Reliability and validity of Turkish version of stress resilience item bank 10 item short form

Abstract

Background

The aim of this study was to adapt the Stress Resilience Item Bank 10-Item Short Form, originally developed in German, into Turkish for patients diagnosed with psychosomatic disorders, and to evaluate its psychometric properties, validity, and reliability.

Method

The study included 507 patients aged 18–65 years whose conditions were historically categorized under psychosomatic disorders and who were diagnosed with Mental Factors Affecting Other Medical Conditions according to DSM-5 criteria, along with 74 healthy controls. Sociodemographic data were collected using a structured data form completed by patients and, when applicable, their relatives. Participants completed the Stress Resilience Item Bank 10-Item Short Form, the Connor-Davidson Resilience Scale, the Brief Resilience Scale, and the ICD-10 Symptom Rating Scale.

Results

The mean age of the patients was 45.9 ± 15 years, and 74.2% were female. In the reliability analysis, Cronbach’s alpha was 0.79. Test–retest reliability indicated a high correlation between the two measurements (r = .675, p < .001). In validity analyses, exploratory factor analysis yielded a two-factor structure explaining 49.2% of the total variance. Confirmatory factor analysis supported this structure, with acceptable model fit indices (RMSEA = 0.062, CFI = 0.942, GFI = 0.962). The Stress Resilience Item Bank 10-Item Short Form showed significant correlations with the Connor-Davidson Resilience Scale and the Brief Resilience Scale (r = .73, p < .001; r = .67, p < .001, respectively).

Conclusion

The Stress Resilience Item Bank 10-Item Short Form, has been identified as a suitable, valid, and reliable measurement tool for use in the Turkish patient population with psychosomatic disorders, the majority of whom are women.

Introduction

Research indicates that stress plays a role in the onset and progression of various medical conditions [1–4]. Nevertheless, studies have shown that some individuals cope with stress more effectively than others [5–7]. To date, limited information is available regarding why certain individuals remain healthy in the face of adversity while others develop illness; resilience appears to be a concept that may help to address this gap [8]. Although the concept of stress resilience has long been recognized, the majority of publications on this topic have emerged over the past 10–20 years. This likely reflects the increasing interest in positive psychology, prevention, and mental health promotion [9].

Resilience may be associated with innate personality traits, and numerous risk and protective factors have been identified in this regard. Prominent individual protective factors include self-confidence, intelligence, internal locus of control, academic competence, problem-solving skills, a positive and optimistic outlook toward the future, positive mood, a sense of humor, and socialization [3]. Resilience varies across age groups and cultures; for example, Asian cultures conceptualize resilience as awareness and acceptance of experience, whereas Western cultures tend to define it in terms of personal choice and environmental mastery [10–12].

Resilience is not solely a psychological phenomenon; it also exerts a strong influence on physical health, as it facilitates the rapid return of bodily functions (e.g., cardiovascular functioning) to baseline levels following adverse experiences [13]. In the literature, stress resilience has been reported to be lower in patients with psychosomatic disorders compared with healthy controls. This may be explained by the fact that resilience is influenced by factors such as social support, coping strategies in stressful situations, and self-esteem, all of which may be impaired in patients with psychosomatic disorders [4, 5]. In a study conducted among African Americans participants examining the independent relationship between stress resilience and psychosomatic complaints, it was shown that as resilience increased, reported stress decreased, and as reported stress increased, psychosomatic complaints also increased [14]. Another study demonstrated that higher stress resilience was associated with psychosomatic well-being and that positive emotions played a mediating role in this relationship [15]. In a systematic review examining the relationship between physical illness and resilience, physical illness was considered a potential adversity following which an individual may or may not demonstrate resilience. The review further suggested that physical illness may function as a variable that either enhances or diminishes resilience following an adverse event [16].

Numerous scales have been developed in the field of resilience; however, due to variations in context, setting, and the nature of threats or events, it is difficult to develop a general scale applicable to all situations [6, 7]. In a review conducted by Salisu et al., which examined frequently used resilience scales according to their psychometric properties, the Connor–Davidson Resilience Scale (CD-RISC) and the 10-item Connor–Davidson Resilience Scale (CD-RISC-10) were identified as the most widely used instruments due to their favorable validity and reliability properties [7]. In a methodological review by Windle et al., 19 resilience scales were evaluated, and a comprehensive set of quality criteria was established. Overall, the CD-RISC, the Resilience Scale for Adults, and the Brief Resilience Scale (BRS) received the highest scores; when all quality criteria were considered, the overall quality of the scales was regarded as moderately acceptable, and it was concluded that no gold-standard measure exists [6].

In 2018, during the development of the Stress Resilience Item Bank (SRIB), a 67-item long form and a 10-item short form were developed, both demonstrating good psychometric properties. This four-point Likert-type scale includes reverse-coded items, and its developers recommended its use across different cultures and populations [8]. Compared with Classical Test Theory (CTT), which has been employed in the development of existing resilience scales, Item Response Theory (IRT) is considered a more robust framework for applications such as test linking/equating, differential item functioning, and computerized adaptive testing [8, 17]. To date, IRT has been used in a limited number of studies aimed at developing or evaluating resilience measures, including the original development study of the SRIB [8, 18, 19].

In the present study, we aimed to translate the Stress Resilience Item Bank 10-Item Short Form (SRIB-10-SF) into Turkish and to examine whether it is suitable for use in a patient population with psychosomatic disorders. Unlike many previous scale development studies, this instrument was developed using a different statistical methodology and has demonstrated favorable psychometric properties. By evaluating its psychometric properties in a Turkish psychosomatic sample, we aimed to contribute to the existing literature in this field.

Methods

Study design and participants

The study group consisted of patients aged 18–65 years who presented to the Gazi University Faculty of Medicine Consultation-Liaison Psychiatry Outpatient Clinic between June 2021 and June 2022, whose conditions were historically evaluated under the heading of psychosomatic disorders, who were diagnosed with Mental Factors Affecting Other Medical Conditions according to DSM-5 criteria, who had a medical illness, and whose course, severity, treatment adherence, and prognosis were affected by psychological factors, and who met the inclusion criteria.

Patients with active suicidal ideation, dementia, intellectual disability, psychotic symptoms, or any condition associated with significant cognitive impairment were excluded from the study.

Before recruitment of the control group, hospital staff members (including physicians, nurses, secretaries, and other employees) were informed about the study. The control group consisted of healthy volunteers aged 18–65 years who were recruited from among the hospital staff. The control group was selected with the aim of achieving a comparable age and sex distribution to the patient group. All participants underwent a clinical psychiatric evaluation conducted by a psychiatrist, and individuals with any psychiatric disorder were excluded from the study. Patients were recruited consecutively from the outpatient clinic, whereas recruitment of healthy controls relied on voluntary participation and stricter exclusion criteria, resulting in a smaller number of eligible control participants.

Measures

Sociodemographic data form

This form was developed for the present study and completed using information obtained from interviews with participants in both the patient and control groups.

Stress Resilience Item Bank – 10 Item Short Form (SRIB-10-SF)

The SRIB-10-SF was developed by Obbarius et al. in 2018 using modern test theory methods and input from the target population [8]. Both a 67-item long form and a 10-item short form were developed and demonstrated strong construct validity and satisfactory psychometric properties. Twenty items in the long form and three items in the short form are reverse scored.

Connor–Davidson Resilience Scale (CD-RISC)

The CD-RISC was developed by Connor and Davidson in 2003 as a valid and reliable measure of resilience based on a sample of more than 1000 participants [20]. This 25-item self-report instrument is rated on a 5-point Likert scale. The Turkish validity and reliability study was conducted by Karaırmak, who demonstrated that the scale is valid and reliable in the Turkish population [21]. In the Turkish version, a Cronbach’s alpha coefficient of 0.92 was reported, and a three-factor structure was identified.

Brief Resilience Scale (BRS)

The BRS was developed by Smith et al. to assess the ability to “bounce back” from stress [22]. This 6-item self-report scale is rated on a 5-point Likert scale, includes reverse-scored items, and has a unidimensional structure. The Turkish validity and reliability study was conducted by Doğan, who demonstrated that the scale is valid and reliable in the Turkish population [23]. Internal consistency coefficients ranged from 0.80 to 0.91.

ICD-10 Symptom Rating (ISR-10)

The ICD-10 Symptom Rating (ISR-10) is a 29-item self-report instrument developed by Tritt et al. to assess psychological symptoms [24]. The Turkish adaptation study conducted by Kızılhan et al. demonstrated very good psychometric properties in the Turkish population [25]. Factor analyses demonstrated satisfactory construct validity. Reported Cronbach’s alpha coefficients ranged from 0.66 for the compulsive syndrome subscale to 0.93 for the total scale score. Patients were diagnosed with psychosomatic disorders based on the results of the clinical psychiatric interview. The ISR-10 was administered as a self-report measure to assess the severity and frequency of psychological symptoms. The ISR-10 was used solely for symptom assessment and was not employed as a diagnostic instrument.

Ethical approval

Ethical approval was obtained from the Ethics Committee of Gazi University (Decision No: E-77082166-604.01.02-72639, dated 12 April 2021). Written informed consent was obtained from all participants prior to their inclusion in the study.

Translation and cultural adaptation of the SRIB-10-SF

Permission to conduct the validity and reliability study was obtained from the original developers of the scale. The short form was selected because of its practical applicability. The translation process was conducted in accordance with the FACIT translation methodology [26]. A forward–backward translation procedure was employed. First, the scale was independently translated into Turkish by two native Turkish-speaking translators who were fluent in German. Subsequently, the translations were evaluated by three independent reviewers with expertise in mental health. The Turkish version was subsequently back-translated into German by two independent bilingual translators who had not seen the original German version. After comparison of the back-translated and original German versions, the Turkish translation was reviewed and revised accordingly. Following the necessary revisions, the pre-final version was administered to 30 individuals from diverse sociocultural backgrounds who were not included in the main study sample. Participants were asked to evaluate the scale in terms of comprehensibility, readability, and item format. Based on participant feedback, final modifications were made and the scale was finalized.

Statistical analysis

Statistical analyses were performed using SPSS version 24 and AMOS version 21. Differences in sociodemographic and clinical characteristics between the patient and control groups were assessed using independent-samples t-tests for continuous variables and chi-square tests for categorical variables. The normality of continuous variables was assessed using the Kolmogorov–Smirnov test.

For reliability analyses, internal consistency was assessed using Cronbach’s alpha coefficient. Item–total correlation analyses were conducted, and Cronbach’s alpha coefficients were recalculated following the deletion of each item. To assess test–retest reliability, the SRIB-10-SF was re-administered to 55 volunteers from the control group 2–4 weeks after the initial assessment, and the correlation coefficient between the two administrations was calculated.

To evaluate construct validity, exploratory factor analysis (EFA) was performed. The suitability of the data for factor analysis was assessed using the Kaiser–Meyer–Olkin (KMO) measure and Bartlett’s Test of Sphericity. Principal component analysis with varimax rotation was performed, and factors with eigenvalues ≥ 1 were retained. Following EFA, confirmatory factor analysis (CFA) was conducted to evaluate the identified factor structure. Model fit was initially evaluated using the chi-square (χ²) goodness-of-fit statistic. After evaluating overall model fit, standardized regression coefficients of the scale items were calculated. To assess convergent validity, the CD-RISC and BRS were administered to all participants, and Pearson correlation analyses were conducted to examine the relationships among the scales.

To assess discriminant validity, independent-samples t-tests were performed to compare total SRIB-10-SF scores between patients with psychosomatic disorders and healthy controls and to determine whether the scale differentiated between clinical and non-clinical samples.

Results

Sociodemographic characteristics

A total of 507 patients and 74 healthy controls were included in the study, yielding a total sample of 581 participants. The mean age of the participants was 44.6 ± 14.9 years. The sociodemographic characteristics of the sample are presented in Table 1.

Table 1.

Sociodemographic Characteristics of the Sample Group

Category	Patients	Controls	Total
Age in years (mean ± SD)	45.9 ± 15	35.7 ± 10.6	44.6 ± 14.9
Number	507 (87.3%)	74 (12.7%)	581
Sex
Female	376 (74.25%)	44 (59.5%)	420 (72.3%)
Male	131 (25.8%)	30 (40.5%)	161 (27.7%)
Education Level
University	154 (30.4%)	50 (67.6%)	204 (35.1%)
High School	157 (31%)	17 (23%)	174 (29.9%)
Middle School	57 (11.2%)	7 (9.4%)	57 (9.8%)
Primary School	114 (22.5%)	0 (0%)	121 (20.8%)
No formal education / Literate only	25 (4.8%)	0 (0%)	25 (4.4%)
Employment Status
Student	47 (9.3%)	7 (9.5%)	54 (9.4%)
Worker	60 (11.8%)	21 (28.4%)	81 (13.9%)
Government employee	50 (9.9%)	39 (52.5%)	89 (15.3%)
Self-employed	9 (1.8%)	1 (1.4%)	10 (1.7%)
Unemployed	215 (42.4%)	3 (4.1%)	218 (37.5%)
Retired	126 (24.8%)	3 (4.1%)	129 (22.2%)
Marital Status
Married	319 (62.9%)	45 (60.8%)	364 (62.7%)
Single	121 (23.9%)	25 (33.8%)	146 (25.1%)
Divorced	38 (7.5%)	3 (4%)	41 (7.1%)
Cohabiting	2 (0.4%)	1 (1.2%)	3 (0.5%)
Widowed	27 (5.3%)	0 (0%)	27 (4.6%)

SD Standart Deviation

ISR-10 symptom scores of the patient group

Psychological symptoms were assessed using the ISR-10, a self-report instrument designed to measure the severity and frequency of psychological symptoms. The ISR-10 was used solely for dimensional symptom assessment in the patient group and was not used as a diagnostic or classification tool. Subscale results are presented in Table 2.

Table 2.

Descriptive Statistics of ISR-10 Psychological Symptom Scores in the Patient Group

ISR-10 Subscales	Mean	SD
Depression	1.90	1.15
Anxiety	1.75	1.17
OCD	1.46	1.16
Somatization Disorder	1.13	1.13
Eating Disorder	0.70	0.93
Additional	1.11	0.83
Total	1.31	0.79

SD Standart Deviation, OCD Obsessive-Compulsive Disorder

EFA indicated a two-factor structure for the scale. These factors were subsequently examined in the reliability and validity analyses reported below.

Reliability

Item analysis

All items showed statistically significant item–total correlations above 0.20. The Cronbach’s alpha coefficient for the SRIB-10-SF was 0.79. In the validity analyses (reported below), the scale was evaluated as a two-factor structure. When analyzed by factor, Factor 1 showed a Cronbach’s alpha of 0.78, whereas Factor 2 showed an alpha of 0.56. Deleting any single item did not result in an increase in the overall Cronbach’s alpha coefficient (Table 3).

Table 3.

Item–Total Correlations and Cronbach’s Alpha Coefficients if Item Deleted for the SRIB-10-SF

Item	Item Total Score Correlation	Cronbach’s Alpha if Item Deleted
1	0.243	0.79
2	0.316	0.79
3	0.469	0.77
4	0.571	0.76
5	0.421	0.77
6	0.625	0.75
7	0.508	0.76
8	0.585	0.75
9	0.555	0.76
10	0.364	0.78

All corrected item–total correlations were statistically significant (p < .0001)

Test–retest reliability

The test was re-admınıstered to 55 volunteer partıcıpants ın the control group after 2–4 weeks to examıne test-retest relıabılıty. The test–retest correlation coefficient was 0.675 (p < .01), indicating satisfactory temporal stability.

Validity

Construct validity – factor analysis

The suitability of the data for factor analysis was evaluated using the KMO measure and Bartlett’s Test of Sphericity. The KMO value was 0.862, and Bartlett’s Test of Sphericity was statistically significant (χ² = 1278.685, p < .001), indicating that the correlations among the variables were adequate for factor analysis. EFA identified two factors with eigenvalues greater than 1 extracted, accounting for 49.26% of the total variance. Factor loadings ranged from 0.481 to 0.763 for Factor 1 and from 0.602 to 0.781 for Factor 2 (Table 4).

Table 4.

Factor Loadings Obtained from Principal Component Analysis of the SRIB-10-SF

Factor	Eigenvalue	Explained Variance (%)	Item	Component
				1	2
1.	3.644	%36.435	1	0.481	− 0.112
			3	0.615	0.160
			4	0.709	0.198
			6	0.676	0.355
			7	0.709	0.111
			8	0.763	0.173
			9	0.507	0.472
2.	1.283	%12.827	2	− 0.028	0.781
			5	0.162	0.721
			10	0.164	0.602

Confirmatory factor analysis

In the CFA, the chi-square goodness-of-fit test was statistically significant (χ² = 108.380, p < .001), and the model fit indices are presented in Table 5.

Table 5.

Goodness-of-Fit Indices for the Confirmatory Factor Analysis of the SRIB-10-SF

Fit Index	SRIB-10-SF	Acceptable Model Fit	Excellent Model Fit
χ² test (p value)	< 0.000	< 0.05	-
χ²/df	3.18	< 5	< 3
CFI	0.942	> 0.9	> 0.95
NFI	0.91	> 0.9	> 0.95
IFI	0.943	> 0.9	> 0.95
GFI	0.962	> 0.85	> 0.9
AGFI	0.939	> 0.85	> 0.9
RMSEA	0.062	0.05–0.08	0.00-0.05
TLI	0.924	> 0.9	> 0.95

χ² chi-square, df  Degrees of Freedom, CFI  Comparative Fit Index, NFI  Normed Fit Index, IFI  Incremental Fit Index, GFI  Goodness-of-Fit Index, AGFI  Adjusted Goodness-of-Fit Index, RMSEA Root Mean Square Error of Approximation, TLI Tucker–Lewis Index

The standardized regression coefficients were statistically significant and ranged from 0.28 to 0.73 (Table 6).

Table 6.

Standardized Regression Weights for the Factor Structure

Latent Factor	Item	Standardized Estimate
Factor 1	SRIB1	0.287
	SRIB3	0.534
	SRIB4	0.645
	SRIB6	0.736
	SRIB7	0.608
	SRIB8	0.726
	SRIB9	0.603
Factor 2	SRIB2	0.502
	SRIB5	0.635
	SRIB10	0.503

All estimates are statistically significant at the p < .001 level

SRIB Stres Resilience Item Bank

Concurrent validity

To evaluate the concurrent validity of the SRIB-10-SF, Pearson correlation analysis was conducted to examine its relationships with the CD-RISC and BRS. Significant positive correlations were found between the total SRIB-10-SF score and the total scores of the CD-RISC and BRS (Table 7).

Table 7.

Inter-Correlations Between SRIB-10-SF, CD-RISC, and BRS

	1	2	3
SRIB-10-SF	-
CD-RISC	0.73**	-
BRS	0.67**	0.58**	-

SRIB-10-SF Stress Resilience Item Bank 10-Item Short Form, CD-RISC Connor-Davidson Resilience Scale, BRS  Brief Resilience Scale

**p < .001

Discriminant validity

Independent-samples t-tests were conducted to examine whether the SRIB-10-SF could differentiate between clinical and non-clinical groups. The results indicated a statistically significant difference in scale scores between the patient group with psychosomatic disorders and the control group (p < .0001; Table 8).

Table 8.

Independent Samples t-Test Results Comparing SRIB-10-SF Scores Between Patient and Control Groups

Scale	Group	N	M	SD	F	t	df
SRIB-10-SF	Patient	507	26.10	5.60	15.845	-5.339	579
	Control	74	29.70	3.60

N  Number, M  Mean score, SD  Standard Deviation, F F-value, t t-value, df  Degrees of Freedom. p < .0001

Discussion

Evaluation of descriptive findings

In the present study, gathering data from two distinct cohorts—patients with psychosomatic disorders and a healthy control group—yielded critical insights into the applicability of the SRIB-10-SF across both clinical and non-clinical populations, thereby enabling meaningful intergroup comparisons. This dual-sample approach distinguishes the current study from the original validation, which relied exclusively on a sample of patients with psychosomatic disorders [8].

Prior to the main administration, the preliminary Turkish translation of the scale was pilot-tested on 30 individuals from diverse sociocultural backgrounds who were external to the final study sample. Reflecting the feedback from these pilot participants, necessary refinements were implemented to finalize the scale. Although the control group exhibited a significantly higher educational level than the patient group, this disparity was deemed unlikely to bias the statistical outcomes, given the scale’s brevity, high comprehensibility, and the adequate literacy levels of all participants. Similarly, variations in employment status were expected to have a negligible impact on scale completion. The predominance of civil servants within the control group is attributable to the recruitment of hospital staff for this cohort. Nevertheless, the potential influence of these demographic differences cannot be entirely excluded and should be considered when interpreting the results. Both the patient and control groups were predominantly female, with no substantial differences in their mean age.

Among the 507 patients, 74.2% were female, reflecting the gender distribution in the psychosomatic patient population, which is similar to the original study (68.8% female). Furthermore, the ISR-10 subscale scores obtained in the present study were highly consistent with those reported in the original validation, further confirming the comparability of the samples [8].

Evaluation of reliability findings

Item-total correlation serves as a core indicator of psychometric reliability, capturing the degree of association between individual items and the composite or subscale score. Methodological literature suggests that item-total correlation values exceeding 0.20 are generally acceptable [27]. In the current study, the item-total correlations for all 10 items surpassed this 0.20 threshold. Although the original validation reported all item-total correlations to be above 0.40, three items in the Turkish version fell below this baseline [8].

The relatively lower item-total correlations observed for SRIB1 (“Nothing can shake my self-confidence”) and SRIB2 (“Many things can upset my emotions”) might be attributable to cross-cultural variations in emotional expression and cognitive appraisal between German and Turkish populations [28]. Additionally, since SRIB2 and SRIB10 are reverse-coded items, subtle nuances in meaning may have been compromised during the cross-cultural translation process, potentially contributing to their attenuated correlations. Nevertheless, it is important to emphasize that item-total correlation values for all items remained within acceptable limits, supporting the overall reliability of the scale.

The Cronbach’s alpha coefficient for the SRIB-10-SF was 0.79, demonstrating satisfactory internal consistency according to established literature standards [29]. In comparison, the original study reported a higher reliability coefficient of 0.98 utilizing IRT frameworks [8]. The relatively lower internal consistency observed for the second factor might be linked to its brief length and the inclusion of reverse-coded items. Extensive psychometric research indicates that short subscales and reverse-worded items frequently attenuate Cronbach’s alpha estimates, even when the underlying factor structure is robust [30, 31].

Standard psychometric protocols dictate that if the exclusion of an item elevates the overall scale’s alpha coefficient, that specific item undermines internal consistency and warrants removal [32]. Conversely, an item may be retained despite an increased alpha-if-deleted value if its item-total correlation is robust or if preserving the conceptual integrity of the scale is prioritized [33]. For the Turkish SRIB-10-SF, the comprehensive structural configuration of the original scale was maintained, as item-deletion analyses revealed no potential increases in the Cronbach’s alpha.

The test-retest reliability, evaluated via Pearson correlation, was 0.675 (p < .001), demonstrating acceptable temporal stability over time [32, 34, 35]. Taken together, these findings provide compelling evidence for the reliability of the Turkish adaptation of the SRIB-10-SF.

Evaluation of validity findings

To ascertain the suitability of the dataset for factor analysis, the KMO measure of sampling adequacy and Bartlett’s test of sphericity were computed. Bartlett’s test being significant indicates that the data were appropriate for factor analysis [36]. The resulting KMO value of 0.862 is classified as “meritorious” or “very good” according to established benchmarks [37, 38].

Various criteria exist for determining factor extraction; among these, the Kaiser criterion (retaining eigenvalues > 1) and the visual inspection of the scree plot remain the most widely utilized [37, 39]. In the present study, both extraction methods concurrently converged on a two-factor solution. In multifactorial designs, it is considered sufficient for the explained variance to be between 40 and 60%. The identified two-factor model accounted for 49.2% of the total variance in the SRIB-10-SF, fitting neatly within this recommended parameter [40].

Based on the factor loading matrix, Items 1, 3, 4, 6, 7, 8, and 9 loaded substantially onto the first factor, whereas Items 2, 5, and 10 clustered under the second factor. Consequently, the first dimension was designated as “Self-Confidence and Resilience,” while the second was labeled “Difficulty and Deficiency in Emotion Control.”

The factor loadings in the first factor range from 0.48 to 0.76, and in the second factor from 0.60 to 0.78. While certain guidelines link the minimum acceptable factor loading threshold (e.g., 0.30 or 0.40) directly to sample size, other methodological perspectives classify loadings of 0.45 and above as a robust measure [27, 38, 41]. In this study, all items exhibited factor loadings exceeding the 0.45 threshold, thereby reinforcing the construct validity of the scale.

In the original instrument development by Obbarius et al., both the long and short forms were structured as unidimensional measures based on IRT modeling [8]. Erkuş noted that factor structures within psychometric scales frequently fluctuate across cultural contexts, necessitating culturally tailored interpretations of factor content [42]. Therefore, the emergence of a two-factor structure in the Turkish sample is likely attributable to these cross-cultural divergences. Furthermore, prior research re-evaluating the psychometric properties of the CD-RISC—originally constructed under CTT principles—via IRT demonstrated a shift in factor structure. Specifically, after the exclusion of three items, the remaining 22 items converged into a unidimensional model [19]. In contrast, while the original CD-RISC development yielded a five-factor model under CTT, its Turkish adaptation identified a distinct three-factor solution [20, 21]. Taken together, these findings illuminate how the factorial architecture of an instrument can fluctuate depending on the underlying psychometric framework and statistical methodologies employed during validation. To the best of our knowledge, no prior study has adapted a scale originally constructed under an IRT framework by utilizing CTT methodologies, thereby highlighting a major methodological strength of the present investigation.

The factor structure of the Turkish SRIB-10-SF was initially explored via EFA and subsequently verified through CFA. Although the chi-square (χ²) statistic was statistically significant, this index is notoriously sensitive to sample size, routinely yielding significant discrepancies in large cohorts. Consequently, the normated chi-square (χ²/df) ratio is widely accepted as a more robust and realistic indicator of model fit [43]. Methodological benchmarks suggest that χ²/df values less than or equal to 2 indicate a good fit, whereas values less than or equal to 5 denote an acceptable fit [44]. In the current study, the obtained χ²/df ratio ( χ²/df = 108.38/34 = 3.18) fell within the parameters for an acceptable model fit.

Additionally, an array of alternative goodness-of-fit indices was examined to comprehensively evaluate the model. Extant literature indicates that different CFA fit indices capture distinct dimensions of model adequacy, ranging from acceptable to excellent fit thresholds [44–46]. In the original validation, a single-factor CFA based on the initial 77-item pool yielded a χ²/df ratio of 2.85, which aligns closely with our observed value. Moreover, their reported RMSEA value of 0.059 is highly congruent with the error variance index obtained in the present investigation. In that original study, the CFI (0.89) and TLI (0.88) values fell below the conventional 0.90 baseline and were classified as a suboptimal fit—an outcome primarily attributed to the inflating effect of a large item pool [47]. Conversely, in the present study, both the CFI and TLI surpassed the 0.90 threshold, indicating a superior and more robust model fit than the original long-form validation.

All standardized regression coefficients in the current model were well within acceptable parameters (greater than 0.30), with the sole exception of Item 1. Although Item 1 exhibited a marginal factor loading, it was deliberately retained due to its substantial conceptual relevance and its critical contribution to the content validity of the underlying construct. Moreover, the global CFA fit indices comprehensively supported the adequacy and stability of this two-factor model. Nevertheless, the performance of this specific item warrants cautious interpretation and should be systematically re-evaluated across future replication studies.

A recent study suggested that in scale adaptation studies, EFA should be performed first, followed by CFA to achieve consistent results and standardization [48]. Research examining scale adaptation studies conducted in Turkey found that both EFA and CFA are generally used, but EFA is preferred over CFA [49, 50]. By rigorously employing both exploratory and confirmatory methods, the current investigation aligns seamlessly with rigorous international and regional psychometric standards.

Furthermore, the concurrent validity of the Turkish SRIB-10-SF was established through cross-correlations with the CD-RISC and the BRS. The composite score of the SRIB-10-SF demonstrated strong, statistically significant positive correlations with both the CD-RISC (r = .73, p < .001) and the BRS (r = .67, p < .001). Although both instruments assess resilience, they represent different conceptual perspectives. Specifically, the CD-RISC operationalizes resilience as a multifaceted construct encompassing adaptability, coping capacity, persistence, and personal competence; conversely, the BRS shifts focus toward the specific dynamic of psychophysiological recovery or “bouncing back” in the immediate aftermath of adversity. Consequently, the strategic inclusion of both instruments facilitated a more comprehensive and layered validation of concurrent validity, anchoring the SRIB-10-SF simultaneously to generalized resilience traits and discrete stress-recovery mechanics. In the original validation study, concurrent validity was restricted solely to correlations with the CD-RISC, which yielded highly convergent results (r = .79, p < .001) [8].

Furthermore, independent samples t-test analyses revealed that the clinical cohort scored statistically significantly lower on the SRIB-10-SF than their healthy counterparts. This pattern aligns with the foundational development data of the CD-RISC, where healthy women scored substantially higher than those diagnosed with fibromyalgia [22]. Likewise, cross-cultural validations of the CD-RISC in Korea and Greece confirmed that healthy populations report significantly higher resilience scores relative to clinical populations diagnosed with psychiatric conditions [33, 51].

Limitations

One limitation of this study is that the majority of the participants were women. his may reflect the higher proportion of women among patients with psychosomatic disorders. Furthermore, differences in sample size, educational level, and gender distribution between the patient and control groups, despite efforts to recruit comparable groups, constitute additional limitations of the study.

Conclusion and recommendations

The lack of previous studies examining the validity and reliability of the SRIB-10-SF made it difficult to compare and interpret the findings; however, this also highlights the novelty of the present study. Another strength of this study is that the sample consisted of patients with psychosomatic disorders. Previous studies on the development and adaptation of stress resilience scales have predominantly included healthy individuals, students, or patients with post-traumatic stress disorder; therefore, the inclusion of patients with psychosomatic disorders may provide a valuable contribution to the literature [21, 23, 52, 53].

The SRIB-10-SF demonstrated satisfactory reliability and validity among Turkish patients with psychosomatic disorders, the majority of whom are women and may serve as a brief and practical instrument for assessing stress resilience in clinical settings. Its use may facilitate the identification of individuals with lower levels of stress resilience and help inform interventions targeting stress-related physical and psychological symptoms.

Future studies should evaluate the Turkish SRIB-10-SF in larger and more diverse samples to further examine the stability of its factor structure and to confirm its reliability and validity across different populations.

Abbreviations

AGFI

Adjusted Goodness-of-Fit Index

AMOS

Analysis of Moment Structures

BRS

Brief Resilience Scale

CTT

Classical Test Theory

CFI

Comparative Fit Index

CFA

Confirmatory Factor Analysis

CD-RISC

Connor-Davidson Resilience Scale

CD-RISC-10

10-Item Connor–Davidson Resilience Scale

df

Degrees of Freedom

DSM-5

Diagnostic and Statistical Manual of Mental Disorders, Fifth Edition

EFA

Exploratory Factor Analysis

GFI

Goodness-of-Fit Index

IFI

Incremental Fit Index

ISR- 10

ICD10 Symptom Rating Scale

IRT

Item Response Theory

KMO

Kaiser–Meyer–Olkin

NFI

Normed Fit Index

RMSEA

Root Mean Square Error of Approximation

SD

Standart Deviation

SPSS

Statistical Package for the Social Sciences

SRIB

Stress Resilience Item Bank

SRIB-10-SF

Stress Resilience Item Bank 10 Item Short Form

TLI

Tucker-Lewis Index

χ²

Chi-square

Authors’ contributions

Berfu Ünal Nural: Conceptualization, Investigation, Methodology, Project administration, Writing – original draft.Muhammed Hakan Aksu: Data curation, Methodology, Research, Project management, Writing – review & editing.Ömer Aydemir: Data curation, Research, Methodology, Writing – review & editing.Behçet Coşar: Data curation, Methodology, Supervision, Writing – review & editing.

Funding

This research did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors.

Data availability

All data generated or analysed during this study are included in this article.Further enquiries can be directed to the corresponding author.

Declarations

Ethics approval and consent to participate

Ethics committee approval was obtained from the Ethics Committee of Gazi University in accordance with the Declaration of Helsinki, dated 12/04/2021 and numbered E-77082166-604.01.02-72639.

Consent for publication

All authors gave approval for publication.

Competing interests

The authors declare no competing interests.