School bullying and students' academic achievement: Propensity score analysis with PISA 2015

• In Event: Poster session III
  In Poster Session: Poster session 3

Proposal

Bullying is an aggressive, intentional set of actions by an individual or a group of students against other students (s) who possess no ability to defend due to the imbalance of physical, mental, or social power (Farrington, 1993; Olweus, 1993, 1999). Traditionally, bullying was enforced in many forms, such as physical, verbal, relational, and social (Wang, Iannotti, & Nansel, 2009¬). In the school context, bullying has long been a severe problem as it may cause both physical injuries and psychological traumas, and further has negative effects on students’ academic achievement (Craig, Pepler & Atlas, 2000).
School bullying has been studied by global scholars, especially those in the U.S. and Europe. However, in developing countries, such as China, bullying is still under investigation though studies have shown the prevalence of traditional bullying victimization and perpetration (Chan & Wong, 2015). Meanwhile, a series of school bullying incidents were reported through mass media, and people are appealing for policies on school bullying. On the other hand, though school bullying has been catching the attention of public since the 1990s, few policies have been made. In recent years, teachers, parents, and school administrations started to address more concerns. As the support and rationale of policymaking, research studies, however, are still limited. Before 2001, there are few research studies on school bullying. During 2001-2005, research studies increased rapidly year by year (Min, 2016). Even though sufficient empirical links among risk factors of school bullying, and its impact are still under established.
In the current paper, we will explore the association between school bullying victimization and academic performance from a comparative perspective. Results of propensity score analysis from China, US, and Japan are compared in the current study since research studies and policies regarding school bullying in the US have been conducted and enforced widely, and Japan shares a similar culture with China.

Review of Literature

Bullying is a common phenomenon, which is a set of aggressive, intentional actions by an individual or a group against other(s) who is in relatively weak position due to the imbalance of physical, mental, or social power (Farrington, 1993; Olweus, 1993, 1999). School bullying frequently happens in schools with prevalence in all countries and is identified to have serious short- and long-term consequences (Eslea et al., 2003; Craig et al., 2009). Researchers found that students in schools with higher levels of bullying performed worse academically (Austin & Joseph, 1996; Juvonen et al. 2000; Nansel et al., 2001; Schwartz & Gorman, 2003).
Researchers also argue that victimization is associated with academic achievement indirectly, specifically through psychological factors (Juvonen, Nishina, & Graham, 2000; Kochenderfer & Ladd, 1996a; Schwartz & Gorman, 2003; Wentzel, 1991). Victims tend to be disliked by others (Graham & Juvonen, 1998). Being a victim of bullying is found to be associated with a number of adjustment problems, such as depression (Neary & Joseph, 1994), anxiety (Olweus, 1978; Slee, 1994), low self-esteem (Boulton & Smith, 1994), loneliness (Boulton & Underwood, 1992), and school avoidance (Kochenderfer & Ladd, 1996). These peer-relationships and psychological conditions are related to students’ academic performance by affecting their motivations (DeRosier, Kupersmidt, & Patterson, 1994; Muma, 1965). Thus, the psychological problems will be the major part of covariates in predicting the propensity score of the school bullying.
The current study established the hypothesized model for examining whether the school bullying will impact the students’ academic achievement, involving the propensity score analysis to reduce the selection bias.

Data and Methods

The data for this research comes from the Program for International Student Assessment (PISA) 2015. In this research, China, Japan, and the United States were selected as targeted countries. In total, 22,200 students from 643 schools are included in our analysis, in which 6,609 are from Japan, 9,778 are from China and 5,566 are from the United States.

Models and Variables

We propose to apply propensity score analysis to study whether being bullied in school will affect the students’ mathematics, science, and reading achievement in PISA 2015, and then compare the effect of the bullying experience across selected countries. Propensity score (Rosenbaum & Rubin, 1983) is the estimated probability that the subjects would be assigned to the treatment group by accounting for the subjects’ covariates. Based on previous research and the variables in PISA 2015, we choose the following variables as the covariates of both students and their schools to predict whether the student will get bullied in school: social economic status (home possessions, parental education, parental occupation), gender, age, repeated a grade, teacher’s attitude, psychological experience, motivated toward schooling, school possessions. In addition, previous researchers suggested that the survey weights should be included in the propensity score model (DuGoff et al. 2014, Ridgeway et al. 2015), thus we also include the final student weights and replicate weights into the model to predict the propensity scores. In the final outcome model, we incorporate survey design elements to study the treatment effect of the bullying experience. Based on Item Response Theory, PISA calculates five plausible values (PV) for each subject score, thus we would run the propensity score analysis five times, and one set of PVs (math, reading, science) would be used in each time. The final results would be the average of these estimates. Standard deviations would be adjusted by the pooled standard deviations. This research aims to answer the following research questions:
1. What factors are associated with the likelihood of being a victim of bullying?
2. Does school bullying make a difference regarding students’ academic achievement?
3. Does the association between school bullying and students’ academic achievement perform differently in China, Japan, and the United States?

Authors

• Chang Liu, Penn State University

• Yu Hu, Indiana University