The Reference Group Effect

Abstract

Academic self-concept and achievement are positively related among students from the same school and country. Yet negative associations between these variables may be found at the level of schools and countries. In the present article, we propose how this apparent paradox can be explained in terms of reference group effects, in which high standards, norms, or benchmarks act to decrease academic self-concept, whereas low ones have the contrary effect. Multilevel regression analyses of the PISA 2006 data consisting of 353,403 students, 13,886 schools from 53 countries revealed an interesting pattern of relationships. Within schools, students’ science achievement and science self-concept were positively related although the size of this relationship varied considerably between countries, whereas between schools and between countries, this association was negative. Consistent with our hypothesis, the size of the between-school relationship was larger in countries with a higher percentage of selective schools. At the country level, the negative relationship between country mean achievement and self-concept was explained by country differences in educational benchmarks, standards, and norms next to country differences in response styles. In this article, we also discuss the implications for the validity of cross-cultural comparisons of self-concept.

Cross-cultural comparability of noncognitive constructs in TIMSS and PISA

“…A classic example of lack of invariance in self-report scales is the paradoxical correlations of students' self-report motivation and achievement in large-scale assessments (e.g. Van de gaer, Grisay, Schulz, & Gebhardt, 2012). In all participating countries, students' self-report learning motivation tended to show a positive correlation with achievement, whereas when scores were aggregated at the country level and the correlation is computed between countries' average levels of motivation and achievement, a negative correlation was found.…”

Agreeableness, antagonism, and mental health across cultures

“…This would explain why Mexicans describe themselves as less sociable than North Americans, despite displaying more sociable behavior (Ramírez-Esparza, Mehl, Á lvarez-Bermúdez, & Pennebaker, 2009). Students in schools with high standards achieve more, but have lower academic self-concepts than lower achieving students in less challenging contexts (Van de Gaer, Grisay, Schulz, & Gebhardt, 2012). Heine et al (2008) found indicators of high national Conscientiousness (e.g., postal workers' speed, per capita GDP, life expectancy) to be positively correlated with perceptions of national character, but negatively with national mean Conscientiousness scores.…”

“…The inventory structure was also stable in a sample of French-speaking African countries, and when translated into Moor e, and administered by interview to villagers in Burkina Faso (Rossier, Ouedraogo, Dahourou, Verardi, & Meyer de Stadelhofen, 2013). Though thresholds for impairment may vary, there are surely individuals who "deviate markedly from the expectations of their culture" in every society (Tseng, 2001). Paris and Lis (2013) argue that social risk factors, including modernity's disruptions to family structures, influence whether the traits that underlie PDs lead to disordered behavior-people with impulsivity and affective instability exist everywhere, but psychosocial risk or protective factors can determine whether their expression becomes pathological.…”

Aspects of students’ civic engagement

“…When interpreting cross-country comparisons of questionnaire data, please be aware that the formats used to gauge respondents' attitudes or perceptions across diverse national contexts may not always measure respondents' beliefs consistently across the different languages and cultures (for evidence of this matter, see, for example, Desa, van de Vijver, Carstens, & Schulz, in press;Heine, Lehman, Peng, & Greenholtz, 2002;van de Gaer, Grisay, Schulz, & Gebhardt, 2012). Although ICCS extensively reviewed issues of measurement invariance during the development variations of scale scores across countries may be partly due to differences related to cultural or linguistic contexts.…”