From October 27 to November 27, 2022, an online questionnaire survey was conducted on students from four medical colleges in Jiangsu Province, China, using a random cluster sampling method. A total of 5,048 participants were included in the final analysis, with 2540 (50.32%) being male and 2508 (49.68%) female. The questionnaire gathered general personal details including gender, grade, major, family economic level, frequency of physical activity, only-child status, and academic stress.

In order to mitigate the potential impact of common method biases, the study employed Harman's single-factor test. This approach entailed conducting an exploratory factor analysis on a combined set of items from three scales, totaling 60 items. The analysis is designed to identify the number of factors with eigenvalues greater than 1. To assess the presence of common method bias, the researchers compared the percentage of variance explained by the first factor against a threshold of 40%. If the variance explained by the first factor exceeded 40%, it indicated a significant common method bias; conversely, if it fell below 40%, it suggested minimal bias.

Different statistical analyses, including descriptive statistical analysis, t-test and Pearson's correlation analysis, were performed using IBM SPSS 26.0. Prior to analysis, all measurements were tested for normality to ensure they followed a normal distribution. The data presentation in the tables incorporated the use of N (%) for categorical variables and mean ± standard deviation (SD) for numerical variables. Pearson's correlation analysis was used to examine the relationship between cell phone addiction, psychological resilience and sleep disturbance. Model 4 (30) of PROCESS macro 4.1 of SPSS 26.0 was used to examine the mediating effect of psychological resilience, while Model 15 (31) was used to explore the moderating role of gender.

Model 4 is used to explore the mediation role of psychological resilience in the association between cell phone addiction and sleep disturbance. This involves analyzing the effect of cell phone addiction on psychological resilience, followed by an evaluation of how psychological resilience influenced sleep disturbance. The significance of the mediation effect was assessed by calculating the product of these effects and constructing confidence intervals for this product. A significant mediation effect was indicated by confidence intervals that did not include zero.

Model 15 was used to investigate the moderating role of gender on both the direct and indirect effects of cell phone addiction on sleep disturbance through psychological resilience. This model accounted for the direct impact of cell phone addiction on sleep disturbance, while also examining how psychological resilience acted as a mediator this relationship. Importantly, interaction terms were introduced to analyze how gender moderated the direct effect (Cell Phone Addiction x Gender interaction) as well as the mediating effect through psychological resilience (Psychological Resilience x Gender interaction). The hypothesis model is shown in Figure 1 .

The Harman's single-factor test was used to conduct exploratory factor analysis on 60 items across 3 scales. Of the total of 8 factors identified, those with an eigenvalue exceeding 1 were considered. Notably, the variance explained by the first factor was 27.399%, falling short of the critical criterion of 40% threshold, indicating an absence of substantial common method bias in this study.

Regarding participant demographics, information encompassed gender, grade, native place, only child or not, family economy, BMI, smoking status, drinking status, and physical exercise frequency. Notably, 59.1% of the participants hailed from urban areas, with a majority not being only (57.8%). Refer to Table 1 for a detailed analysis.

The overall mean MPAI score was 43.79 ± 13.84, while the mean CD-RISC score averaged at 61.72 ± 19.01, and the PSQI score at 5.15 ± 2.87. The primary focus of our inquiry pertained to gender difference, with descriptive analyses indicating that females registered significantly higher mean scores on both the MPAI and PSQI scales compared to males. Moreover, female scores were higher across all dimensions of the MPAI and PSQI scales, except for Sleep latency, Sleep duration and Use of sleep medication. The total CD-RISC score and its dimensions were significantly lower in females than in males (P< 0.05). Detailed information is shown in Table 2 .

Controlling for gender, grade, BMI, native place and being only child or not, partial correlation analysis was used. Findings revealed a negative correlation between the total MPAI score and the total CD-RISC score (r=-0.099, P<0.001), alongside positive correlation with the total PSQI score (r=0.273, P<0.001) and gender (r=0.105, P<0.001). Furthermore, the total CD-RISC score was negatively correlated with total PSQI score (r=-0.203, P<0.001), gender (r=-0.057, P<0.001), native place (r=-0.079, P<0.001), only child or not (r=-0.056, P<0.001). Total PSQI scores were positively correlated with gender (r=0.061, P<0.001), grade (r=0.111, P<0.001), native place (r=0.046, P<0.01) and only child or not (r=0.047, P<0.01), while negatively correlating with BMI (r=-0.036, P<0.05) ( Table 3 ).

Model 4 in SPSS Macro PROCESS 4.1 was used to examine the mediating effects. Controlling for grade, BMI, native place, and being only child or not, cell phone addiction was significantly negatively associated with psychological resilience (β= -0.132, P< 0.001). Cell phone addiction was found to be significantly positively associated with sleep disorder (β= 0.053, P< 0.001). Psychological resilience was significantly negatively associated with sleep disorder (β= -0.026, P< 0.001).

The results derived from the 5,000 bootstrap samples showed that all indirect effects were statistically significant, with the 95% confidence interval excluding 0. The total effect of cell phone addiction on sleep disturbance was estimated at 0.0560 (P< 0.001). The indirect effect of psychological resilience was scaled as 0.0034, with a 95% CI (0.0019,0.0052), accounting for 6.07% of the total effect (0.0034/0.0560). This analysis suggests that psychological resilience partially function as a mediator in the relationship between cell phone addiction and sleep disorder ( Table 4 and Figure 2 ).

After controlling for grade, BMI, native place, and being only child or not, Model 15 of the SPSS Macro PROCESS was applied to evaluate the proposed moderating mediator model, with psychological resilience as the mediator and gender as the moderator. Specifically, gender was incorporated as a dummy variable (male = 0, female = 1). The results showed that gender moderated the direct path but not the second half of the mediated path, as shown in Table 5 . Bootstrap test results showed that the interaction term between MPAI and gender significantly predicted PSQI (Coeff.= -0.0215, P<0.001), indicating that gender plays a moderating role in the relationship between MPAI and PSQI. Further simple slope analysis was performed, as shown in Figure 3 . The level of cell phone dependence was positively correlated with the sleep disturbance in both males (simple slope = 0.0616, t = 16.66, P< 0 .001) and females (simple slope = 0.0401, t = 9.51, P< 0 .001), revealing that as cell phone dependence increased, sleep disturbance worsened for both genders.

The current study has several limitations that need to be addressed. Firstly, the study only gathered questionnaires from medical college students in Jiangsu Province, China. To enhance the study's scope and generalizability, future research should consider expanding the survey to include a more diverse population. Secondly, the survey's reliance on personal subjective assessments may compromise the objectivity of the study results. Subsequent studies could benefit from incorporating objective assessment methods like clinician diagnosis and polysomnography (PSG) examination. Thirdly, the study did not thoroughly examine various sleep patterns such as onset latency or intermittent awakenings. To address this limitation, future research should conduct a comprehensive analysis of these aspects to deepen the understanding of factors affecting sleep quality in university students. Additionally, the online survey format hindered the accurate calculation of response rates, potentially biasing the sample toward individuals more inclined to complete the survey, thus affecting the generalizability of the findings, and we did not collect data on smoking status, alcohol and caffeine intake, and physical activity, all of which are important variables that may need to be adjusted. These factors could potentially have a significant impact on our study results. Despite these limitations, the results of this study remain significant. The findings reaffirm the pivotal role of cell phone addiction in predicting sleep disturbance among medical college students. Furthermore, the mediating effect analysis demonstrated that psychological resilience mitigated the impact of cell phone addiction on sleep disturbance. Additionally, the study elucidated the moderating role of gender as an individual characteristic in the relationship between cell phone addiction and sleep disturbance.