The relationship between reflective ability and professional identity: the mediating effect of self-directed learning and self-efficacy for junior clinical nurses Authors: Lijie Huang, Xiangsheng Zhang, Feijie Wang, Songbo Zhang, Xiaoxia Chang, Yinping Chu, Lu Wang, Wenwen Jia, Beibei Zhang Published: 2024 ABSTRACT BackgroundNursing is crucial for delivering healthcare services; however, nursing shortages pose a significant global challenge, which the high turnover rate of junior nurses is expected to exacerbate. Enhancing nurses’ professional identities is likely to be an effective strategy for reducing turnover rates.MethodsWe recruited 865 junior clinical nurses from three tertiary hospitals in Henan Province, China. The participants completed a demographic questionnaire along with the Reflective Ability Scale for Clinical Nurses, the Nursing Profession Self-efficacy Scale, the Nursing Staff Self-learning Ability Evaluation Scale, and the Professional Identity Scale for Nurses. The relationships between reflective ability, self-directed learning, self-efficacy, and professional identity were evaluated using SPSS26.0, with PROCESS version 3.5 used to construct a chain-mediation model.ResultsPositive correlations were found between reflective ability, self-directed learning, self-efficacy, and professional identity. Chain mediation model analysis showed that reflective ability, self-directed learning, and self-efficacy directly and positively affected professional identity (β = 0.22, β = 0.30, β = 0.66, all p < 0.001). Self-directed learning and self-efficacy exerted a significant chain mediating effect between reflective ability and professional identity (95% confidence interval [CI]: 0.06–0.13). The total effect of reflective ability on the professional identity of clinical nurses was 0.51 (p < 0.001, 95%CI: 0.42–0.59), the direct effect was 0.22 (p < 0 0.001, 95%CI: 0.14–0.29), and the indirect effect was 0.29 (p < 0.001, 95%CI: 0.22–0.37).ConclusionSelf-directed learning and self-efficacy partially mediated the relationship between reflective ability and professional identity. Multifaceted strategies targeted at enhancing reflective ability, self-directed learning, and self-efficacy are recommended to improve junior clinical nurses’ professional identities. Introduction Nurses comprise the largest professional group in most countries, accounting for approximately half of the global healthcare workforce [1], and they play a pivotal role in delivering high-quality healthcare services [2]. Despite their recognized importance in healthcare delivery, persistent nurse shortages continue to challenge health systems worldwide, which a high turnover rate among nurses is anticipated to exacerbate even further [3, 4]. In particular, the large number of junior nurses with high turnover rates warrants attention from healthcare administrators and policymakers [5]. Professional identity, which reflects individuals’ understanding of their profession’s societal impact and the significance of their work [6], serves as a crucial psychological foundation for effectively fulfilling duties and maintaining organizational commitment [7]. It is also considered to be a key factor in reducing nurses’ turnover intention [8]. Sabanciogullari et al. [9] reported that a reduced sense of professional identity among nurses could predict higher turnover intentions in Turkey. Similarly, professional identity is a predictive factor of nurses’ intentions to stay in clinical roles in China [8]. Therefore, identifying the various factors that influence professional identity is likely to be of considerable importance for developing strategies to reduce turnover intentions among clinical nurses. Reflection involves a cognitive process through which individuals examine their distinctive experiences gained through interactions, infusing these experiences with meaning for thought and action while actively developing practical knowledge [10]. A growing body of literature has shown that student nurses’ awareness of their own reflective processes, along with the integration of reflection and scientific theories of care, can enhance their capacity to care for individuals experiencing distress [11]. Reflection has also been shown to improve emergency department nurses’ communication skills when they interact with patients [12]. Zarrin [13] reported that improving nurses’ reflective abilities enhanced their empowerment and professionalism, such that their improved sense of empowerment then strengthened their professional identity [14]. Nishimoto further confirmed that reflection is a direct predictor of nurses’ professional identity [10]. Therefore, we propose the following hypotheses: Hypothesis 1 Reflective ability significantly predicts professional identity. Self-directed learning is a personal endeavor wherein individuals take the initiative to identify their learning requirements, establish goals, find resources, devise and execute appropriate strategies, and assess the results of their learning while retaining the option of seeking external guidance or assistance [15]. This process encompasses various elements, including self-observation, social interaction, motivation, planning, and execution. Alves and Gazzola [16] identified self-directed learning as a key factor influencing the development of professional identity among counselors, and it has been shown to contribute significantly to the formation of professional identity [17]. Furthermore, reflection promotes the growth of self-directed learning [18], and is positively associated with it [19]. Hypothesis 2 Self-directed learning mediates the relationship between reflective ability and professional identity. Self-efficacy refers to an individual’s confidence in their ability to perform the tasks necessary to achieve specific goals. It reflects a belief in one’s capacity to manage motivation, behavior, and interactions within a social context, and plays a pivotal role in motivating individuals and achieving desired outcomes [20]. In nursing, self-efficacy shapes beliefs, actions, and behaviors when caring for patients, and is a key determinant of career success [21]. Increased self-efficacy among nurses is linked to improved patient care quality, as nurses with higher self-efficacy consistently demonstrate superior job performance [22]. As a fundamental aspect of the nursing profession, self-efficacy is closely tied to nursing actions and has been recognized in the literature as influencing nurses’ professional identity [23]. Additionally, nurses’ reflection is positively correlated with self-efficacy [13]. Therefore, we propose the following hypotheses: Hypothesis 3 Self-efficacy mediates the relationship between reflective ability and professional identity. Research has shown that reflection, self-directed learning, and self-efficacy are positively correlated with professional identity [17, 24, 25]. Self-directed learning has also been identified as a predictor of self-efficacy among nursing students [26]. Reflection helps individuals make sense of their experiences, potentially deepening their professional identities. Self-directed learning and self-efficacy may further enhance the reflective processes. Hypothesis 4 Self-directed learning and self-efficacy jointly mediate the relationship between reflective ability and professional identity. The current literature suggests that reflective ability, self-directed learning, and self-efficacy independently influence nurses’ professional identities. However, few studies have examined how reflective ability affects professional identity through the mediating effects of self-directed learning and self-efficacy. This study aimed to explore the interactions among these four factors within a single model and to verify the mediating effects of self-directed learning and self-efficacy. Materials and methods Aims The primary objective of this study was to investigate the relationship between reflective ability and professional identity among Chinese junior clinical nurses and the impact of self-directed learning and self-efficacy on this relationship through a chained mediation model analysis. Design In this study, we employed a cross-sectional correlational research design and followed the STROBE (Strengthening the Reporting of Observational Studies in Epidemiology) checklist to ensure the rigor and quality of our investigation. Participants From January to March 2023, convenience sampling was used to recruit junior clinical nurses from three tertiary grade A hospitals in Henan Province, China. The inclusion criteria for participants were as follows: (a) successfully passed the Chinese nurses practice examination, completed registration, and currently practicing as clinical nurses; (b) having worked ≤5 years in the current hospital; (c) possessing a nurse qualification certificate; and (d) willing to participate in the study (written informed consent provided). Exclusion criteria included training nurses, assistant nurses, and nurses on leave for more than 3 months. According to the recommended criteria, the sample size should be 5 to 10 times larger than the total number of variables, with an allowance for a 20% invalid response rate [27]. Considering that the study considered 107 variables, including gender, age, educational level, working years, and title (5 variables) in a questionnaire, and 19, 34, 19, and 30 variables (items), respectively, from four scales, the calculated sample size ranged from 642 to 1284 participants. A total of 865 participants from three hospitals in Henan Province, China completed the survey, with 853 valid responses, resulting in an effective response rate of 98.6%. Data collection Permission to conduct the study was obtained from the relevant hospital nursing department before the investigation. On the Wen Juan Xing platform (https://www.wjx.cn/), where the questionnaires for the survey were uploaded, quick-response (QR) codes were generated to access the questionnaires, which were distributed by the ward head nurses, as designated by the nursing department. Nurses could scan the QR codes and complete the questionnaires through WeChat, a widely used social networking application in China. Respondents were required to complete the online survey within one week while following the questionnaire instructions. The researcher downloaded the completed questionnaires from the platform. To ensure data accuracy, it was mandatory for each question to be answered and participants were allowed only one opportunity to submit their responses. Variable measures General demographic questionnaire General demographic information, including gender, age, educational level, working years, and professional title, was obtained using a self-designed questionnaire. The reflective ability scale for clinical nurses The Reflective Ability Scale for Clinical Nurses [10] was developed to assess the reflective ability of clinical nurses. Shao et al. [28] translated the scale into Chinese and administered it to clinical nurses, demonstrating that it had good reliability and validity. This tool consists of 19 items in three subscales: recall their own nursing practice (8 items), reflect on their own nursing practice (6 items), and expand their own nursing practice (5 items). In this 6-point Likert scale, items are scored from 1 (“totally disagree”) to 6 (“totally agree”), with higher scores indicating better reflective ability. In the present study, Cronbach’s alpha for the overall scale was 0.98. The nursing staff self-learning ability evaluation scale The Nursing Staff Self-learning Ability Evaluation Scale was developed by Xiao and Li [29] and includes 34 items and four principal components: self-motivational beliefs (14 items), task analysis (6 items), self-monitoring and regulation (10 items), and self-evaluation (4 items). This 5-point Likert scale scores items through total scores ranging from 34 to 170 points, and the higher the score, the stronger the nurse’s self-directed learning competence. This scale has shown excellent reliability and validity and can be used to objectively measure nurses’ self-directed learning competence. In this study, Cronbach’s alpha for the overall scale was 0.96. The nursing profession self-efficacy scale The Nursing Profession Self-efficacy Scale was developed by Caruso [30] based on Bandura’s social cognitive theory and has been translated by Zhang [31]. This scale consists of 19 items categorized into two subscales. The two subscales assess nursing practice status (12 items) and professional nursing status (seven items). This 5-point Likert scale scores items ranging from 1 (“no confidence at all”) to 5 (“full confidence”), and the higher the score, the higher the self-efficacy level. In the present study, Cronbach’s alpha for the total scale was 0.97. The professional identity scale for nurses The Professional Identity Scale for Nurses was developed by Liu in Mandarin [32] to evaluate the professional identity of nurses. This scale comprises 30 items categorized into five subscales: occupational cognitive evaluation, occupational social support, professional social skills, coping with career setbacks, and professional self-reflection. This 5-point Likert-scale scores items ranging from 1 (“totally disagree”) to 5 (“totally agree”), with the total final score ranging from 30 to 150. Higher scores indicate a higher level of professional identity. A score of 30–60 indicates low professional identity, 61–90 indicates lower professional identity, 91–120 indicates medium professional identity, and 121–150 indicates a high level of professional identity. In the present study, Cronbach’s alpha for the overall scale was 0.98. Statistical analysis Data analysis was performed using IBM SPSS Statistics for Windows, version 26.0. Count data are expressed as frequency and percentage (%), and comparisons between groups were conducted using t-tests and variance analysis. Normality tests were performed on the measured data. Data conforming to normal distribution are presented as mean ± standard deviation \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:\left(equation\right)$$\end{document}. The Pearson coefficient was used to analyze the relationships among reflective ability, self-efficacy, self-directed learning, and professional identity. The same measurement environment, data source, or evaluator can introduce errors, which is known as common method bias. A common method bias test was conducted to assess the possibility of bias. Pearson’s correlation analysis was performed to examine the relationships between reflective ability, self-efficacy, self-directed learning, and professional identity. Three models were used to evaluate the chained mediation effects of self-efficacy and self-directed learning on the relationship between reflective ability and professional identity. Model 1 included professional title and reflective ability as independent variables and self-directed learning as the dependent variable. Model 2 incorporated professional title, reflective ability, and self-directed learning as independent variables and self-efficacy as the dependent variable. Model 3 included professional title, reflective ability, self-directed learning, and self-efficacy as independent variables, and professional identity as the dependent variable. Model 6 in PROCESS version 3.5 was employed, with reflective ability as the independent variable, professional identity as the dependent variable, and self-directed learning and professional self-efficacy as mediating variables, to construct a structural equation model. The mediating effect was evaluated using the bias-corrected nonparametric percentile bootstrap method, with 5,000 sampling repetitions. A mediating effect was confirmed if the 95% confidence interval (CI) did not include 0. In this study, a p-value of less than 0.05, derived from a two-sided test, was considered statistically significant. Results Descriptive statistics A total of 853 junior clinical nurses participated in this study. Their ages ranged from 22 to 35 years, with an average age of 26.43±1.77 years; 820 (96.1%) were women and 33 (3.9%) were men. The correlation between age and professional identity was r = 0.002 (p > 0.05). Detailed general information is given in Table 1. Table 1The detailed general information of clinical nurses(N)Reflective abilitySelf-directed learningSelf-efficacyProfessional identityScoreP-valueScoreP-valueScoreP-valueScoreP-value Gender Male (33)86.06±17.330.80128.09±22.650.7475.88±11.950.43113.70±21.400.54Female (820)86.74±14.74126.95±19.2777.43±10.97111.48±20.08 Working years ≤3 (489)86.15±15.360.20126.57±19.310.4576.81±11.390.08110.79±20.180.204–5 (364)87.48±14.08127.58±19.5378.12±10.44112.57±19.85 Educational level College degree or below (177)87.51±13.450.43126.36±18.560.6977.06±11.000.61110.76±19.740.20Undergraduate (668)86.57±15.22127.22±19.7177.49±11.07111.92±20.26Master’s degree or above (8)81.13±10.06122.38±9.5974.00±4.6999.88±12.98 Title Primary title (470)86.40±14.890.49126.27±19.620.2276.95±11.550.22110.26±20.690.04Intermediate title (383)87.111±14.78127.90±19.1177.88±10.30113.13±19.14 Table 1The detailed general information of clinical nurses(N)Reflective abilitySelf-directed learningSelf-efficacyProfessional identityScoreP-valueScoreP-valueScoreP-valueScoreP-value Gender Male (33)86.06±17.330.80128.09±22.650.7475.88±11.950.43113.70±21.400.54Female (820)86.74±14.74126.95±19.2777.43±10.97111.48±20.08 Working years ≤3 (489)86.15±15.360.20126.57±19.310.4576.81±11.390.08110.79±20.180.204–5 (364)87.48±14.08127.58±19.5378.12±10.44112.57±19.85 Educational level College degree or below (177)87.51±13.450.43126.36±18.560.6977.06±11.000.61110.76±19.740.20Undergraduate (668)86.57±15.22127.22±19.7177.49±11.07111.92±20.26Master’s degree or above (8)81.13±10.06122.38±9.5974.00±4.6999.88±12.98 Title Primary title (470)86.40±14.890.49126.27±19.620.2276.95±11.550.22110.26±20.690.04Intermediate title (383)87.111±14.78127.90±19.1177.88±10.30113.13±19.14 The detailed general information of clinical nurses Common method bias test results All items were analyzed using Harman’s single-factor test. Ten factors were identified with eigenvalues greater than 1, with the first factor accounting for 39.93% of the explained variance, which is below the 40% threshold. This finding indicated that common method bias was not a significant issue in this study. Pearson’s correlation results in relation to reflective ability, self-efficacy, self-directed learning, and professional identity Reflective ability was positively correlated with self-directed learning, professional self-efficacy, and professional identity; self-directed learning was positively correlated with professional self-efficacy and identity; and self-efficacy was positively correlated with professional identity (see Table 2). Table 2Correlation coefficients among reflective ability, self-directed learning, self-efficacy and professional identityVariables12341.Reflective ability-2.Self-directed learning0.32**-3.Self-efficacy0.34**0.63**-4.Professional identity0.38**0.57**0.600**-**The correlation is significant at the 0.01 significance level, (1, 2, 3, 4 represents reflective ability, self-directed learning, self-efficacy and professional identity respectively) Table 2Correlation coefficients among reflective ability, self-directed learning, self-efficacy and professional identityVariables12341.Reflective ability-2.Self-directed learning0.32**-3.Self-efficacy0.34**0.63**-4.Professional identity0.38**0.57**0.600**-**The correlation is significant at the 0.01 significance level, (1, 2, 3, 4 represents reflective ability, self-directed learning, self-efficacy and professional identity respectively) Correlation coefficients among reflective ability, self-directed learning, self-efficacy and professional identity **The correlation is significant at the 0.01 significance level, (1, 2, 3, 4 represents reflective ability, self-directed learning, self-efficacy and professional identity respectively) Chained mediation analysis results As shown in Table 3, the findings indicated that reflective ability had a notably positive impact on self-directed learning (β = 0.32, p < 0.001) in model 1 and a notably positive impact on self-efficacy (β = 0.16, p < 0.001) in model 2. Self-directed learning positively predicted self-efficacy (β = 0.58, p < 0.001) in model 2. Additionally, reflective ability (β = 0.16, p < 0.001), self-directed learning (β = 0.29, p < 0.001), and self-efficacy (β = 0.37, p < 0.001) all significantly contributed to the prediction of professional identity in model 3. Table 3Chained mediation models of self-efficacy and self-directed learning in the relationship between reflective ability and professional identity (n = 853)Predictor variableOutcome variable R R 2 FβtBoot LLCIBoot ULCI Model 1 TitleSelf-directed learning0.320.1049.01***0.020.49-1.232.06Reflective ability0.329.890.340.50 Model 2 TitleSelf-efficacy0.650.42201.26***-0.01-0.49-0.940.57Reflective abilitySelf-directed learning0.160.585.5920.870.070.300.160.36 Model 3 TitleProfessional identity0.670.44168.87***0.041.68-0.192.49Reflective ability0.165.790.140.29Self-directed learning0.298.600.230.36Self-efficacy0.3710.900.550.79***p < 0.001. R represents the relationship between independent and dependent variables; R² represents the extent to which the independent variables explain the variance in the dependent variable; The F-statistic is used to test the overall validity of the model; β represents the expected change in the dependent variable for each one-unit change in the independent variable; Boot LLCI: Lower Limit of Confidence Interval, Boot ULCI: Upper Limit of Confidence Interval Table 3Chained mediation models of self-efficacy and self-directed learning in the relationship between reflective ability and professional identity (n = 853)Predictor variableOutcome variable R R 2 FβtBoot LLCIBoot ULCI Model 1 TitleSelf-directed learning0.320.1049.01***0.020.49-1.232.06Reflective ability0.329.890.340.50 Model 2 TitleSelf-efficacy0.650.42201.26***-0.01-0.49-0.940.57Reflective abilitySelf-directed learning0.160.585.5920.870.070.300.160.36 Model 3 TitleProfessional identity0.670.44168.87***0.041.68-0.192.49Reflective ability0.165.790.140.29Self-directed learning0.298.600.230.36Self-efficacy0.3710.900.550.79***p < 0.001. R represents the relationship between independent and dependent variables; R² represents the extent to which the independent variables explain the variance in the dependent variable; The F-statistic is used to test the overall validity of the model; β represents the expected change in the dependent variable for each one-unit change in the independent variable; Boot LLCI: Lower Limit of Confidence Interval, Boot ULCI: Upper Limit of Confidence Interval Chained mediation models of self-efficacy and self-directed learning in the relationship between reflective ability and professional identity (n = 853) Reflective ability Self-directed learning 0.16 0.58 5.59 20.87 0.07 0.30 0.16 0.36 ***p < 0.001. R represents the relationship between independent and dependent variables; R² represents the extent to which the independent variables explain the variance in the dependent variable; The F-statistic is used to test the overall validity of the model; β represents the expected change in the dependent variable for each one-unit change in the independent variable; Boot LLCI: Lower Limit of Confidence Interval, Boot ULCI: Upper Limit of Confidence Interval The results of the chain-mediation model are shown in Fig. 1; Table 4. The results showed that none of the 95% CIs included zero and that the model exhibited partial mediation effects. The cumulative indirect effect encompassing the three distinct pathways was 0.29 (95%CI: 0.22, 0.37), accounting for 56.86% of the overall effect. The combined mediating effect of self-directed learning and self-efficacy accounted for 31.03% of the total indirect effect (Table 4). Fig. 1The chained mediation model (n = 853), ∗∗∗p < 0.001. e1-e6 represent the Path Coefficient: a numerical representation of the strength and direction of the relationships between variables in the model The chained mediation model (n = 853), ∗∗∗p < 0.001. e1-e6 represent the Path Coefficient: a numerical representation of the strength and direction of the relationships between variables in the model Table 4The direct and indirect effects in chained mediation modelEffectSELLCIULCIRatio of indirect to total effect (%)Total effect0.510.040.420.59-Direct effect0.220.040.140.29-Total indirect effect0.290.040.220.3756.86 \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$X \to M1 \to Y$$\end{document} 0.120.020.090.1823.53 \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$X \to M2 \to Y$$\end{document} 0.080.020.040.1115.69 \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$X \to M1 \to M2 \to Y$$\end{document} 0.090.020.060.1317.65X: Reflective ability; M1: Self-directed learning; M2: Self-efficacy; Y: Professional identity Table 4The direct and indirect effects in chained mediation modelEffectSELLCIULCIRatio of indirect to total effect (%)Total effect0.510.040.420.59-Direct effect0.220.040.140.29-Total indirect effect0.290.040.220.3756.86 \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$X \to M1 \to Y$$\end{document} 0.120.020.090.1823.53 \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$X \to M2 \to Y$$\end{document} 0.080.020.040.1115.69 \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$X \to M1 \to M2 \to Y$$\end{document} 0.090.020.060.1317.65X: Reflective ability; M1: Self-directed learning; M2: Self-efficacy; Y: Professional identity The direct and indirect effects in chained mediation model X: Reflective ability; M1: Self-directed learning; M2: Self-efficacy; Y: Professional identity Discussion The results indicated that reflective ability positively influenced the junior clinical nurses’ professional identity, supporting Hypothesis 1. Stronger reflective ability was associated with a higher level of professional identity. Reflection serves as a key internal motivator for enhancing professional identity, enabling individuals to strengthen their sense of identity by reflecting on their actions, values, and attitudes [33]. Reflection fosters personal insights and offers an alternative perspective on patient interaction, providing valuable insights into patient care. It also reinforces professional values when integrated into clinical practice, potentially enhancing the development of professional identity [34, 35]. Reflection is considered a prerequisite for developing a high level of professional identity [36], and reflective ability can be improved through training [37, 38]. Implementing effective strategies to enhance reflective abilities can significantly improve professional identities. The findings revealed that self-directed learning mediated the relationship between reflective ability and professional identity, supporting Hypothesis 2. The mediating effect accounted for 23.53%. Reflection and self-directed learning play significant roles in enhancing professional identity [16, 39]. Self-directed learning serves as a bridge between reflective ability and professional identity, contributing to the development of nurses’ professional identity. This finding aligns with previous studies that show that reflective practices significantly enhance students’ self-directed learning skills [40] and that self-directed learning is a key factor in shaping professional identity [41]. Additionally, the results revealed that self-efficacy partially mediated the relationship between reflective ability and professional identity, with a mediating effect accounting for 15.69%, thus supporting Hypothesis 3. Stronger reflective ability has been associated with higher self-efficacy and professional identity. Reflection has been shown to predict self-efficacy positively. Bandura emphasized the role of self-reflection in reshaping an individual’s cognition and behavior, identifying it as a cornerstone of the self-efficacy concept [42]. The positive correlation found between self-efficacy and professional identity aligns with the findings of Cai’s study [43]. Additionally, self-efficacy has been found to mediate the relationship between the clinical learning environment and professional identity among nursing students [44]. Self-efficacy enhances nursing competence and improves the quality of care, enabling nurses to perform their duties more effectively. This, in turn, fosters a sense of achievement, deepens passion for the profession, and strengthens professional identity [45]. These findings support the idea that interventions aimed at increasing self-efficacy can promote the development of professional identity. This study showed that reflective ability and professional identity were mediated by self-directed learning and self-efficacy, accounting for 17.65% of the total effect, thus supporting Hypothesis 4. Self-directed learning and self-efficacy enhance professional identity. As nurses develop stronger reflective abilities, self-directed learning and self-efficacy improve, leading to an increased sense of professional identity. Reflection is deeply intertwined with personal understanding and growth [11], and this continuous process of assimilation transforms junior clinical nurses both professionally and personally. Reflection fosters a deeper understanding of caring, which is the cornerstone of nursing and is essential for becoming a compassionate nurse. This cognitive process contributes to the development of high-quality nursing practice [45]. Nurses with higher levels of self-directed learning and self-efficacy tend to exhibit higher levels of performance and accomplishment [46], which, in turn, strengthens their professional identities. Limitations Although this study boasts certain strengths, it has some limitations, and caution should be exercised when interpreting its findings. First, it employed a cross-sectional research design that precluded the establishment of causal relationships. Second, as the data were gathered from a province in central China, generalizing the findings to other regions within or outside China warrants caution. Finally, while the mediating effects of self-directed learning and self-efficacy accounted for approximately 56.86% of the total effect in this study, further research is needed to explore additional mediating factors to provide a more comprehensive understanding. Conclusion This study is the first to use a chain-mediation model to examine the intricate relationships between reflective ability, self-directed learning, self-efficacy, and professional identity. The findings highlighted the critical role of reflective ability in fostering self-directed learning and self-efficacy, both of which are essential for developing a strong professional identity among junior clinical nurses. This suggests that interventions aimed at enhancing reflective ability could have a significant cascading effect, fostering greater self-directed learning and self-efficacy and ultimately leading to the development of a strong professional identity. Therefore, nursing managers should provide junior nurses with systematic professional development training programs that cover reflective practice, self-directed learning, and self-efficacy enhancement, thereby improving their professional identity. Nursing managers must provide effective guidance during the implementation of such training programs and offer positive directions and reinforcement. Specifically, they need to encourage junior nurses to develop their own reflective approaches and engage in regular self-reflection in the workplace. They also need to offer constructive feedback that highlights both the achievements of nurses and areas for growth that can reinforce nurses’ intrinsic motivation for learning and their sense of intrinsic value, which would help foster a supportive environment that encourages the continuous development of nurses’ professional identity.