Can self-perceptions serve as mechanisms of achievement inequality? We addressed this question by analyzing data from the 2018 PISA survey, including 520,729 students across 70 countries. We identified five dimensions of children’s self-perceptions: self-perceived competency, self-efficacy, growth mindset, sense of belonging, and fear of failure. As predicted, across countries, children’s self-perceptions partially mediated the association between socioeconomic status and academic achievement. The effect of self-perceived competency on achievement was stronger in countries with higher social mobility and lower income inequality, indicating the importance of environments that “afford” the use of beneficial self-perceptions. By contrast, self-efficacy, growth mindset, and sense of belonging were less strongly positively related to achievement in countries with higher social mobility, suggesting that those self-perceptions might become less important positive predictors for achievement under more supportive, equalizing external conditions.
Associations Between Ses, Self-perceptions, And Achievement
Our findings substantiate the idea that children’s socioeconomic background can become ingrained in their self-perceptions 46–49 and confirm the importance of these self-perceptions for academic achievement 50,51. Although all self-perception variables were associated with achievement 22,23,26,29, this bivariate association was strongest for self-perceived competency and almost negligible for fear of failure. Self-perceived reading competency is the self-perception variable most directly related to academic achievement in general and performance in the domain of reading in particular. The domain-specificity of this variable 52 can explain the especially strong association with reading achievement in our analyses. The weak association with fear of failure however has a more complex interpretation. We consider fear of failure as “persistent and irrational anxiety about failing to measure up to the standards and goals set by oneself or others” 53. While this self-perception variable sometimes seems to impair academic outcomes 32, it has also been found to contribute to an increase in performance 54. Children may respond differently to fear of failure. For example, some children may try to avoid failure by succeeding and striving, whereas others may deal with their fear through counterproductive activity aimed primarily at self-protection 55,56. The measure used in this study does not distinguish between those contradictory reactions to fear of failure, which may have obscured actual associations with SES and achievement. Future research might try to identify systematic differences in self-perception that underlie those divergent reactions to fear of failure.
Do Self-perceptions Mediate Ses Disparities In Achievement?
Our findings show that self-perceptions, both considered jointly and separately, partially mediate the association between SES and achievement. In line with our bivariate analyses, self-perceived competency was the strongest mediator variable, whereas self-efficacy turned out to be the only slightly negative mediator variable, resulting from a positive mediation in some countries and a negative mediation in others. We will discuss this finding in more detail in the next section.
In addition to the significant indirect effect, the direct effect of SES on achievement was still significant and explained a larger proportion of the variance in reading achievement as compared to the indirect effect. From a global perspective, our analyses provide evidence that children’s socioeconomic background can indeed become ingrained in their self-perceptions, which then influence their academic achievement, substantiating self-perceptions as mechanisms of achievement inequality. However, the partial mediation clearly shows that other mechanisms might also play a significant role here. Moreover, the mediation via self-perceptions might be based on different assumptions depending on the broader societal environment in different countries. The large variations in the size of the correlation between SES, self-perceptions, and achievement across countries support this interpretation. SES-related structural barriers such as reduced access to educational materials or extracurricular activities might be perceived differently depending on what you can expect to achieve in life through performance as well as on your comparison group. Children from different countries can hence be expected to differ in the way they make sense of their everyday experiences.
In recent literature, self-perceptions in youth have received some attention for promoting deficit perspectives of individuals 57,58. Deficit perspectives in research can lead to a focus on isolated variables and specific behaviors, and the extent to which youth exhibit them. Deficit models are guided by research that focuses on children and adolescents who fail to meet predetermined standards for an isolated skill or characteristic 59. While we have some evidence that self-perceptions partially mediate the association between SES and achievement, we do not ascribe to deficit models of youth that attribute individual lack as the factor for achievement. Instead, we believe there are possibilities for future work to use innovative quantitative (e.g., person-centered analyses) and qualitative research to move beyond a deficit model. Qualitative methods, for instance, can provide ways for researchers to discover what youth can do and achieve beyond the isolated variables 60. Research yielding positive findings about children's accomplishments has the potential to move educators beyond the deficit perspective that pervades much of education research in general.
Moderation By Country-level Income Inequality And Social Mobility
We had two opposing hypotheses about how the effect of SES on self-perceptions is moderated by country-level income inequality and social mobility. Our results provided some evidence for the hypothesis that SES exerts a weaker effect on children’s self-perceptions in countries with higher levels of income inequality and lower levels of social mobility, presumably due to segregation and fewer opportunities for social comparisons. More specifically, in countries with higher income inequality and lower social mobility, the effect of SES on self-perceived competency was weaker. For higher income inequality, we found the same effect for self-efficacy and for lower social mobility, for fear of failure.
Similarly, we had two opposing hypotheses about how the effect of self-perceptions on reading achievement is moderated by country-level income inequality and social mobility. We found partial support for both perspectives. The positive mediation effect of self-perceived competency was more pronounced in countries with higher social mobility and lower income inequality, indicating the importance of environments that “afford” the use of beneficial self-perceptions. Fear of failure was also more positively related to achievement in countries with higher social mobility. Following the same line of argumentation, an environment that offers more opportunities to evolve, might also “afford” more opportunities to fail. Sense of belonging, self-efficacy, and growth mindset, however, turned out to be more strongly related to achievement in countries with lower social mobility, suggesting that those self-perceptions might become less important for achievement under more supportive external conditions.
The results for self-perceived reading competency were consistent across social mobility and income inequality and pointed in the same direction for both its relation to SES and reading achievement. Consequently, this self-perception variable seems to be an important lever to address inequality, especially in countries with higher social mobility and lower income inequality. Under these conditions, children’s self-perceptions seem to strongly reflect SES differences. At the same time, those conditions can be considered fertile soil that provides a good basis for educational growth and the development of individual potential 40. Under those conditions, a strong belief in one's own domain-specific abilities seems to drive students to higher levels of performance and motivate them to take advantage of the resources offered 61. Interventions targeting self-perceived competency 62 might hence be particularly effective in counteracting educational inequalities in countries with rather high social mobility and low income inequalities. In the present study, we provide evidence for this mechanism in the domain of reading. However, based on the broad literature on self-perceived competency (or self-concept) and its relation to performance in different domains 63, we would expect similar results in other domains as well.
With the exception of fear of failure (we discussed the downsides of this variable above), the mediation effects of self-efficacy, growth mindset, and sense of belonging were indeed significantly moderated by social mobility. These effects are primarily based on the associations between self-perceptions and achievement, given that the moderation effects for those variables’ associations with SES were less consistent. Our findings are based on a comprehensive measure of social mobility reflecting multiple intergenerational outcomes (e.g., health, working conditions, technology access). These outcomes represent the country-level resources available to individuals to exploit their potential. Accordingly, self-perceptions indeed seem to become less important for achievement when external resources are abundant. At the same time, children’s internal resources become more important for achievement when you have to strengthen yourself in the face of demotivating and restrictive external conditions. The belief that you are socially accepted and able to handle difficult situations, grow, and develop even in less supportive and less permeable environments might provide the motivation needed to make the most of the learning opportunities offered 64. This interpretation does not apply to narrower measures focusing on a specific aspect of educational equity at the country level (e.g., educational mobility as implemented in 42).
Importantly, for a number of countries, the relation between self-efficacy and achievement was negative and becoming more negative with increasing social mobility, indicating that self-efficacy was associated with increasingly lower (not higher) achievement. Researchers note perceptions of self-efficacy or resilience do have problematic issues, discussing, for example, the notion that emphasizing too much self-efficacy can be troubling for individuals 65. For instance, while exercise can put good strain on the heart, too much exercise can actually be damaging. Extreme self-efficacy can drive learners to overly persistent, but unattainable goals. Called the “false hope syndrome”, too much self-efficacy can make people overly tolerant of adversity 66. Future research might identify country-level characteristics that allow to distinguish between countries where self-efficacy and achievement are positively vs. negatively associated. Then we can try to better understand the moderating effect of social mobility for self-efficacy.
On a general note, simply changing low-SES children’s self-perceptions, without addressing the multitude of structural barriers that gave rise to them, may not be effective and may inadvertently convey that the children are themselves to blame for their predicament. For instance, school SES 67 or home and school learning environments 68–70 can also play an important role in reading achievement. Our work looks into the notion of country-level moderators and the global level in trying to understand self-perception variables as mediators. We believe that future research in this area can combine the country level, the community level, proximal environments, and the individual to flush out a broader picture of self-perceptions and their role in explaining socioeconomic achievement gaps.
Limitations
Our study has several strengths, including its examination of self-perceptions as mechanisms of inequality across the globe, its novel meta-analytic technique to model cross-country heterogeneity, and its investigation of both social mobility and income inequality as potential country-level moderators. Despite the robust results presented here, our study has some limitations. First, there are still intrinsic limitations to using survey-based and self-report measures, such as demand characteristics, response biases, test effort, and fatigue 71. Second, PISA is also cross-sectional in nature, only looking at data in 15-year-old adolescents. Future studies need to make use of longitudinal data and methods in order to further explore these findings at different developmental time periods. Third, the self-perception variables and measures were also limited to what was available in the PISA student questionnaire, which fared well for the exploratory nature of this study. However, future studies need to draw on both experimental and qualitative methods to further unpack the relationships between self-perceptions and achievement inequality found here.
Moreover, an important limitation of PISA is that it includes no country from Africa. Consequently, all countries from the African region were excluded from our global analyses. Future studies could make use of cross-sectional and longitudinal datasets specific to Africa 72 to compare the results of this study.