Subjective well-being

The amount of research on subjective well-being in the context of education has significantly increased in recent years. In 2015, a well-being survey was added to the Programme for International Student Assessment (PISA) questionnaire, and later, the Organization for Economic Cooperation and Development (OECD) identified well-being as a valuable goal of education [12]. However, research on well-being is complicated by the large number of terms used to describe well-being in different theoretical traditions [4,5,13] as well as by the broad theoretical framework used in international documents. For example, in the OECD framework, well-being is described through five domains: cognitive well-being, psychological well-being (which includes life satisfaction), physical well-being, social well-being, and material well-being [12]. To understand how well-being is viewed in a particular study, it is often necessary to analyze how the authors have operationalized the construct. In this study, we examine well-being from a hedonistic approach, where well-being is considered as subjective beliefs that an individual’s life is pleasant and good. Three components of subjective well-being are recognized: cognitive or life satisfaction, and positive and negative affect [14]. In this study, the special focus is on subjective well-being in school, treated as satisfaction with different school aspects and affect toward school [15].

The study of the relationship between well-being and academic achievement is the most intensively researched field on well-being in children and adolescents [e.g., 16–18]. However, the results of such studies are often contradictory. Some studies show a positive relationship between subjective well-being and academic achievement [e.g., 18,19], while others do not find such a relationship [20] or find differences in well-being only between high and low performers [21]. Inconsistent results could be explained by differences in methodological approaches such as instruments used to measure well-being (e.g., general well-being or well-being in a specific domain; one item or a scale) and academic achievement (GPA, self-evaluation, or standardized test), as well as level of education (elementary or secondary school). The emergence of meta-analyses has made a significant contribution to the debate on the relationships between well-being and academic achievement.

In meta-analysis by Bücker et al. [4], the correlations between academic achievement and subjective well-being were small to medium in magnitude, with overall mean effect r = 0.16. Very similar results were obtained in meta-analysis by Kaya and Erdem [5]; although, they included in the analysis not only studies of subjective well-being as life satisfaction in Diener’s conceptualization [14], but also within different domains of subjective well-being such as psychological well-being, social well-being, cognitive well-being, and physical well-being. Kaya and Erdem [5] also demonstrate positive but small mean effect size of the relationship between students’ well-being and their academic achievement (r = 0.17).

The results of meta-analyses suggest that, although not to a strong degree, children’s subjective well-being relates to their academic achievement. However, there may be several different mechanisms that explain this relationship. For example, the weak association between well-being and academic achievement may be explained by the inflexibility of the school environment, which links a child’s success at school and personal development only to learning outcomes. In this case, such a link between well-being and achievement would have a rather negative connotation. There could also be an alternative positive explanation. Since the main purpose of school is to promote learning, the possibility to fulfill the psychological need of competence through academic achievement could bolster students’ well-being.

In the Russian context, there are almost no studies on the relationship between subjective well-being and academic achievement in schoolchildren. The only available study, conducted among primary school students (n = 144), found differences in school satisfaction and affect toward school between students with low and high levels of math achievement [21]. Therefore, the present study could provide significant insights into the phenomena of well-being and academic achievement in Russia.

Students’ beliefs and perceptions about themselves and their abilities, such as self-efficacy, could represent another source of explanation. Therefore, to better understand the mechanism of the relationship between well-being and academic outcomes, this study will test the moderating effect of self-efficacy on the relationship between well-being and academic achievement.

Self-efficacy

Self-efficacy is another source of motivation that relates to competence. Students’ self-efficacy is their perceived confidence to do a task and study well [22–24]. Research studies show that self-efficacy positively relates to students’ academic outcomes and psychological constructs of motivation, self-regulated learning (SRL), and subjective well-being [6,8,23,25–27]. Self-efficacy changes depending on tasks, mastery experiences, and successes or failures [25]. In addition, self-efficacy and academic achievement show a reciprocal relationship [6]. That is, high academic achievement contributes to an increase in self-efficacy, and high self-efficacy relates to higher academic achievement, as is evidenced by the results of the meta-analysis of studies (n = 11) examining self-efficacy and academic outcomes [6].

High self-efficacy tends to be a strong predictor of students’ achievement and success [6,27]. Generally, students with high self-efficacy exhibit higher motivation and academic achievement, seek more opportunities to learn, regulate their own learning, interpret their academic failures as a lack of effort, and perceive learning difficulties as challenges to resolve than students with low self-efficacy [6,23–25,27].

Researchers debate over how to conceptualize and measure self-efficacy, stating that self-efficacy can be domain-specific or domain-general [27]. Bandura [23] outlined three levels of specificity when discussing self-efficacy. The general level refers to students’ beliefs about their general academic self-efficacy to study or general self-efficacy within a domain, such as “I am sure I can do well in math.” The intermediate level refers to beliefs about certain skills or competencies, such as “I am sure I can solve math equations.” Finally, the specific level of self-efficacy focuses on a task, such as “I am sure I can solve this quadratic equation.” Irrespective of the views, Bandura [23] emphasized the importance of assessing “… the multifaceted ways in which efficacy beliefs operate within the selected activity domain” (p. 310). Talsma and colleagues [6] identified in their meta-analysis that the effect sizes for scales measuring domain-specific self-efficacy were higher than for scales measuring general self-efficacy.

In the Russian context, a few studies examined self-efficacy among university students [28–30] and focused on foreign students learning Russian [31]. For instance, the study of the third-year university students (n = 558) indicated medium levels of self-efficacy; however, the group with the medium level of self-efficacy showed higher levels of job stability than students with low levels of self-efficacy [29]. Several studies examined Russian middle school students’ self-efficacy and academic achievement [32,33], including relationships with SRL [34] and science identity [35]. Belova and colleagues [35] examined the relationship of science learning self-efficacy with science identity, checking for the moderation effects of student gender and science success among Russian high school students (n = 519). The results of the study supported the significant link between science communication efficacy and science identity. Also, gender did not show a moderating effect between self-efficacy and science identity. Finally, science success did not significantly moderate this link; however, interaction of science success with science communication efficacy resulted in the negative effect on science identity of high school students. Belova and colleagues [35] speculate that a nuanced moderation effect of science success may suggest that students who perform well in science but do not feel confident in communicating about science may experience negative effects on developing their science identity.

A brief overview of studies in the Russian context show that self-efficacy is researched among middle or high school students, and mostly cover the samples of university students. Hence, similar to the foreign research studies, Russian findings focus mostly on relationships of self-efficacy with the academic results, job satisfaction, coping strategies, but not really on well-being and samples of elementary school students.

Subjective well-being and self-efficacy

Research studies in the field of education provide compelling evidence of the effects and correlations of self-efficacy with academic outcomes [25,27,34,35]. In recent years, researchers have also shifted their attention to other psychological constructs, such as subjective well-being.

Research studies show that self-efficacy has a mediating effect between students’ self-concept and subjective well-being on a sample of migrant and native Chilean adolescents [n = 406; 8]. A study conducted among Italian high school and university students (n = 465) during the COVID-19 lockdown identified that self-efficacy for self-regulated learning and self-efficacy for positive and negative emotions were related to cognitive and emotional well-being [7]. Yet another study of German 8^th and 9^th grade students (n = 767) identified that positive school climate, self-efficacy, and test anxiety predicted subjective well-being and academic achievement [9].

In the Russian context, we could identify a comparative study examining well-being and self-efficacy in Russian and Kazakh students [36] and a master’s degree thesis that examined the well-being and self-efficacy of Russian adolescents [10]. The master’s thesis examined direct and indirect effects between self-efficacy in reading and subjective well-being on a sample of more than 15,000 Russian adolescents from the PISA-2018 assessment and showed that general self-efficacy served as a mediator between self-efficacy in reading and subjective well-being [10]. In addition, reading self-efficacy had a positive effect on students’ subjective well-being.

Similar to the studies conducted around the world, there was an attempt to examine the mediating effect of self-efficacy between SRL and academic results in a sample of Russian elementary students [11]. Akhmedjanova [11] examined the relationships among SRL, self-efficacy, and academic outcomes (math and reading) of the fourth-grade students (n = 2,158) using structural equation modelling [11]. While the results indicated the negative direct effects of SRL on academic outcomes, self-efficacy served as a mediator between SRL and academic achievement in both math and reading models in the analyses by gender and location. In this way, self-efficacy serves as a mediator between SRL and academic outcomes of elementary school students in Russia. However, it is not clear how it relates to subjective well-being.

Current study

Studies in other countries provide evidence of relationships between self-efficacy and subjective well-being, stating specifically that it serves as a mediator between the psychological constructs and academic outcomes. In addition, most of these studies both in Russia and across the world are conducted with samples of adolescents and young adults. While there are many studies examining self-efficacy and various academic and non-academic outcomes with older students, there is a research gap on how these variables relate to each other in elementary school students. Therefore, the goal of this paper was to examine the role of subjective well-being in school and self-efficacy in predicting academic achievement in math and reading among elementary school students in Russia (Fig 1).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 1. The Moderating Effect of Domain-specific Self-efficacy between Subjective Well-being in School and Academic Outcomes of Elementary School Students.

https://doi.org/10.1371/journal.pone.0341318.g001

To achieve the research goal, we addressed the following research questions:

1. What is the relationship between students’ outcomes in math and reading and their subjective well-being in school and domain-specific self-efficacy?
2. Is there a moderation effect of domain-specific self-efficacy on the relationship between students’ results in math and reading and their subjective well-being in school?
3. Are there gender differences in the relationship between the student’s test scores in math and reading and their subjective well-being in school and domain-specific self-efficacy?

Study design

This study is part of the longitudinal project examining the factors contributing to students’ academic failure in Russian primary schools. For the purposes of this study, we used survey data from elementary students in the first wave of data collection in fall 2022.

Procedures

The data collection took place online in 40 public schools in one of the regions of central Russia after receiving approval from the Ethics Committee (#19). Parents were informed about the purpose of the study and signed online consent forms. Before starting the assessment, children provided their assent to participate in the study. Data were collected on two separate days: one day for math and reading assessments and another day for survey data.

Sample

To recruit participants, we used the random stratified proportional sampling at the level of schools. That is, the unit of selection was an educational organization – school. For school to be included into the sample, it should have had a within school support system to address school failure across urban and rural areas, and 42 schools fit this criterion. As a result, nine municipalities in a central region of Russia were included in the study across three types of locations: a large city, small towns, and rural areas, resulting in 40 schools that agreed to participate and were able to provide research data. The sample included responses from the fourth graders, 9 and 10 years old. The missing data analysis revealed that the dataset included 13% or less of missing cases in variables underpinning constructs of self-efficacy in math and reading, affect towards school, and satisfaction with school. The missing values originated due to students skipping specific questions in these constructs. Rigorous approaches such as multiple imputation require appropriate auxiliary data that are predictive of both the missingness and the missing values to produce reliable imputations [37]. The dataset provides only limited auxiliary information – students’ gender, outcome variables (test scores), and location. Given the limited auxiliary variables, we opted for listwise deletion to preserve the validity and interpretability of the findings while acknowledging its limitations. The missing data were removed using listwise deletion, resulting in the 1,962 students’ responses (50.4% girls, 7.7% in rural areas), which were used for the analysis. The listwise deletion was chosen since the sensitivity analyses indicated that the main results were robust to this choice.

Instruments.

To measure students’ well-being, two subscales from the Survey of subjective well-being in school (SSWBS) were used: satisfaction with school (k = 7) and affect toward school (k = 3) [15]. The Likert-type agreement scale ranging from 1 to 4 was used for satisfaction with the school scale, and the Likert-type frequency scale ranging from 1 to 4 was used for affect toward the school scale. The reliability indices revealed appropriate estimates (α = 0.87; α = 0.89).

The domain-specific self-efficacy scales were developed for the longitudinal project [38]. In this study, we used two scales: self-efficacy for mathematics (k = 4) and self-efficacy for reading (k = 4) on a four-point Likert-type scale (4 – I can do it well, 1– I cannot do it at all). An example item: “Can you solve a math problem?”. The internal consistency for both scales was good: α_math = 0.8; ω _math = 0.81; and α_read = 0.78; ω_read = 0.8.

To measure students’ academic achievement in mathematics and reading, the Progress assessment was used. Progress is the computer adaptive test to measure students’ mathematics and reading achievement [39,40]. The mathematics domain measures spatial concepts, measurement, patterns and sequences, modeling, and data handling and includes 30 dichotomous multiple-choice items with acceptable psychometric properties [40]. The internal consistency of the whole scale indicated high reliability estimates, α = 0.88; ω = 0.89. The reading test measures students’ skills of searching for information, making interpretations, analyzing, and synthesizing, and includes 3 texts with 17 items [39]. The internal consistency of the whole scale indicated acceptable reliability estimates, α = 0.63; ω = 0.67. Generally, the reliability estimates of > 0.70 are acceptable for conducting further analyses. However, some scholars establish acceptable reliability estimates at 0.65 [41] or at 0.6 [42]. Due to the reading test reliability estimates falling between 0.63 and 0.67, we concluded that the estimates were acceptable to conduct further analyses; however, the results of the reading model should be interpreted with caution.

Data also provides some students’ characteristics. In our analysis we controlled for students’ gender (1 = girl) and school location in rural area (1 = yes). In the models with school fixed effect we used schools’ ids. Descriptive statistics are presented in Table 1.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 1. Descriptive Statistics (n = 1962).

https://doi.org/10.1371/journal.pone.0341318.t001

Data analysis

To conduct the analysis, the authors accessed the dataset on 10/04/2024. The analysis was done in three consecutive steps. First, to estimate the relationship between students’ outcomes in math and reading, their self-efficacy in these subjects, and their subjective well-being, we applied regular linear regression (ordinary least squares, or OLS):

(1)

where i denotes a student, s denotes a school. represents a student’s test score in math or reading. is a student’s well-being score (satisfaction with school or affect toward school). is student’s self-efficacy in a subject (math or reading). represents a vector of covariates including students’ gender and school location in rural area. is a school fixed effect. is a vector of regression coefficients. is an error term.

The intra-class correlation coefficient (ICC), estimated from random-intercept null models, was 0.375 for math scores and 0.115 for reading scores, indicating substantial and moderate between-school variance, respectively. Although this suggests clustering at the school level, particularly for math scores, the primary analyses focused on models with school fixed effects to facilitate interpretation of moderation effects at the individual level. The use of school fixed effects is a common and rigorous approach in education research, as it controls for observed and unobserved differences between schools.

Second, to estimate the moderation effect of students’ self-efficacy, we added an interaction between variables of self-efficacy and well-being to the right side of the equation (1):

(2)

Third, to test gender differences in the relationship between the student’s math and reading scores and their subjective well-being and self-efficacy, we ran the analysis (equations 1 and 2) on subsamples of girls and boys.

At all stages, the analysis was done separately for each well-being variable (satisfaction with school or affect toward school) due to a high correlation between them in this study (r = 0.97). For the same reason, we did not use both self-efficacy in math and reading (r = 0.82) in the same model. We opted not to include both well-being subscales or both self-efficacy domains simultaneously in the same models to avoid issues of multicollinearity. By modeling these constructs separately, we aimed to provide clearer, more reliable estimates of their distinct associations and potential moderation effects. Other predictors were added stepwise. This strategy allowed for determining the changes in the coefficients of the main predictors.

In the tested models, coefficients could be regarded as unbiased estimates of causal effect, based on a strict assumption that well-being and self-efficacy variables are not related to the error terms. To decrease bias, we controlled for available students’ characteristics (gender and rural area) because they could relate to both students’ outcomes and the main predictors. We also fixed the school effect to account for all differences between schools. However, there could still be extraneous variables related to the outcomes and predictors, which is a limitation of the analysis.

The data analysis was conducted in R using the lm function.

Limitations

While this study explained the nature of the relationship between subjective well-being, self-efficacy, and academic performance among Russian schoolchildren, it has some inherent limitations. The study was cross-sectional and included only the results of the fourth-grade students in one of the regions of Russia. In addition, self-report questionnaires were used to measure subjective well-being and self-efficacy, which typically identify average results across participants and have low to medium correlations with academic outcomes [23]. Also, the reliability estimates of the reading assessment were below the established > 0.70 estimate; therefore, the results of the reading models should be interpreted with caution. For the analysis, we used only children’s outcomes, which might limit the role of parental and teacher characteristics, socio-economic status, school context, and other important variables. We did not measure or inquire about how teachers promoted and developed students’ subjective well-being and self-efficacy in various school contexts. Finally, some research evidence suggests that high academic performance leads to increases in domain-specific self-efficacy [25], suggesting that there can be bidirectional effects and calling for further research.

Future research

Therefore, as future research, we recommend examining the development of students’ well-being and self-efficacy as well as their interplay in relation to academic results longitudinally, using more psychometrically robust instruments to measure reading skills and checking for their bidirectional associations. The substantial between-school variance observed for math scores suggests that future research could further examine these relationships using multilevel models. It is also worth examining the school context and educational programs to identify how they promote the development of well-being and self-efficacy. Along these lines, it will be beneficial to conduct intervention studies with students of various ages to target the development of well-being and self-efficacy. Also, adding school, teaching, and parental characteristics to children’s outcomes might provide a more comprehensive understanding of the interplay between well-being, self-efficacy, and academic outcomes. We also call for qualitative studies which might potentially shed light on possible gender roles in school contexts, focusing on parental, teachers’, and children’s perceptions, behaviors, and practices in and outside of class in relation to self-efficacy and subjective well-being.

Implications for practice

Irrespective of the limitations outlined above and small effect sizes, we confirmed the role and importance of self-efficacy in academic results as well as its relationship to subjective well-being for elementary school students in Russia. Based on the results, self-efficacy for math is equally important for girls’ and boys’ results in math. In contrast, for girls, self-efficacy in reading is more important for their results in reading than for boys. Hence, both types of self-efficacy are important for girls and moderate the relationship between subjective well-being and their academic results. Therefore, one of the practical implications is to target the development and increase in self-efficacy for math, especially for elementary school girls. That is, it is important to create opportunities of professional development for teachers so they can learn about how to increase and sustain high levels of self-efficacy through such practices as creating situations of academic success by assigning tasks within students’ zone of proximal development, providing supportive feedback on how to do tasks, and acknowledging students’ emotional reactions to successes and failures, to name a few [23].

Hence, teachers should be trained on how to deliver constructive feedback regarding math to girls, since it might potentially lead to an improvement in their self-efficacy for math and, further, increase their subjective well-being and math performance. It is especially important since research findings suggest a decreasing trend in subjective well-being in adolescence [47,48]. It is also worth improving self-efficacy for reading since it is one of the basic literacy skills that should be mastered to tackle tasks across subjects in school. To support the proposed practical implications, it is worth drawing the attention of policy makers to assist with the development and application of professional development opportunities for elementary school teachers.