A Multilevel Analysis of Student Perceptions of School Climate: The Effect of Social and Academic Risk Factors.

Overview 
School climate can influence students’ experiences in school as well as their academic performance, so investigating which factors contribute to a positive school environment can help explain variation in student success.  According to Fan, Williams, and Corkin, most empirical studies of school climate focus solely on how school characteristics are related to school climate, relying on the assumption that the organization and structure of a school necessarily influences its culture.  However, focusing only on school-level characteristics, to the exclusion of the characteristics of students themselves, relies on the (likely incorrect) assumption that all students within a school experience its climate in the same way.  Previous studies using multilevel models fail to shed light on the link between the student and school levels because they do not account for the possibility that individual factors could influence perceived climate at the school level, nor do they include individual risk factors within family or academic settings.  Therefore, Fan, Williams, and Corkin examine a larger range of school and student-level variables in order to better understand how theses factors influence students’ perceptions of school climate.

 

About the Study
Data from the Educational Longitudinal Study of 2002 allow the authors to examine variation in students’ perception of school climate given ten individual-level variables – race, gender, family structure, family income, parent's education, parent's nationality, siblings dropped out of school, changed schools, repeated a grade, and behavior problems at school – in addition to three school-level variables – enrollment, subsidized meals, and school sector (public, Catholic, and non-Catholic private).  Students' climate perceptions are measured along three separate axes: order, safety and discipline; fairness and clarity of school rules; and teacher-student relationships.  The sample is nationally representative and includes 16,168 tenth graders from 757 different high schools.

 

Conclusions
Fan, Williams, and Corkin find that the majority of variation in perceived school climate (over 80% in each category) is accounted for using data from the individual level.  They also find that the associations at the individual level are independent of any school level factors.  For example, being black, male, or lower class is negatively correlated with perceived climate regardless of school characteristics. The remaining variation in perceived climate is explained through school-level characteristics.  Of the school level predictors, the authors find that attending a private institution has the strongest positive influence all three measures of perceived climate.

The relatively small influence of school-level variables in comparison with individual-level variables illustrates the weakness in previous research exclusively engaged with school-level data since the majority of variation in perceptions of student climate was unaccounted for due to of the omission of individual-level variables.  While the individual-level variables explained four times the variation, school sector differences independently explained some fluctuations in climate perceptions, highlighting the importance of including both levels of analysis in the investigation on students’ experiences in school.  

 

Evaluation
This paper uses a multilevel analysis to fill the analytic holes found in previous research, yet the study raises a number of methodological questions itself.  The authors use 10 individual characteristics and only three school-level characteristics to predict perceptions of school climate.  The individual-level analysis covers a broad range of concepts that are generally independent of each other.  However, the negative effect of having a limited number of school factors is exacerbated by the variables’ inter-dependence.  For example, private schools generally have lower enrollment and a lower proportion of students receiving subsidized meals than public schools.  Therefore, the school-level variables that were included in the analysis likely all had similar influence on students’ perceptions of school climate.  Including so few school-level variables, and in particular by including those variables likely to explain the same type of variation in students’ perceptions of school climate, severely limits the ability of the school-level portion of the analysis to contribute much variation in the outcome.  The authors were likely limited by the data at hand, but future analyses that incorporate other school-level variables such as neighborhood socioeconomic status, teacher experience, and average class sizes will likely find that school-level factors explain more variation in school climate perceptions, and better capture how school-level factors impact perceived school climate.

Keywords: Behavior, Student Attitudes, Educational Environment, Multilevel Analysis

Sector: Catholic,  Private,  Public

Outcomes: Academic,  Morality and Character Formation,  Peers and Social Networks

Date Posted: 2015-11-18