Psychological resilience can be defined as the capacity to cope with challenges. In adolescents, resilience involves response and protective factors they exercise when confronting difficulties. In the past two decades, resilience during adolescence has gained considerable research attention. However, most relevant studies in Taiwan have employed translated or revised non-Taiwanese scales with high reliability and validity to measure resilience among adolescents. Furthermore, few resilience scales targeting factors specific to junior high school students have been reported in Taiwan. The current study filled this gap by developing an adolescent resilience scale (ARS) and then verifying its reliability and validity. It also explored the adolescent resilience differences associated with individual variables such as age and gender.

We employed the qualitative and quantitative methods and developed the ARS in the three phases: an open-ended questionnaire, a pilot test, and a formal test. All three phases were included to ensure the reliability and validity of not only the pilot but also the final version of the ARS.

In the first phase, we applied convenience sampling to recruit 80 adolescents from two junior high schools in Northern Taiwan. After the exclusion of 18 questionnaires for incomplete data, we included 50 valid responses from 23 boys and 27 girls, with an 85% recovery rate. The questionnaire items included “What personality traits or abilities have helped you cope with difficulties?” “What strategies do you apply to solve problems, and why?” “What stops you from quitting learning when things get tough?” Next, we applied the inductive method to generate items from the obtained responses. Keywords and significant phrases related to adolescent resilience were summarized and generalized. They were then categorized into six factors: Show of Empathy (putting yourself in someone else’s position and understanding the difficulty of others), Problem-Solving Competency (hard work, persistence, responsibility, and problem-solving), Self-Emotion Regulation (calmness and self-encouragement), Establishment of Life Goals (living up to expectations, planning for the future, and goal achievement), Competence Perception (independence, self-confidence, strength, ability to face challenges, and bravery), and Awareness of Positive Support (encouragement from friends and parents and help from teachers). We used the extracted keywords and phrases to draft a pilot ARS questionnaire, containing 53 items, rated on a 5-point Likert scale–with 1 indicating “completely disagree” and 5 indicating “completely agree”; higher the score, the higher was the level of resilience.

In the second phase, the pilot ARS questionnaire was distributed to 200 grade 8 and 9 students in a junior high school in Taiwan based on convenience sampling. Consequently, we obtained 145 valid samples from 52 grade 8 students and 93 grade 9 students (recovery rate = 83.8%; 69 boys and 76 girls). The data were analyzed using SPSS Amos (version 26; IBM). We performed item discrimination analysis, exploratory factor analysis (EFA), and reliability analysis to evaluate the internal consistency reliability and validity of the questionnaire. Item analysis was performed using corrected item–total correlation and the significance of the mean item difference of the top 27% and bottom 27% of the data was determined using a t test. Two items with a critical ratio of < 3 and one item with item–total correlation of < 0.4 were excluded to increase the overall Cronbach’s α. Our EFA results confirmed the factor structure of the ARS. A high Kaiser–Meyer–Olkin (KMO; 0.922) indicated the suitability of the data for EFA, and a χ2 of 6173.745 (p < 0.001) proved the factorability of the data. In the EFA, all 23 items were subjected to principal component analysis, and only the final six factors were able to explain 66.363% of the total variance. After repeated EFA, all 23 items (representing all six factors; KMO = 0.918, χ2 = 2227.545, p < 0.001) could explain 74.489% of the total variance. The factor loadings ranged from 0.562 to 0.816, indicating that all items were good measures of their respective factors. All factors were internally consistent (α = 0.946) as well.

In the third phase, we distributed 1,600 of the final questionnaire to grade 7–9 students from five junior high schools. We collected 1,513 valid responses from 546 grade 7, 419 grade 8, and 548 grade 9 students (recovery rate = 94.6%; 730 boys and 783 girls). The data were analyzed through confirmatory factor analysis (CFA) and cross-validation to verify the validity of the final version of the ARS. The data of the 1513 samples were randomly split in half. The first half (N = 756), considered an estimate sample, was subjected to CFA, whereas the second half (N = 757), considered a validity sample, was subjected to cross-validation. The CFA results for all items indicated that the model demonstrated a good fit [comparative fit index (CFI) = 0.95, root mean square error of approximation (RMSEA) = 0.05, Tucker–Lewis index (TLI) = 0.94, goodness of fit index (GFI) = 0.93] and was free from offending estimates. The internal consistency estimates in scale construction were satisfactory. The correlation analysis demonstrated a consistent pattern of relationships, suggesting convergent and discriminant validity. Next, to confirm the stability of the scale for different samples, the second half of the sample (N = 757) was analyzed through crossvalidation. The results demonstrated a good data fit in the resilience model (χ2 = 726.844, p < 0.001; CFI = 0.93, RMSEA = 0.05, TLI = 0.91, GFI = 0.92). As such, both CFA and EFA indicated that the 23-item ARS, with a 6-factor structure, has good construct validity for our single factor model.

Significant age differences were observed in the overall ARS scores and scores on the Show of Empathy, Self-Emotion Regulation, and Establishment of Life Goals subscales. Specifically, the grade 7 students scored higher on the subscales than did the grade 8 and 9 students. Although no sex differences were found on the overall ARS score, the girls scored significantly higher on the Show of Empathy subscale, whereas the boys scored significantly higher in the Self-Emotion Regulation subscale. Moreover, the Show of Empathy subscale scores did not increase with age; nevertheless, it was the highest in the girls in grade 9 but the lowest in the boys in grade 8. The Self-Emotion Regulation subscale scores decreased with aged in the girls but not in the boys; moreover, the Self-Emotion Regulation scores were the highest in boys in the grade 7 boys but the lowest in the grade 9 girls.

In conclusion, our results indicated that our 23-item 6-domain ARS is a reliable and valid instrument for the measurement of adolescent resilience and these subfactors: show of empathy, problem-solving competency, self-emotion regulation, establishment of life goals, competence perception, and positive support awareness. The ARS’s psychometric properties, structural validity and internal consistency reliability, were explored. The results demonstrated that the ARS has great reliability and validity and thus is recommended for analyzing adolescent resilience. In the present study, we focused on the effects of individual traits related to resilience. Future studies including environmental factors, such as those related to peers, family, school, community, and society, which also influence adolescent resilience, are warranted. Moreover, these studies may involve individual personalities as inner protective factors and add external protective factors and risk factors for a more comprehensive understanding of adolescent resilience. Other future research directions include expanding the study scope to participants from different cultures, age groups, or socioeconomic backgrounds. Moreover, in future, we will expand on the current results by clarifying the effects of local Chinese cultures on adolescent resilience.