy for a teacher's ability to spiritually form their students.We recognize that many Christians attend secular schools or postsecondary institutions and receive spiritual formation during their career preparation years through their church, campus ministries, or interpersonal relationships (Johnson & Lee, 2023b), and therefore having a Christian education may be a noisy proxy for spiritual leadership.However, research has shown that teachers in Christian schools who did not attend Christian universities may feel underprepared to contribute toward the unique missions of Christian schools (Cooling et al., 2016).Mitchell (1982, p. 6) notes that Christian educator preparation programs "should offer education courses that (a) provide the required distinctive content and approach; (b) allow for student teaching experiences at Christian schools; and (c) furnish information on topics deemed essential, such as Biblical integration procedures, personal commitment, classroom discipline, and purposes of Christian schools." We additionally assigned information about each fictitious candidate's primary and secondary schooling background.In particular, each was either a graduate of the survey participant's school or not.We narrowed this variable for primary and secondary schooling to the administrator's school, not just a Christian school, because this signals that the administrator would likely have personal knowledge of the teacher candidate's spiritual leadership and overall character, and the teacher candidate would have a personal knowledge of the school culture, theoretically making that teacher's transition into teaching at the school easier (a homegrown effect).Therefore, we hypothesize that being a graduate of a Christian college or university and a graduate of the administrator's school will increase the likelihood of hire.

Data Sample
The sample of 170 school administrators for which we have descriptive data is broadly representative of ACSI in terms of most school administrator characteristics.As the largest Christian school organization in the world, ACSI represents a large portion of the Christian school sector.Therefore, we have reason to believe our results may be applicable to the Christian school sector at large.Nineteen percent of school administrators reported to be a head of school, 33 percent reported to be a campus, upper-level, or lower-level principal, and 48 percent reported to be in another leadership position such as an assistant or vice principal, a dean of students or academics, or a school chaplain (see Table 2).Forty-two percent of respondents reported to be male and 58 percent reported to be female.Ninety-five percent reported to be white, with the remaining 5 percent identifying as Black, Hispanic, Asian, Native American, or some other ethnicity.The mean age was reported to be 49 years and the median age 50 years, with the sample ranging from age 22 to 73 years.Ninety-three percent of respondents reported being married.
With respect to administrator educational attainment and prior experience, 27 percent of survey respondents reported their highest degree was a bachelor's degree, 54 percent a master's degree, and 13 percent a specialist or doctoral degree.Survey respondents reported a mean of 10 years of teaching experience before becoming a school administrator, with a median of nine years and respondents ranging from no teaching experience to 33 years of teaching experience.
Similarly, they reported a mean of 10 years of school administration experience, with a median of six years and a range from one to 44 years of experience.As for years in their current role, survey respondents reported a mean of six years, with a median of three years and a range of one to 34 years in that role.
Finally, with respect to administrator school characteristics, the reported mean school size in terms of enrollment is 741 students, with a median of 602 students and a range of 117 to 1,671 Note: While 189 administrators participated in the conjoint experiment section on the survey, only 170 participated in the demographic data collection section.

Analytic Strategy
We assume an administrator  will hire candidate  from among  = 1, … ,  options based on an unobserved utility model, which is a function of teacher candidate 's academic qualifications (  ) and experience (  ), as well as whether the candidate is a graduate of the administrator's school (  ) and a graduate of a Christian higher education institution (  ).We assume the administrator's choice will maximize utility within a set.
Following Hainmueller et al. (2014), since the qualities of the fictitious teachers within the sets in the experiment were randomly assigned, we can nonparametrically estimate the causal effect of each teacher characteristic on likelihood of being hired by administrators in our experiment.We do so using the Conjoint package in Stata,3 employing the following model: In Equation 1, ℎ  is a binary variable that equals one if administrator a chooses to hire teacher t in set s.This dependent variable is expressed as a function of teacher t's characteristics.  ′ is a vector of teacher t's academic qualifications and includes a dummy variable for whether the candidate has below average academic qualifications or strong academic qualifications, with modest academic qualifications as the omitted category.  takes a value of 1 if the candidate t is a graduate of the administrator's school and 0 otherwise.Similarly,   takes a value of 1 if candidate t is a graduate of a Christian college or university and 0 otherwise.
represents candidate t's experience and takes a value of 1 if the candidate has extensive experience and 0 otherwise.Lastly,   is the conventional error term.Because each respondent indicated preferences for four different choice sets of teachers, we cluster the standard errors at the respondent level to correct for the non-independence of observations in our data that originate from the same respondent.Because each characteristic is randomly assigned,   ,  2 ,  3 , and  4 capture the causal effect of each component on the likelihood a candidate will be hired.
In the following section, we present the estimated -coefficients or average marginal component effects, which can be interpreted as the effect that a particular teacher characteristic has on the change in likelihood that teacher t is hired.

Overall Results
We begin by presenting the average marginal component effects for the entire sample of school administrators for which we have both experimental and descriptive data (n = 170) in Table 3 and Figure 1.These results are robust to expanding the sample to include all 189 administrators who participated in the conjoint experiment (see Appendix A).We find that relative to a teacher with modest academic achievement and qualifications, administrators are 26 percentage points less likely to hire a teacher with lower-than-average and 19 percentage points more likely to hire a teacher with higher-than-average academic achievement and qualifications, all else equal.We also find that administrators are 13 percentage points more likely to hire a teacher who is a graduate of their school, relative to a teacher who graduated from a different school.Administrators are 24 percentage points more likely to hire a teacher who graduated from a Christian college or university, relative to a teacher who attended a secular college or university.Finally, school administrators are 26 percentage points more likely to hire a teacher who has extensive teaching or leadership experience, relative to a teacher who has limited teaching or leadership experience (all p < 0.001).

Results by Administrator Characteristics
Next, we examine whether administrator preferences differ by subgroup, requiring us to examine only administrators for whom we have descriptive data.To do this, we estimate the marginal means (the probability of being hired when a characteristic is held fixed, with all other teacher characteristics held equal) for each possible teacher characteristic for administrators within each subgroup.We then compare the confidence intervals (set at the 95 percent confidence level) for these means, examining whether they overlap; non-overlapping confidence intervals can be interpreted as significant differences.Because variation in cultural contexts may However, because Christian and other religions schools do not exist to simply replicate the results of secular public schools, it is important to understand the value Christian school leaders place upon possible indicators of teachers' spiritual leadership.Christian school administrators do prefer to hire teachers who graduated from their K-12 school, relative to a different one (by 13 percentage points), and who graduated from a Christian postsecondary institution, relative to a secular one (by 24 percentage points).However, we do see evidence that school administrators outside the U.S. are less likely to hire teacher candidates based on the religious character of their postsecondary schooling institution.We suspect that this finding may be due to these schools' lack of proximity to Christian colleges and universities.For example, only 22 percent of member schools in the Council for Christian Colleges and Universities, an international organization, are located outside of the United States. 4ur research has a few limitations.First, administrators may conceptualize "belowaverage," "modest," and "strong academic achievement and qualifications" differently.
Researchers themselves have conceptualized teacher academics differently-for example, college GPA, teacher certification exam scores, and self-reported content knowledge (Burroughs et al., 2019;Pelayo & Brewer, 2010;Wayne & Youngs, 2003).It is not clear which measures of academic qualifications and achievement the administrators in our sample value; however, it may be important to word this variable in such a way that maintains a broad interpretation because schools across different countries operate within different educational systems.Likewise, the definition of the experience variable is not very specific and may be interpreted differently by each school administrator: "limited" versus "extensive."Additionally, the experience may not necessarily be teaching experience, as we defined experience as being in the fields of "teaching or leadership." In future research, we would recommend more specificity of definitions for teacher characteristics so that we can better ascertain what it is, exactly, about a teacher candidate's

Figure 1 :
Figure 1: Average Marginal Component Effects for Teacher Characteristics affect administrator preferences, we first examine marginal means between administrators in the United States versus schools outside the United States.Although the confidence intervals for the marginal means for administrators outside the United States are quite large because of the smaller sample size (n = 17), there is very little overlap of confidence intervals for the postsecondary variables, suggesting a nearly-significant difference.Whereas a Christian postsecondary graduate has a 45 percent chance, a secular postsecondary graduate has only a 22 percent chance of being hired by an administrator in a school in the United States, a difference of 23 percentage points.However, these probabilities are closer together for school administrators outside the United States; a Christian postsecondary graduate has a 38 percent chance of being hired, while a secular postsecondary graduate has a 28 percent chance of being hired, a difference of 10 percentage points.Additionally, while the confidence intervals for U.S. administrators preferring secular and Christian postsecondary education do not overlap with one another, suggesting a clear preference for graduates of Christian postsecondary institutions, the confidence intervals for school administrators outside the United States preferring secular and Christian postsecondary education overlap, suggesting no clear preference for alumni of7.0 DiscussionIn this study, we examined Christian school administrators' preferences toward certain characteristics of teacher candidates being considered for employment at the administrators' schools.Using experimental data from a conjoint analysis, we found that having extensive (compared to limited) teaching and/or leadership experience and modest (compared to belowaverage) academic achievements and qualifications increased the likelihood of a prospective teacher being hired, by about 26 percentage points for both characteristics.Having strong (compared to modest) academic qualifications increased the likelihood of being hired by about 19 percentage points.While teacher quality literature is inconclusive on nearly every predictor of teacher quality, teacher experience and academic achievement are most consistently shown to be somewhat important for student success across a large body of studies; our findings seem to show that school administrators are thinking in terms of academic effectiveness.Our findings also seem consistent withGiersch and Dong (2019), who found that public school administrators value academic quality and work experience more than other teacher characteristics.
experience that is attractive to school administrators.Future research should also qualitatively examine how Christian school administrators value teachers to allow researchers and leaders in Christian education to better understand what variables school administrators consider to be proxies for spiritual leadership.Additionally, researchers should investigate the extent to which teacher preparation in Christian colleges and universities differ from other colleges and universities.How different are the programmatic elements and do those differences have an effect on teacher quality, especially with respect to the capacity to form students spiritually?If spiritual formation is a central goal of Christian schools, improving the effectiveness of teachers-whether by pre-service training, professional development, or better screening during the hiring process-to attain this goal is paramount.

Figure B. 2 :
Figure B.2: Marginal Means by Administrator's Highest Degree

Figure B. 3 :
Figure B.3: Marginal Means by Administrator Degree Area

Figure B. 4 :
Figure B.4: Marginal Means by Administrator's Teaching Experience

Table 1 : Possible Teacher Attributes in the Conjoint Experiment
We chose the teacher characteristics listed in Table1because they are all teacher

Table 2 :
students.This sample is skewed toward larger schools; the average ACSI member school's size is 200 students. 2This skew toward larger schools is also reflected in the survey respondents' reported school budgets, where the mean is $6.3 million, the median is $5.3 million, and the range is from $1.0 million to $13.4 million in U.S. dollars.Ten percent of survey respondents reported working at a school outside of the United States.Sample 2 Statistic from NCES' Private School Universe Survey: https://nces.ed.gov/pubs2021/2021061.pdf

Table 3 : Average Marginal Component Effects for Teacher Characteristics
Note: ** p<.001.Std.errors clustered by survey respondent and set.For academic achievement/qualifications, "modest" academic quality is the omitted category."Extensive teaching/leadership experience" is a binary variable where the alternative is "limited experience."