The questionnaire and tests were successfully completed by 16,479 ATHSSs from 242 top senior high schools across 20 provinces (including municipalities and autonomous regions) resulting in a response rate of 88.50%. For the 242 sample schools in this study, the average admission rate (64.55%) into first-tier universities was almost four times the average admission rate in the 20 provinces. The average number of students enrolled in Peking University and Tsinghua university in 2017 and from 2013 to 2017 were 9.35 and 46.21 respectively, which are both about nine times the national average.
Among the respondents, 26.42% chose the liberal arts academic track, 53.78% were male and 77.67% were from an urban area of China (Table 1).
Current status of medical career expectations of ATHSSs
Table 2 shows the distribution of career expectations among the ATHSSs. From the 18 career choices, ‘economist or financial analyst’ (33.49%) ranked first, followed by ‘college/university faculty’ (33.07%) and ‘freelancer or start-up pioneer’ (32.24%). In contrast, ‘doctor’ only accounted for 20.60% and ranked seventh overall.
Potential factors affecting medical career expectations of ATHSSs
ATHSSs with and without MCE differed significantly in many aspects (Table 3). The ratio of male ATHSSs with MCE was significantly lower than those without MCE (t = 8.52, p < 0.01). The proportion of students from urban areas with MCE was significantly lower than those without MCE (t = 7.38, p < 0.01).
Among the ATHSSs with MCE, the parents’ years of schooling were significantly lower than those without MCE (father: t = 7.77, p < 0.01; mother: t = 8.76, p < 0.01). For parents’ ISEI, there were also significant differences between ATHSSs with and without MCE (father: t = 9.55, p < 0.01; mother: t = 7.88, p < 0.01). Specifically, both the fathers and mothers of ATHSSs with MCE were significantly lower on the ISEI than those without MCE. In addition, the proportion of ATHSSs with MCE who came from high-income families was significantly lower than those without MCE (t = 6.50, p < 0.01). However, no difference was found for middle-income families (t = 0.59, p = 0.28).
On academic performance, ATHSSs with MCE performed significantly worse on the academic tests than those without MCE (Chinese: t = 7.00, p < 0.01; English: t = 6.84, p < 0.01; liberal arts: t = 6.50, p < 0.01; science: t = 11.01, p < 0.01), with the exception of mathematics (t = 0.63, p = 0.26). The biggest gap between ATHSSs with and without MCE was found in the liberal arts, which on average reached 8.49 points.
Considering that p-values are affected by sample size in t-tests, we calculated the effect size of Cohen’s d for each t-test to help interpret the importance of differences observed. The results in Table 3 showed that all the significant t-tests were with a small effect size except the t-test for the liberal arts, which had a small to medium effect size.
Four logit models were conducted to examine the potential factors that affected the ATHSSs’ MCE. ATHSS with MCE was viewed as a dependent variable, and the stepwise method was used to obtain a summary of the results (Table 4). Model 1 controlled the demographic characteristics and family background of the ATHSSs, and model 2 further controlled the factors of academic performance on Chinese, mathematics and English. Considering that the students took either the liberal arts or science test, two separate analyses were carried out as model 3 and model 4.
The ATHSSs with MCE were more likely to be female, and this was exacerbated when academic performance was controlled, becoming 35.4% (1 - (e-0.436) = 0.354) higher than males. The logit models showed that ATHSSs with MCE were less likely to be from an urban area. Specifically, the percentage of ATHSSs with MCE from urban areas was significantly lower than those from rural areas (1 - (e-0.103) = 0.098), when controlling for demographic characteristics, family background and academic performance. However, this difference was more nuanced when the science test score was controlled. The results also showed that father’s years of schooling had no significant impact on ATHSSs’ MCE. In contrast, mother’s years of schooling had a significantly negative effect on MCE. In addition, high-income families showed a remarkably negative attitude towards MCE whereas middle-income families did not.
In terms of the academic tracks of liberal arts and science, the models revealed that both tracks had a significant influence on the MCE of ATHSSs. In general, ATHSSs with MCE did much worse on the liberal arts test compared to the science test. In addition, for every 1-point increase in the liberal arts test, the possibility of an ATHSS having MCE decreased by 1.3% (1-e-0.013). However, the higher the mathematics score an ATHSS on the liberal arts track had, the higher the possibility the individual would have MCE, whereas it was the opposite with the science track. Whatever the academic track was, English learning achievement had a significantly negative impact on MCE, but no difference was found for the Chinese language.