Journey of reintegration: developing and validating an index for assessing the reintegration status of migrant workers

Abstract

Background

Over 1.5 million Sri Lankans are employed overseas, predominantly as temporary migrant workers in the Middle East. Despite their substantial contribution to the national economy, empirical evidence on their reintegration experiences after return remains limited. This study aimed to develop and validate an Index of Reintegration (IR) to assess the reintegration status of workers who returned from the Middle East and are living in the Kurunegala District, Sri Lanka.

Methods

The IR was developed using a three-phase process: item development, scale development, and scale evaluation. Item development utilised deductive (literature review) and inductive (interviews with experts, returnees and their families) methods. A community-based survey among 650 returnee migrant workers in Kurunegala District collected data for factor analysis in the scale development phase. Developed scale was evaluated for its reliability and validity.

Results

Exploratory Factor Analysis revealed a 20-item IR with six factors, explaining approximately 60% of the total variance. Confirmatory Factor Analysis affirmed significant factor loading for all items to their respective factor and supported the goodness of fit for the six-factor model. The factors encompass Economic Self-Sufficiency, Family Reacceptance, Physical-Psychological Health, Social Stability, Support Networks, and Cultural Adaptation. Scale evaluation for reliability indicated Cronbach’s alpha values ≥ 0.7 for all components. The scale underwent validation for content validity and exhibited good construct validity.

Conclusions

The IR is a reliable and valid tool to assess the reintegration status of Sri Lankan workers who returned from Middle East. The scale has broad applications for research and practice. Further research across diverse migrant worker groups is recommended to enhance the scale’s generalisability and robustness.

Supplementary Information

The online version contains supplementary material available at 10.1186/s12889-026-26704-x.

Background

Temporary migrant workers embark on their migration journey with the intention of eventually returning to their country of origin [1, 2]. Over the past three decades, temporary labour migration from Sri Lanka and other South Asian countries to the Middle East has expanded substantially, driven by employment opportunities in low- and semi-skilled sectors [1]. Currently, over 1.5 million Sri Lankans constituting nearly one-fifth of national workforce, are employed overseas, contributing significantly to the Sri Lankan economy through remittances [3, 4]. There are many related policies and government, and non-government organisations aimed at assisting the welfare of outbound migrant workers and their families left behind. However, comparatively little attention has been paid to the post-return phase of migration, particularly how Sri Lankan workers who returned from Middle East reintegrate into their home communities and the factors that influence successful reintegration [2, 5, 6].

Reintegration is widely recognised as a multidimensional “process which enables individuals to re-establish the economic, social and psychosocial relationships needed to maintain life, livelihood and dignity and inclusion in civic life” [7]. The successful reintegration, of migrant workers is crucial not only for the individual welfare of those returnees but also for the national development when considered collectively [8]. Economic theories like Human Capital Theory [9] and Social Capital principles related to migration [10] emphasise that sustainable reintegration depends on the effective utilisation of skills, knowledge, financial assets, and social networks accumulated during migration to support livelihoods and social participation after return [6]. In contrast, unsuccessful reintegration, particularly when migrants’ economic or social aspirations remain unmet may increase the likelihood of repeat or circular migration [6, 11]. Moreover, social reintegration failures can strain relationships between returnee migrant workers and their family members, weaken community ties, and disrupt informal social support systems that are critical for long-term stability and well-being [6, 12].

Despite the growing recognition of reintegration as a critical phase of the migration cycle, empirical research examining reintegration experiences in low- and middle-income countries remains limited. Moreover, scales used in previous studies were designed for populations such as asylum seekers and forced returnees [13, 14], which differ substantially from the target population of this study. Additionally, such tools did not adequately capture socio-cultural aspects of reintegration that are particularly salient in the Sri Lankan and other similar developing country contexts, including extended family structures, community-based social support, informal employment arrangements, and culturally specific notions of social acceptance and belonging [12, 14]. Nevertheless, relevant methodological features of existing scales [13, 14], such as item structure, scoring approaches, terminology, and conceptually related items were reviewed and selectively adapted to inform the development of the new scale.

This highlights a significant knowledge gap, necessitating the development of a new, culturally appropriate tool for assessing reintegration among Sri Lakan workers who returned from the Middle East [6, 15]. Moreover, this gap has practical implications. Although the Sri Lankan government has introduced initiatives to support the economic and social reintegration of returnee migrant workers, the absence of a standardised, culturally appropriate evaluation tool constrains the ability to systematically assess programme effectiveness and inform evidence-based policy development [12, 16].

In conclusion, there is a need for a tool to objectively monitor and evaluate the reintegration of workers who returned from the Middle East to their home countries, such as Sri Lanka. Such a tool could provide valuable insights for researchers, academics, service providers, and policymakers involved in migration management. To address this, the study aimed to develop and validate an Index of Reintegration (IR) to assess the reintegration status of Sri Lankan workers who returned from the Middle East.

Methods

The development and validation of the Index of Reintegration (IR) to assess the reintegration status of Sri Lankan workers who returned from the Middle East were conducted in three phases: item development, scale development, and scale evaluation. As illustrated in Table 1, each phase comprised of several sequential steps. Data collection for the literature review, qualitative interviews, and cross-sectional survey conducted across these phases (Table 1) took place between August 2021 and April 2022.

Table 1.

Illustration of the steps followed for the development & validation of the Index Reintegration (IR)

Item Development	Step 1: Identification of domain/construct and item generation	Literature review,
		Interviews with returnee migrants,
		Interviews with family members of returnee migrants,
		Key informant interviews with experts,
	Step 2: Assessment for Judgemental Validity: Assessing if the items adequately measure the domain of interest
Scale Development	Step 3: Formulating draft index to measure the status of reintegration
	Step 4: Pre-testing of the draft index
	Step 5: Cross-sectional survey of Field testing of the index for factor analysis
Scale Evaluation	Step 6: Assessment of the reliability and validity of the ‘Index of Reintegration’
	Step 7: Finalising the ‘Index of Reintegration’ (IR)

Phase 1-Item development

Item development began with defining the domain or construct to be measured, namely the “reintegration status of Sri Lankan workers who returned from the Middle East to the home community”. Next, the International Organization for Migration (IOM) definition of reintegration was chosen as the operational definition due to its comprehensive coverage of multiple important dimensions and global acceptance. This study employed both deductive and inductive methods for item development.

Literature review

This involved an extensive literature search across electronic databases (“PubMed”, “Embase”, “SCOPUS”, “JSTOR”, and “Web of Science”). using relevant keywords and controlled vocabulary searches. Literature searches were also made in “Google Scholar” and by manually reviewing the reference lists of identified articles. A grey literature search covered unpublished works, reports, theses, dissertations, government documents, and policies, utilising bibliographies available at relevant institutional Libraries. This iterative approach yielded 63 abstracts, 42 full articles, 27 full reports, and four books, which were subsequently reviewed.

Qualitative interviews

This involved conducting one-to-one semi-structured interviews with technical experts, returnee migrant workers, and their family members (step 1). A total of 19 interviews were conducted, including 12 Sri Lankan workers who returned from the Middle East (7 female and 5 male), three spouses, and four other adult family members, including parents and parents-in-law of the returnees. A total of 12 key informant interviews were conducted with individuals possessing academic, policy, and field-based experience in working with migrant populations. While many informants were affiliated with academic institutions, they also held substantive practitioner roles and engaged closely with migrant communities through applied research, programme implementation, policy advisory work, and collaboration with migrant networks and civil society organisations. The informants represented a range of relevant disciplines and practice areas, including labour and migration studies, law, economics, sociology, and public health, thereby providing a transdisciplinary perspective on migrant reintegration. The first author conducted interviews using an interviewer guide to ensure consistency and comprehensive coverage of the study’s scope. Qualitative interviews typically lasted 45–60 min, and were audio-recorded, transcribed, and analysed using thematic analysis.

Assessment for judgemental validity

Judgemental validity encompasses face, content, and consensual validity and is a critical early step in instrument development to ensure that scale items adequately represent the construct of interest, are meaningful to the target population, and are grounded in both theory and expert knowledge [17]. Face validity was assessed by a group of returnee migrant workers. Content validity was evaluated by a team of two measurement experts and five content experts in the fields of labour studies, migration studies, law, sociology, and public health, to ensure relevant domains and items were included. Each expert independently reviewed the pool of items and rated the importance of each item using a ten-point scale. Items with high expert ratings (≥ 5) were included in the draft IR. Furthermore, experts provided feedback on clarity, relevance and potential additions, edits, or deletions of items. Subsequently, the entire expert panel employed a consensus approach to establish the consensual validity for items that received lower scores (< 5), yet were considered important in the literature, as well as for items that newly emerged during the content validity assessment.

Phase 2- Scale development

Formulating the draft IR

The draft IR incorporated shortlisted items that underwent content and consensual validation. By using the methodology described by Koo and Yang [18], for questionnaire design and development a Likert scale with five points was adopted. Both positively and negatively worded items were randomly interspersed within the draft IR. A scoring system was developed to produce a composite IR score ranging from 0 to 100.

The draft IR was originally developed in English and then forward-translated into local languages (Sinhala and Tamil) and again back-translated into English to ensure translation accuracy. Minor revisions to the translated items were made through a consensus process. The draft IR was pre-tested and necessary adjustments were made.

Cross-sectional survey for factor analysis

A community-based cross-sectional survey was conducted in five Divisional Secretary divisions of the Kurunegala District in Sri Lanka. Kurunegala District, the third-largest district in Sri Lanka, with a population exceeding 1.6 million, has been consistently sending a significant number of migrant workers to the Middle East over the past two decades [4, 19]. The study population included both male and female Sri Lankan workers who returned from Middle Eastern countries after being employed in skilled, semi-skilled, housemaid, and unskilled manpower-level occupations.

The sample size for Exploratory Factor Analysis (EFA) was determined considering the Sample to Item Ratio and recommendations from scholars on selecting sample sizes for factor analysis [20, 21]. The newly developed draft IR used for EFA consisted of 24 items. For the EFA, participant to scale item ratio was set as 12:1, resulting in the required sample size of 288 participants, with a 10% non-response rate rounded up to 320. The sample size for Confirmatory Factor Analysis (CFA) was set as 330, considering a participant-to-scale item ratio of 15:1, and accounting for an estimated 10% non-response rate. The overall study sample was recruited using a multi-stage cluster sampling method. The first author and seven other trained data collectors collected data using a pre-tested draft IR as an Interviewer Administered Questionnaire.

The EFA was performed using Principal Component Analysis (PCA) to determine the factor structure of the IR. An oblique factor rotation method was employed with the expectation that the factors could exhibit correlations. The Kaiser–Meyer–Olkin (KMO) measure and Bartlett’s test of Sphericity tests were applied to determine the sample adequacy to perform EFA. Further multiple criteria: Eigenvalue-based factor selection, the Scree plot-based factor selection and Parallel analysis were used to determine the optimum number of factors to be retained. Thereafter, CFA was conducted to construct a measurement model and assess model fit to ascertain the alignment between observed variables and hypothesised latent constructs/factors.

Phase 3- Scale evaluation

In Phase 3, the Index of Reintegration (IR) was evaluated for reliability and validity. Test–retest reliability was assessed using the Intraclass Correlation Coefficient (ICC) by re-administering the IR to a randomly selected subsample of 50 returnee migrant workers, 14 days after the initial administration in the original survey setting. These 50 respondents were part of a larger sample used for factor analysis. The internal consistency of each latent construct was determined using Cronbach’s alpha values. Composite reliability was examined within the developed measurement model from the CFA in line with established literature [22]. In addition to the content and consensual validity assessment, the developing IR was evaluated for construct validity (convergent and discriminant validity).

Once all stages of development and validation were completed, the IR was finalised by organising items, response options, and user instructions into a practical and user-friendly format. SPSS version 22 was used for the EFA, while SPSS AMOS software was used for the CFA. Informed written consent was obtained from each participant, and steps were taken to safeguard their privacy and confidentiality. Ethical clearance was obtained from the Faculty of Medicine, University of Kelaniya, Ethics Review Committee (P/22/03/2021).

Results

Item development encompassed a comprehensive literature review and one-to-one semi-structured interviews with experts, returnee migrant workers and selected adult family members of returnees. The preliminary item pool consisted of 53 items (Supplementary file 1). Following expert assessment for content validity, the preliminary item pool was reduced to 18 items. These 18 items, together with several additional items proposed by experts during the content validity assessment, were subsequently re-evaluated using a consensus-based approach. Through this process 24 items were shortlisted for inclusion in the draft Index of Reintegration (IR) for factor analysis.

For scale development, a cross-sectional survey was conducted among returnee migrant workers. A total of 300 eligible participants were approached for the exploratory factor analysis (EFA) component, yielding a response rate of 93.7%. For the confirmatory factor analysis (CFA) component, 322 eligible returnees were approached, with a response rate of 97.6%. The demographic characteristics of the study samples are summarised in Table 2.

Table 2.

Frequency distribution of demographic characteristics of the EFA and CFA study samples

Characteristics	Study respondents of EFA (n = 300)		Study respondents of CFA (n = 322)
	Number	Percentage (%)	Number	Percentage (%)
Sex
Male	137	45.7	132	41.0
Female	163	54.4	190	59.0
Age
35 to 39	85	28.3	92	28.6
40 to 44	103	34.3	99	30.8
45 to 49	112	37.3	131	40.7
Sector
Urban	73	24.3	76	23.6
Rural	227	75.7	246	76.4
Host Country Employed*
Kuwait	102	34.0	70	21.7
Qatar	6	2.0	21	6.5
UAE	11	3.7	66	20.5
Saudi Arabia	84	28.0	112	34.8
Oman	14	4.7	17	5.3
Bahrain	69	23.0	30	9.3
Missing	14	4.7	6	1.9
Manpower level of the foreign job**
Skilled	78	26.0	87	27.0
Semi-skilled	22	7.3	53	16.5
Unskilled	65	21.7	41	12.7
Housemaid	126	42.0	137	42.5
Missing	9	3.0	4	1.2
Period employed in the Middle East
5 to 10	181	60.3	159	49.4
11 to 15	68	22.7	101	33.7
16 to 20	43	14.3	49	16.3
> 20	4	1.3	3	1.0
Missing	4	1.3	10	3.3

Missing refers to the number not responded to the question

Host Country Employed* classified according to last Middle East country of employment

Manpower level of the foreign job** was categorised according to SLBFE classification

The factorability of the sample was gauged using the Kaiser–Meyer–Olkin (KMO) measure and a KMO value of 0.69 (> 0.5) indicated sample adequacy for EFA. Furthermore, Bartlett’s Test yielded statistically significant results: X² (n = 300) = 1356.309 (p < 0.0001), confirming adequacy for factor analysis. To ensure EFA quality, the minimum extraction communality value was set to 0.4. Four items with lower commonality values (< 0.4) were observed to exhibit cross-loading or poorly loading. These four items with < 0.4 communality values were removed from 24 items, yielding 20 items. The pattern matrix obtained through Direct Oblimin rotation (Table 3) revealed that all 20 items showed factor loadings exceeding 0.5.

Table 3.

Pattern matrix and communality values for the final index of reintegration

Items	Pattern coefficients						Communalities
	Component
	1	2	3	4	5	6
Economic self-sufficiency (ESS)
ESS01 Overseas employment enabled me to acquire assets	0.74						0.57
ESS02 Adjusted* Facing difficulties with loan settlement or redeeming pawned valuables	0.71						0.50
ESS03 Family members are satisfied with my financial contribution to the family	0.67						0.53
ESS04 Satisfied with my current standard of living	0.63		0.32				0.68
ESS05 My current income adequately covers my basic financial needs	0.57						0.66
Family Reacceptance (FR)
FR01 Feel loved and valued within the family		0.73					0.54
FR02 Satisfied with the way my family accepted me back into the family		0.72					0.62
FR03 Satisfied with intimate relationships		0.60					0.52
Physical-psychological Health (PPH)
PPH01 Adjusted* Felt the need to seek psychological support			0.76				0.60
PPH02 Adjusted* Physical illness interfered with my regular daily activities		0.31	0.62				0.60
PPH03 Adjusted* Struggle to cope with past or present experiences of harassment			0.59				0.56
PPH04, I have a consistently sought timely medical attention for my health needs			0.59				0.51
Social Stability (SS)
SS01 The skills I acquired during overseas employment were useful in generating an income back in home county				0.76			0.67
SS02 Feeling safe and secure in daily activities				0.71		0.31	0.61
SS03 Adjusted* I encounter challenges in maintaining habits/behaviours I acquired in abroad				0.53			0.55
Support Networks (SN)
SN03 Have reliable support networks (e.g., family, friends, relatives) which I can turn to in times of difficulty					0.78		0.66
SN02 I am satisfied with the quality of available public services & infrastructure (e.g., healthcare, education)					0.69		0.51
SN01 Able to regain my rights upon return to my home country					0.66		0.57
Cultural Adaptation (CA)
CA01 Actively participate in community-level-organisations, societies						0.71	0.66
CA02 Enjoy interacting with local people at my workplace/community gatherings						0.68	0.52

Extraction Method: Principal Component Analysis. Rotation Method: Oblimin with Kaiser Normalisation

Coefficients were sorted by size and small coefficient values below 0.3 were suppressed

“Adjusted” refers to the negatively worded items in the IR that were modified/adjusted during the analysis phase

The optimum number of factors to be retained for the novel IR was determined using three methods: Eigenvalue-based factor selection (Kaiser Criterion), the Scree plot, and Parallel analysis. The eigenvalue-based factor selection (Table 4) indicated that six factors/components with Eigenvalues greater than one (> 1), which were chosen for further analysis. Factor-wise, the first, second, and third factors exhibited 16.3%, 13.1%, and 8.1% of the variance, respectively.

Table 4.

Total variance explained and results of parallel analysis to determine the optimum number of factors to be retained for the IR

Component number	Total initial eigenvalues from EFA			Random data eigenvalues from parallel analysis*	The decision to retain the factor
	Total	Percentage of variance	Cumulative percentage
1	3.27	16.34	16.34	1.56	Accept
2	2.62	13.08	29.42	1.44	Accept
3	1.62	8.09	37.51	1.37	Accept
4	1.55	7.76	45.27	1.32	Accept
5	1.30	6.50	51.77	1.26	Accept
6	1.25	6.25	58.02	1.21	Accept
7	0.94	4.72	62.74	1.17	Reject
8	0.88	4.42	67.16	1.12	Reject
9	0.86	4.28	71.44
10	0.80	4.02	75.47

Extraction Method: Principal Component Analysis

^*Specifications for this run: Ncases 300, Nvars 20, Ndatasets100, Percent 95

Next, the Scree plot (Fig. 1) displayed the inflection point or the bend at component 7. Therefore, six factors above the inflection point were subjectively selected to be retained. Parallel analysis, as the third method for determining the optimal number of components, compared the initial eigenvalues of the real data used for EFA with random data eigenvalues of a random dataset for 300 observations and 20 variables (Table 4).

Fig. 1.

Exploratory factor analysis—scree plot

Based on the parallel analysis, six factors (Components 1–6) with initial eigenvalues from EFA exceeding the corresponding eigenvalues derived from random data were retained (Table 4). All three methods used to determine the optimal number of factors consistently recommended retaining six factors. Together, the first six factors explained 58.0% of the total variance.

In CFA, all items included in the final model (Table 5) demonstrated standardised factor loadings greater than 0.5. Multiple goodness-of-fit indices were examined to assess the overall model fit, including Chi-square, minimum discrepancy of confirmatory factor analysis/degrees of freedom (CMIN/df), goodness of fit index (GFI), comparative fit index (CFI), Tucker-Lewis Index.

Table 5.

Confirmatory Factor Analysis (CFA) loadings, reliability statistics and convergent validity of the index of reintegration

Description items/components	CFA loadings	Cronbach's alpha	Composite reliability	Test–retest reliability ICC (95%CI)	Average variance extracted
Economic Self-sufficiency		0.83	0.83	0.88 (0.78–0.93)	0.49
ESS01	0.66
ESS02	0.58
ESS03	0.57
ESS04	0.84
ESS05	0.82
Family Reacceptance		0.84	0.84	0.87 (0.72–0.93)	0.64
FR01	0.71
FR02	0.85
FR03	0.83
Physical and Psychological Health		0.76	0.81	0.91 (0.84–0.95)	0.52
PPH01	0.55
PPH02	0.91
PPH03	0.73
PPH04	0.64
Social Stability		0.86	0.87	0.79 (0.64–0.88)	0.69
SS01	0.82
SS02	0.92
SS03	0.74
Support Networks		0.82	0.70	0.90 (0.78- 0.95)	0.61
SN01	0.76
SN02	0.70
SN03	0.88
Cultural Adaptation		0.66	0.72	0.90 (0.82–0.94)	0.57
CA01	0.73
CA02	0.78

(TLI), standardised root mean squared residual (SRMR), and Root Mean Square Error of Approximation (RMSEA) [22]. These values (X² = 413.519(151), p < 0.001, CMIN/df = 2.74, GFI = 0.90, CFI = 0.90, TLI = 0.88, SRMR = 0.05, and RMSEA = 0.07) were within or very close to their acceptance levels, supporting a good fit for the six-factor model.

Factor 1 (Economic Self-sufficiency) consisted of five items, while Factor 3 (Physical and Psychological Health) consisted of four items. Factor 6 (Cultural Adaptation) consisted of two items, while each of the remaining factors consisted of three items (Fig. 2). Overall, the six-factor measurement model identified through factor analysis corresponded well with the theoretical proposition of this study.

Fig. 2.

Measurement model of the Index of Reintegration (IR). Footnote: ESS = Economic Self-Sufficiency; FR = Family Reacceptance; PPH = Physical-Psychological Health; SS = Social Stability; SN = Support Networks; CA = Cultural Adaptation

Scale evaluation involved assessing the reliability and validity of the IR. Internal consistency reliability was examined using data from 322 participants (Table 5). Cronbach’s coefficients indicated good internal consistency for most factors, with values exceeding 0.7: Economic Self-sufficiency (α = 0.83), Family Reacceptance(α = 0.84), Physical Psychological Health (α = 0.76), Social Stability (α = 0.86), and Support Networks (α = 0.82). Factor 6, Cultural Adaptation, which comprised two items, demonstrated a slightly lower alpha value (α = 0.66). However, no items were removed, as item deletion did not result in an improvement in the alpha coefficient.

Composite reliability further supported the scale’s reliability, with values ranging from 0.72 to 0.87. All factors exceeded the benchmark of 0.7, except Support Networks, which was 0.70. Test–retest reliability was good, with intraclass correlation coefficient (ICC) values ranging from 0.793 to 0.908, indicating strong temporal stability of the IR (Table 5).

Construct validity, including convergent and discriminant validity, was evaluated during the scale’s evaluation phase. Convergent validity for scale factors was determined using Average Variance Extracted (AVE) [22]. The AVE values, ranged from 0.57 to 0.69, surpassing the 0.5 threshold for all constructs except “Economic Self-Sufficiency” (0.49). Therefore, the IR scale demonstrated the required convergent validity (Table 5).

Discriminant validity was assessed using the Fornell and Larcker Criterion, which is satisfied when the square root of the average variance extracted (AVE) for each construct exceeds its correlations with other constructs [22, 23]. All constructs met this criterion, with the square root of AVE consistently greater than the inter-construct correlations, confirming adequate discriminant validity (Table 6).

Table 6.

Component correlation matrix for discriminant validity assessment

Components/factors	Economic self-sufficiency	Family reacceptance	Physical and psychological health	Social stability	Support networks	Cultural adaptation
ESS	0.7	0.28	0.10	−0.23	0.17	0.10
FR	0.28	0.80	−0.07	0.14	0.28	0.27
PPH	0.10	−0.07	0.72	0.05	0.01	0.09
SS	−0.23	0.14	0.05	0.83	0.04	0.24
SN	0.17	0.28	0.01	0.04	0.78	0.34
CA	0.10	0.27	0.09	0.24	0.34	0.75

The final IR consists of 20 items in the form of a Likert scale (Supplementary file 2), organised into six components identified through EFA and CFA. The IR generates a composite score ranging from 0 to 100, with higher scores indicating better reintegration status.

Discussion

This study aimed to develop and validate an Index of Reintegration (IR) to assess the reintegration status of Sri Lakan workers who returned from the Middle East. Overall, the findings demonstrate that the IR possesses good psychometric properties. EFA identified a 20-item IR with six factors that together explained nearly 60% of the total variance. CFA further supported the six-factor structure, with all items demonstrating significant loadings on their respective constructs and satisfactory overall model fit. The six factors include Economic Self-Sufficiency, Family Reacceptance, Physical-Psychological Health, Social Stability, Support Networks, and Cultural Adaptation. Reliability testing showed acceptable to good internal consistency across components, and the scale demonstrated adequate content and construct validity.

The IR was developed using a rigorous three-phase stepwise approach encompassing item development, scale development, and scale evaluation. This method consistent with best practices in scale development and validation [17, 24]. Furthermore, the study adhered to established reporting guidelines for scale development, as recommended by Boateng et al. [17].

Streiner and Kottner [25], emphasised the importance of initiating scale development with a gap analysis to identify existing instruments that measure similar constructs. As described in the introduction the literature review identified two international reintegration scales previously applied to different populations [14]. Moreover, neither was cross-culturally validated nor conceptually appropriate for Sri Lankan returnee migrant workers. Nonetheless, selected technical elements, such as relevant items and the scoring mechanism, were adapted from the existing scales.

Consistent with methodological recommendations, this study employed both deductive and inductive approaches for item development [17, 26]. Deductive methods involved a comprehensive literature review, while inductive methods included semi-structured interviews with returnee migrant workers, their family members, and subject-matter experts. This mixed approach offered two key advantages. First, it enabled the IR to extend beyond the predominantly economic and socio-cultural focus of earlier instruments by incorporating a distinct physical–psychological health dimension. Second, it facilitated the development of a context-specific scale tailored to Sri Lankan workers who returned from overseas employment, a population that constitute nearly one-fifth of the national workforce [3, 4].

Reintegration is widely recognised as a complex, multi-dimensional process including economic, social and psychosocial relationships that enable individuals to maintain life, livelihood and dignity and inclusion in civic life [7]. The Integrated Approach to Reintegration theory, highlights the need to consider reintegration across individual, community, and structural levels [27]. Reliance on a single data source may therefore result in a partial understanding of this complex construct. In line with recommendations by Kuschminder [28], this study employed a combined approach that incorporated perspectives from returnee migrant workers, family members, and experts, alongside a literature review. Importantly, the IR includes items that capture community- and structural-level factors, rather than focusing solely on individual characteristics, thereby reflecting a more comprehensive conceptualisation of reintegration.

Hair et al. [29] suggested using EFA to identify the underlying dimensions of the scale. This study employed EFA using the Principal Component Analysis (PCA). Sample adequacy for EFA, was assessed using Kaiser–Meyer–Olkin (KMO) measure and Bartlett’s Test of Sphericity. The KMO value of 0.7 exceeded the recommended minimum threshold of 0.5, affirms sample adequacy for factor analysis [29]. Bartlett’s test result was significant, indicating data do not produce an identity matrix, hence suitable for factor analysis.

Multiple factor retention methods, including Eigen value-based, Scree plot-based, and Parallel analysis converged on retaining six factors. The first six factors explained 58.01% of the total variance, closely approaching 60%. While some researchers advocate for higher total variance, caution is advised against inflating the number of retained factors solely to increase variance explanation [29]. Although retaining an additional factor (seven factors) would have marginally increased the total variance explained to 62.7%, this approach was inconsistent with the unanimous recommendations of the factor retention methods and adversely affected item loadings, resulting in substantial cross-loading issues. Therefore, the six-factor solution was retained as the most parsimonious and theoretically coherent model.

CFA results further confirmed the validity of the six-factor structure. All item loadings exceeded the recommended threshold of 0.50, indicating strong associations with their respective latent constructs [22, 29]. Model fit measures, including CMIN/df, GFI, CFI, TLI, SRMR, and RMSEA, provided supportive evidence for the goodness of fit for the six-factor model. These findings aligned with recommendations for an adopted model in CFA [22, 29].

In the third phase developed scale was evaluated for its reliability and validity. Internal consistency reliability was acceptable for all factors, with Cronbach’s alpha values exceeding 0.70 for most components [29, 30]. The Cultural Adaptation factor yielded a slightly lower alpha (α = 0.66), which may be attributable to the small number of items within this construct. As item removal did not improve internal consistency and the factor was theoretically important for content validity, all items were retained.

Convergent validity evaluated using Average Variance Extracted (AVE) [22], satisfactorily exceeded 0.5 threshold for all factors, except “Economic Self-Sufficiency” (AVE = 0.49). However, given the strong composite reliability, “Economic Self-Sufficiency” was deemed valid. Therefore, the IR scale demonstrated the necessary convergent validity (Table 3). Discriminant validity was assessed using the Fornell and Larcker Criterion [29] and all the constructs demonstrated discriminant validity.

Several limitations of this study should be acknowledged. First, the qualitative component that informed item development involved relatively small samples across stakeholder groups, including 19 Sri Lankan workers who returned from the Middle East, their spouses and other adult family members, and 12 key informant interviews with experts. These sample sizes were primarily constrained by time and resource limitations. Although the interviews yielded valuable insights into reintegration experiences, the limited number of participants within each subgroup may not have fully captured the diversity and heterogeneity of perspectives among workers who returned from the Middle East, their families or the full spectrum of expert views on reintegration. Second, the study focused on labour-oriented Sri Lankan workers who returned from the Middle Eastern countries, who represent the majority of Sri Lanka’s migrant workforce [4, 19]. Consequently, the findings may not be fully generalisable to the entire spectrum of returnee migrant workers, as certain professional groups and returnees from other regions were excluded. Third, data collection was confined the Kurunegala District, potentially compromising representativeness compared to an island-wide survey. Therefore, future research is recommended to test and validate the IR in more diverse samples of Sri Lankan workers who returned from foreign employment across multiple districts and occupational groups, including cross-cultural settings, to enhance its generalisability and robustness.

Implications for research and policy

Returnee migrant workers are widely recognised as a hard-to-reach population worldwide, including in Sri Lanka [6, 12]. This study contributes to the limited empirical literature on reintegration by providing a validated, multidimensional measurement tool tailored to this population. In addition to reaffirming the importance of economic and social reintegration, the IR introduces a novel dimension by exploring the health and service needs for successful reintegration [7]. Although physical and psychosocial health needs have been increasingly recognised as important aspects of reintegration in recent literature they have not been sufficiently incorporated into existing reintegration assessment scales [13, 14]. This finding underscores the significance of considering the physical and psychological health needs of Sri Lankan workers who returned from the Middle East, aligning with recommendations from the National Migration Health policy of Sri Lanka and supporting existing theories on return migration and reintegration [12, 16, 31].

The research findings lay the groundwork for future studies on reintegration, including the new migration health dimension. Researchers can expand upon the identified dimensions and measurement tools to deepen the understanding of the reintegration process, assess the implications of reintegration interventions, and investigate the experience of specific subgroups/categories of workers who returned from foreign employment in Sri Lanka and similar developing countries.

Implications for practice

The IR has several practical implications. First, government agencies such as the Sri Lanka Bureau of Foreign Employment (SLBFE), international organisations (e.g. IOM), and non-governmental actors involved in voluntary reintegration assistance (VRA) for migrants can use the IR to systematically assess reintegration status and identify individuals requiring targeted assistance. Second, the IR scale can be employed to evaluate the effectiveness of reintegration programmes and services. Third, comparing pre and post IR scores allows assessment of the intervention’s impact on returnee’s reintegration status. Overall, the IR offers a practical, evidence-based tool to inform policy, programme design, and service delivery aimed at improving reintegration outcomes for workers who returned from foreign employment in Sri Lanka and similar developing countries.

Conclusion

This study describes the development and initial validation of a 20-item Index of Reintegration (IR) designed to assess the reintegration status of Sri Lankan workers returned from Middle East. The findings indicate that the IR demonstrates acceptable psychometric properties, including satisfactory reliability and evidence of construct validity, within the study sample. As such, the IR may be a useful measurement tool for assessing multiple dimensions of reintegration among workers who returned from foreign employment in Sri Lanka and similar developing countries. The IR offers a structured approach to capturing economic, social, health-related, and community-level aspects of reintegration, thereby contributing to the limited empirical literature in this area. While further validation in diverse populations and settings is warranted, the IR has the potential to support future research examining reintegration experiences of migrant workers and to inform the evaluation of VRA related reintegration programmes and policies.

Abbreviations

AVE

Average Variance Extracted

CFA

Confirmatory Factor Analysis

EFA

Exploratory Factor Analysis

ICC

Intraclass Correlation Coefficient

IR

Index of Reintegration

KMO

Kaiser-Meyer-Olkin

SLBFE

Sri Lanka Bureau of Foreign Employment

Authors’ contributions

Pasindu Dilshan Wijeratne (Corresponding author): PDW developed the research proposal, data collection, data analysis, interpretation, and writing the manuscript. Arunasalam Pathmeswaran: AP contributed to the development of the research design, supervised data collection, participated in data analysis and interpretation.

Funding

This manuscript is based on the Post graduate Thesis in Public Health of the principal author. This study was self-funded by the principal author.

Data availability

The dataset used for analysis during this study is available from the corresponding author on reasonable request.

Declarations

Ethics approval and consent to participate

Ethical clearance for this study was obtained from the Ethics Review Committee, Faculty of Medicine, University of Kelaniya (Ref. No: P/22/03/2021). Informed written consent was obtained from all participants, and all methods involving human participants were conducted in adherence to ethical standards.

Consent for publication

This manuscript does not contain data from any individual person, hence consent for publication is “Not applicable”.

Competing interests

The authors declare no competing interests.