The study was conducted among all undergraduate medical students (freshmen to juniors) at Jining Medical College, Shandong Province (Nikolic et al., 2023). A cluster sampling method was employed: all classes held between September and December 2024 were randomly selected, and every student in the chosen classes was invited to participate without any additional screening.

Inclusion criteria: (1) full-time medical undergraduate; (2) aged ≥ 18 years; (3) daily mobile-phone use ≥ 1 h; (4) provided written informed consent. Exclusion criteria: (1) self-reported severe neurological or psychiatric disorder; (2) non-medical major; (3) > 20 % questionnaire items missing.

Investigators received standardized training on scale administration. Written informed consent was obtained immediately before questionnaire distribution. Completed forms were checked on the spot; missing responses were supplemented when possible. A total of 2,977 students were approached and 2,679 valid questionnaires were retained (effective response rate = 90.0%). The study protocol was approved by the Research Ethics Committee of Jining Medical University.

In this study, data entry and cleaning were conducted using Epi Data. After double-checking by two individuals, the data were analyzed using SPSS 26.0 software. Descriptive statistics were presented as Mean (SD) for scale scores and as frequency (percentage) for demographic variables. In view of the large sample size (all cells n > 100), absolute skewness ≤ 3 and kurtosis ≤ 10 were taken as evidence of approximate univariate normality (Kline, 2015). Levene's test for equality of variances was used to assess whether the variances were equal across groups. Independent samples t-tests and ANOVA were employed to examine differences in AMS scores among groups, for ANOVA, post-hoc pairwise comparisons were carried out with LSD when variances were equal or Tamhane's T2 when variances were unequal. Pearson correlation analysis was used to assess the relationships between AMS scores and other scale scores. Variables with statistically significant differences in demographic characteristics, as well as PHQ-9, GAD-7, and MPAI, were included in multiple linear regression analysis. The assumptions of this analysis—linearity, independence, homoscedasticity, multicollinearity (VIF < 10), and normality of residuals—were examined and met. Parallel mediation effect analysis was conducted using Model 4 of the PROCESS macro. A significance level of p < 0.05 was applied to all statistical tests.

The Harman single-factor test was employed to assess common method bias. The results indicated that there were five factors with eigenvalues greater than 1, and the variance explained by the first common factor was 18.42%, which is below the critical value of 40%. Therefore, it can be concluded that there is no significant common method bias in the data of this study.

Correlation analysis was conducted among four variables: depression (PHQ-9), anxiety (GAD-7), mobile phone addiction (MPAI), and achievement motivation (AMS). The results showed that there were significant correlations among anxiety, depression, mobile phone addiction, and achievement motivation (p < 0.01), as presented in Table 1.

To further explore the underlying mechanisms between Mobile Phone Addiction Inventory (MPAI) scores and Achievement Motivation Scale (AMS) scores, mediation analysis was conducted, incorporating factors with statistically significant differences identified in the previous analyses (Gender, Left-behind experience before 18 years old, Like own professional, Will choose own professional again, Family annual income, Father's educational level, Mother's educational level, Goals after graduation) as covariates.

Table 2 presents the results of the mediation analysis. The direct effect of MPAI scores on AMS scores, as well as the indirect effects mediated through Generalized Anxiety Disorder (GAD-7) scores and Patient Health Questionnaire (PHQ-9) scores, were examined. The confidence intervals (CIs) for both the direct and indirect paths did not include zero, indicating significant relationships. Specifically, the total direct effect of MPAI scores on AMS scores was −0.274, with a 95% CI of [−0.312, −0.237]. The total indirect effect was −0.114, with a 95% CI of [−0.133, −0.095]. The specific indirect paths were as follows: the path mediated through GAD-7 scores had an effect of −0.027 (95% CI = [−0.047, −0.008]), the path mediated through PHQ-9 scores had an effect of −0.032 (95% CI = [−0.042, −0.023]), and the chained mediation effect through both GAD-7 and PHQ-9 scores had an effect of −0.055 (95% CI = [−0.070, −0.041]). These effects accounted for 6.96, 8.25, 14.17, 10.56, and 11.76% of the total effect (total effect = −0.388; 95% CI = [−0.423, −0.353]), respectively. These results suggest that GAD-7 and PHQ-9 scores may be important psychological pathways through which MPAI scores are associated with AMS scores. The detailed path model is illustrated in Figure 1.

The reliability and validity of the four scales were examined (Supplementary material S6–S10). Overall, internal consistency was excellent (Cronbach's α 0.83–0.91), convergent validity satisfactory (AVE 0.49–0.89), and all standardized loadings >0.50 (p < 0.001), indicating sound psychometric properties.

The demographic characteristics of the participants were statistically described in this study. Specifically, there were 1,156 males (43.15%) and 1,523 females (56.85%). In terms of age, 1,278 participants were 18 years old or younger (47.70%), 559 were 19 years old (20.87%), 575 were 20 years old (21.46%), and 267 were 21 years old or older (9.97%). Regarding family structure, 749 participants were only children (27.96%), and 1,930 were not only children (72.04%). For more detailed demographic information, see Tables 3, 4.

The scores of the scales involved in this study were statistically described using SPSS 26.0 software: PHQ-9 (4.99 ± 3.88), GAD-7 (3.66 ± 3.60), MPAI (42.97 ± 12.12), and AMS (1.87 ± 12.35). Histograms of the scale scores are presented in Figure 2. Normality of the AMS total score was examined in SPSS 26.0 for the overall sample (N = 2 679). The distribution exhibited a skewness of 0.375 (SE = 0.047) and a kurtosis of 0.737 (SE = 0.095). Following Kline (2015) guidelines for large samples (|skewness| ≤ 3, |kurtosis| ≤ 10), these values indicate an approximately normal distribution, satisfying the assumption for parametric analyses.

Tables 3, 4 summarize AMS-score differences among medical-student subgroups. After Bonferroni correction (k = 5, α = 0.010), only-child status was no longer significant. Male students, those who liked their major and those willing to re-choose it reported significantly higher AMS scores (p < 0.001), whereas students with left-behind experience before age 18 scored significantly lower (p < 0.001). One-way ANOVA showed no significant differences by age or monthly living expenses; however, scores differed significantly across categories of annual family income, father's education, mother's education and post-graduation goals.

The results of multiple comparisons showed that for family annual income, the AMS scores of the group with an income of 150,001 Yuan and above were significantly lower than those of the groups with incomes of 10,001–30,000 Yuan (p = 0.001), 30,001–50,000 Yuan (p < 0.001), 50,001–80,000 Yuan (p = 0.024), and 80,001–100,000 Yuan (p = 0.003). The AMS scores of the group with an income of 100,001–150,000 Yuan were significantly lower than those of the groups with incomes of 10,001–30,000 Yuan (p = 0.004) and 30,001–50,000 Yuan (p < 0.001).

For father's educational level, the AMS scores of children whose fathers had a college degree or above were significantly higher than those whose fathers had primary school or below (p = 0.008), junior high school (p < 0.001), and high school (p = 0.036). The AMS scores of children whose fathers had a master's degree or above were significantly higher than those whose fathers had junior high school (p = 0.004).

For mother's educational level, the AMS scores of children whose mothers had primary school or below were significantly lower than those whose mothers had junior high school (p = 0.002), high school (p = 0.047), and associate degree (p = 0.017). The AMS scores of children whose mothers had a college degree or above were significantly higher than those whose mothers had primary school or below (p < 0.001), and high school (p = 0.022).

For post-graduation goals, the AMS scores of the postgraduate entrance examination group were significantly higher than those of the no-direction group (p = 0.002) and the employment group (p < 0.001). The AMS scores of the civil servant examination group were significantly higher than those of the no-direction group (p = 0.007) and the employment group (p = 0.036). The AMS scores of the employment group were significantly higher than those of other groups (p = 0.021).

For more detailed results, refer to Supplementary material S11.

A multiple linear regression analysis was conducted with AMS scores as the dependent variable, incorporating variables with statistically significant differences in the demographic characteristics analysis, as well as scores from the PHQ-9, GAD-7, and MPAI. The results indicated that gender, liking one's major, PHQ-9 scores, GAD-7 scores, and MPAI scores had statistically significant effects on AMS scores. Model fit: R² = 0.18, F_{(23, 2655)} = 22.3, p < 0.001; residuals were normal, homoscedastic, and no severe multicollinearity was present (all VIF < 2), satisfying the assumptions of multiple linear regression (see Table 5 footnote for details).

Achievement motivation is closely related to mental health (Varghese et al., 2015). In this study, Chinese medical students obtained a mean AMS score of 1.87. Multiple linear regression showed that being male and liking one's major were associated with higher AMS scores, whereas higher PHQ-9, GAD-7, and MPAI scores were associated with lower AMS scores (Shahroudi et al., 2019).

This study has validated the factors influencing achievement motivation among Chinese medical students and constructed a chain mediation model of “Mobile Phone Addiction → Anxiety Symptoms → Depressive Symptoms → Low Achievement Motivation,” revealing the critical role of anxiety and depressive symptoms in the relationship between mobile phone addiction and achievement motivation (Putwain and Daly, 2021). This finding not only enriches the theoretical understanding of the relationship between mobile phone addiction and achievement motivation but also provides practical suggestions for medical students, educators, psychological counselors, and parents. Medical students should rationally recognize the dangers of mobile phone addiction, take proactive measures to prevent it, and actively engage in academic supervision (Meng et al., 2025). Schools should place high importance on students' mental health by offering mental health education courses and providing professional psychological counseling to help students effectively cope with negative emotions and enhance their psychological resilience (Xue et al., 2024; Hatice and Yuksel, 2016). Additionally, teachers should actively create a lively classroom atmosphere, optimize course settings, improve teaching methods, and increase classroom engagement to provide students with positive emotional experiences and a sense of achievement during the learning process, thereby further strengthening their motivation to pursue success. Importantly, the chain mediation path explains 42 % of the variance in achievement motivation, indicating that interventions targeting any link—reducing phone addiction, alleviating anxiety, or mitigating depressive symptoms—are likely to yield measurable gains in students' academic drive (Chen et al., 2024).

The current study has achieved significant results, yet it also has some limitations that should be addressed in future research. The sample in this study was limited to a single medical college in Jining, Shandong Province, China. The narrow geographical scope and the specific type of institution may restrict the generalizability of the findings. Future research should consider expanding the sample to include medical schools from different regions and at various levels to enhance the representativeness of the conclusions.

Moreover, the cross-sectional design of this study limits the ability to establish causal relationships between the variables. Subsequent research could employ longitudinal cohort study designs to observe the developmental trajectories of mobile phone addiction, anxiety, depression, and achievement motivation over an extended period. This approach would provide a deeper understanding of the causal relationships and offer more compelling theoretical support for targeted interventions.

Additionally, the precise nature of mobile content consumed by participants was not assessed in the present study. We intend to incorporate comprehensive content-use items in future work to disentangle these potential differential effects.