Educational level is an important aspect of one's life and is associated with biopsychosocial and economic factors. When high school students determine whether they enter college, university, or not, they may think by themselves and consult their teachers, friends, and parents. In the course of decision-making, their temperament and their relationship with parents may be crucial. Moreover, their intelligence may be one of the most important factors for educational level.

Liu et al. (1) indicated that, in China, when mothers hold dominant positions in their children's educational decisions, they are more likely to adopt a “tiger mom” approach. Recently, maternal dominance in household decision-making has increased due to social development and evolving gender roles. This shift is reflected in the emergence of various family models, such as the dynamics of “Tiger Mom and Cat Dad”, “Wife Dominant and Husband Accompanying”, and “Supermom.” Therefore, exploring the impact of maternal dominance in children's educational decision-making on the development of adolescents' human capital is essential (1).

As for parental bonding, Ling et al. (2) showed that parental antipathy and neglect and mother negative loving (a form of insecure attachment) sequentially mediated the association between maternal control and adulthood antisocial personality disorder features. Geng et al. (3) suggest that it is important to pay attention to medical students (especially women) with low maternal care and low empathy (cognitive empathy in particular), and that these students are more likely to have depressive symptoms during their medical studies and clinical work. In UK, Reyes et al. (4) showed that lower maternal attachment in early life predicted increased adolescent multiple risky behaviors, and that almost a third of the excess risk was attributable to child, maternal and socioeconomic factors, with over a quarter explained by maternal depression.

In USA, Wegemer and Vandell (5) investigated early childhood antecedents of adult political orientation using longitudinal data, and found that mothers' beliefs and parenting behaviors, children's temperament, and attachment security during early childhood were related to political ideology and party affiliation at age 26. Young children's fearful temperament and anxious attachment security, as well as mothers' authoritarian parenting beliefs in early childhood, predicted conservative political orientations at age 26. Children's abilities to focus attention and avoidant attachment security predicted liberal orientations (5).

These findings are not directly associated with educational level, but possibly indirectly associated with the level via decision-making, which may be affected by intelligence, affective temperament, and parental bonding. In the present study, we hypothesized that higher educational level may be associated with higher intelligence, hyperthymic temperament, and more maternal care. The aim of this study was to investigate this hypothesis.

Table 1 shows the data of age, parental bonding style, affective temperament, intelligence across the three groups along with the results of ANOVA. Age, paternal overprotection, maternal care, irritable temperament, and intelligence were significantly different across the three groups.

Table 2 shows the results of multiple regression analysis, where intelligence, maternal care, and male sex were significantly and positively associated with educational level, whereas age was significantly and negatively associated with educational level. On the other hand, paternal overprotection and irritable temperament were excluded from the model.

The major findings of this study can be summarized as follows: (1) higher intelligence and more maternal care were significantly associated with higher educational level. (2) male sex and younger age were significantly associated with higher education level. To our knowledge, this is the first study investigating associations between intelligence, affective temperament, and parental bonding concurrently with educational level.

With regard to our hypothesis that higher educational level may be associated with higher intelligence, hyperthymic temperament, and more maternal care, the involvement of higher intelligence and more maternal care with higher education level is in agreement with the present results. Moreover, the importance of maternal care is in agreement with the previous reports (2–5). However, unexpectedly, hyperthymic temperament was not associated with educational level. Although the reason is uncertain, hyperthymic temperament may be independent from any educational level. As shown in Table 1, however, the scores of hyperthymic temperament and cyclothymic temperament increased along with the increase of educational level, although they were not significant. Possibly, this may suggest the modest involvement of common factor such as ambition between hyperthymic and cyclothymic temperament with educational level. Moreover, the “modest” involvement may be derived from the fact that majority of the subjects were females (108 females and 22 males). This is probably because hyperthymic male individuals can express themselves as they want, whereas hyperthymic female individuals can express themselves within the limited social culture where male dominance is still prevalent at least in Japan. Interestingly, this is contrast to “tiger mom” in China (1).

As for the involvement of male sex and younger age with higher education, in old Japan, male dominance was prevalent and female individuals were less likely to enter into higher education level. Since the present subjects had mainly females, the results may reflect such tendency in older individuals. Also, the fact that three medical students were included in the subjects may contribute to the negative association between age and educational level and the positive association between intelligence and educational level. In order to investigate this possibility, we excluded these medical students, but the results were almost unchanged (data not shown).

This study had some limitations. First, this was a cross-sectional study, and causal relationships could not be determined. Second, the sample size was relatively precluded to establish a definite conclusion. Third, JARTIQ could not measure total intelligence. Forth, the sample was heavily skewed towards females (108 females vs. 22 males), which limits generalizability. Because of small sample of males, however, we could not perform additional sex-stratified analysis. Fifth, recruitment through bulletin boards and flyers might have introduced selection bias. Sixth, the negative association between age and educational level could be due to cohort effects in access to education. Seventh, further studies are required to clarify why male sex and younger age were significantly associated with higher education level and conversely why female sex and older age were not associated with higher educational level. Finally, this study focused on only psychological aspects, and biological, social, economic, and cultural aspects should be considered as well. Particularly, important potential confounders such as socioeconomic status, parental occupation, and geographic location were not included. These factors could influence both maternal care and educational attainment.

In conclusion, the present findings suggest that maternal care as well as intelligence may be associated with higher educational level. Further prospective studies are required to determine any causal relationship and investigate other factors related to educational level stratified by sex.