Assessing the Affective Organizational Commitment Scale: A Multilevel Psychometric Study among Prison Staff

Conceptualizing Affective Organizational Commitment

The idea of affective organizational commitment originates from a more general concept of organizational commitment. The literature on organizational commitment and its history is extensive and complex, mostly due to the large number of (sometimes conflicting) conceptualizations and operationalizations of organizational commitment (Mercurio, 2015). Despite the many interpretations of the concept, Meyer and Herscovitch (2001) argued that organizational commitment fundamentally represents a work attitude that psychologically binds a worker to the employing organization. Most recent approaches to organizational commitment involve multidimensional conceptualizations, of which Allen and Meyer’s (1990) three-component framework remains the most influential.

Allen and Meyer (1990) developed their model in an effort to build upon research and overcome conceptual confusion. Depending on how the bond is formed, three types of commitment can be distinguished: affective, normative or continuance commitment. Affective commitment refers to the individual’s “emotional attachment to, identification with, and involvement in, the organization” (Meyer & Allen, 1991, p. 67). Normative commitment relates to a moral obligation to the organization due to social pressure. Continuance commitment is an attachment created by a recognition of the costs associated with leaving the organization. Research indicates that not all types of commitment influence measures of organizational effectiveness to the same extent, with affective commitment being the most beneficial (Iverson & Buttigieg, 1999; Meyer et al., 2002). In addition, empirical researchers have criticized the definition of continuance and normative commitment and the overlap between normative and affective commitment (Solinger et al., 2008; Stazyk et al., 2011). For these reasons, studies have been advised to take an exclusively affective approach to organizational commitment (Mercurio, 2015).

The concept of affective commitment involves a number of theoretical propositions. First, it is believed that the congruence between the values of the worker and the organization forms the basis for the development of the affective bond (Meyer & Herscovitch, 2001). Affectively engaged workers are not only likely to share the company’s values, but “experience self-set and assigned goals as autonomously regulated (. . .) and as ideals to be achieved” (Meyer et al., 2004, p. 1001). Second, affective commitment specifically reflects the desire to remain working for the organization, as opposed to a continued involvement based on a moral obligation or a recognition of the costs associated with leaving the organization (Allen & Meyer, 1990). Third, affective commitment is expected to be the result of the interplay between a worker and their work environment, and to lead to desirable outcomes through a motivational mechanism (Cropanzano & Mitchell, 2005; Ostroff & Aumann, 2004). In the early 1990s, prison researchers interested in the work environment of prison staff started to include organizational commitment as a predictor of turnover and prosocial organizational behavior in their research (Camp, 1994; Culliver et al., 1991. In the Prison Social Climate Survey, used to assess the social environment of correctional institutions in the United States, organizational commitment was operationalized as the general extent to which prison workers had a positive outlook on the prison organization. Research on affective organizational commitment in a prison setting gained momentum after the turn of the century, when Eric Lambert and his colleagues (1999, 2005) underlined the importance of affective organizational commitment of prison staff in a number of influential studies. A systematic and meta-analytic review of the literature on correctional officers conducted by Butler and colleagues (2019) shows that organizational commitment is a frequently studied outcome in correctional research today.

Measuring Affective Organizational Commitment

The most well-known instruments to measure affective organizational commitment are the Organizational Commitment Questionnaire (OCQ) (Mowday et al., 1979) and the Affective Organizational Commitment Scale (AOCS) (Allen & Meyer, 1990). Although other measures have been designed (see, for instance, Dick & Metcalfe, 2001; Sergeant & Frenkel, 2000), they have not been as thoroughly psychometrically tested, or align poorly with the applied definition of commitment (Cook & Wall, 1980; Meyer & Allen, 1991).

Almost all prison researchers use the Organizational Commitment Questionnaire [OCQ] (Mowday et al., 1979) to measure affective organizational commitment (Lambert et al., 2010, 2016, 2017). The OCQ is a 15-item scale reflecting an acceptance of organizational values, a willingness to exert effort on behalf of the organization and a desire to maintain membership in the organization. While the instrument is generally seen as a reliable and valid measurement of affective organizational commitment (Meyer & Allen, 1991), some criticism has been provided. First, the OCQ does not exclusively include items assessing an affective attitude toward the organization, but also items that reflect moral obligations and behavioral intentions. Therefore, in the scale, the construct of affective organizational commitment appears to merge with related but different attitudes and behaviors. Second, some have argued that the OCQ is not suitable for predicting outcomes such as turnover or performance, as certain scale items already reflect these outcomes (Griffeth et al., 2000). This overlap raises some questions about the validity of the scale in regard to its prediction.

The limitations of the OCQ resonate with developments in the broader literature outside the prison context, where the measurement of affective organizational commitment has been refined. Allen and Meyer’s AOCS was designed to build upon existing instruments and overcome some of their limitations (Allen & Meyer, 1990). First of all, as the scale is based on a multidimensional conceptualization of organizational commitment, scale items are not mistakenly intertwined with other components of organizational commitment. Second, the AOCS measures emotional attachment and involvement rather than turnover or performance, and is therefore better suited to predict these outcomes. Researchers have subjected the scale to scrupulous psychometric evaluation, confirming its validity and reliability in various occupational contexts (Allen & Meyer, 1996; Iverson & Buttigieg, 1999; Meyer et al., 1993). Although the AOCS is the most established instrument for assessing affective organizational commitment outside the prison context (Mercurio, 2015), to our knowledge, only three prison studies have used the scale to measure affective organizational commitment of prison workers (Aube et al., 2007; Karaaslan & Aslan, 2019; López-Cabarcos et al., 2016). The AOCS provides potential for strengthening the study of affective organizational commitment in prison settings, but its psychometric properties have never been tested in a prison context. Prison organizations have a unique and demanding work environment that differs from more traditional workplaces such as schools, offices, or hotels. The prison work environment, similar to the military (Gade, 2003), is characterized by strict hierarchical structures, security protocols, and a focus on risk management. Prison staff are expected to work under stressful conditions and need to rely on their colleagues when facing potential safety threats (Liebling et al., 2010). This unique context is not only likely to impact how prison staff perceive their organization but also the psychological mechanisms underpinning their commitment.

Factor Structure

Affective organizational commitment traditionally measures a commitment toward the entire organization (Allen & Meyer, 1990). However, as organizations consist of multiple layers, commitment researchers have long since established that workers can be (differently) committed to various parts of their organization (Becker, 1992; Wombacher & Felfe, 2017). In this paper, these different entities to which a person can be committed are labeled as commitment foci (Meyer & Allen, 1997). Research outside the prison context indicates that commitment foci are factorially distinct, can relate to different antecedents, and that commitment to a more direct focus (such as the team) might reinforce feelings of commitment to a more distant focus (such as the entire organization) (Vandenberghe et al., 2004; Wombacher & Felfe, 2017). While a scarcity of prison scholars considers the prison agency and facility as different foci (Garland et al., 2009; Lambert et al., 2008), they have not assessed the factor structure of the included commitment scale, studied team as a focus of commitment, or considered that specific antecedents might be related to specific foci.

Construct Validity

In light of the favorable outcomes as established in other fields, prison researchers have focused on studying predictors of affective commitment (Lambert et al., 2021). These predictors are usually grouped into characteristics related to the person, job or organization. Generally, job and organizational variables correlate more strongly with affective commitment than personal characteristics (Meyer & Allen, 1991; Meyer et al., 2002). Social exchange theory can help explain how especially positive work experiences, such as supervisor consideration, have been found to contribute to the development of affective commitment of prison staff (Lambert et al., 2021). According to social exchange theory, relationships between parties may “evolve into trusting, loyal, and mutual commitments,” depending on the perceived effort demonstrated by these parties (Cropanzano & Mitchell, 2005, p. 875). In this regard, employees may reciprocate organizational efforts such as supervisor consideration with affective attachments, bringing about strong and lasting relationships.

In addition, as discussed, employees are more likely to develop affective organizational commitment when they perceive congruence between their values and those of the organization (Meyer & Herscovitch, 2001). Given that affective organizational commitment is characterized by a voluntary and emotional psychological bond, this tendency is expected to be particularly strong when the organization’s values reflect socially driven themes such as humanity (Abbott et al., 2005; Finegan, 2000). This proposition is illustrated by the person-environment fit theory that underlines the match between a person and their work environment (Ostroff & Aumann, 2004). Accordingly, affective organizational commitment can arise when the employee’s value system is in agreement with that of the organization. A handful of studies investigated whether affective organizational commitment of prison workers could be predicted by their correctional orientation (Lambert et al., 2021). However, there has not yet been an explicit effort to examine whether different values of prison agencies are associated with affective organizational commitment of prison workers.

Current Study

The aim of the current study is twofold: it (a) strives to strengthen the study of affective organizational commitment in prison by using a measurement that overcomes some of the shortcomings of other scales, and (b) seeks to assess the psychometric qualities of this measurement in a prison context by testing its factor structure, reliability, and construct validity in this unique setting. Therefore, it seeks to answer the following sub-questions:

Building on the identified theoretical propositions and gaps in empirical literature, the current study seeks to test several hypotheses related to these research questions. We hypothesize that affective organizational commitment foci are factorially distinct, that workers differ in their levels of affective organizational commitment dependent on these foci, and that antecedents differently relate to different foci. Furthermore, we expect that positive job and organizational factors raise affective organizational commitment, with positive work experiences providing the strongest associations. Finally, we hypothesize that organizational values specifically communicated by prison organizations are associated with affective commitment, with humanity-related values providing the strongest associations.

The innovative aspects of this study lie in validating an affective commitment scale that addresses the previously identified limitations of other measures, including three different foci, and applying a multilevel approach that accounts for the structure of prison organizations. We use a combination of self-reported and administrative data from a large and nationwide sample of Dutch prison employees across different job roles. By evaluating the psychometric properties of the scale, we aim to inform the field of prison research and provide a clear foundation for the study of affective organizational commitment in prison.

Dutch Prison Context

As part of the Dutch Ministry of Justice and Security, the Custodial Institutions Agency (DJI) is responsible for the execution of all sentences and custodial measures on a national scale. The Agency is in charge of Juvenile detention centers, psychiatric penitentiary facilities, detention centers for foreign citizens, and correctional facilities. The Netherlands has 23 overarching penitentiary institutions (with regimes varying from minimum to maximum security) spread across 28 locations, as some institutions consist of multiple physical sites that are managed under a single administrative umbrella. Some staff members can be assigned across these locations within the same institution. Their work should be considered against the backdrop of increasingly challenging working conditions (Van Ginneken et al., 2020). In recent years, the Dutch prison sector has been dealing with high workload, absenteeism, and staff shortages (FNV, 2020). Due to rising demands and a lack of personnel, many staff members feel overworked and have experienced an increase in dangerous work situations. Staff shortages have forced the government to take measures, such as cutting activities in prison, releasing certain detained individuals earlier than planned, and temporarily suspending calling in self-reporters (Roks et al., 2024; Tweede Kamer, 2024). According to its mission statement, the Agency “contributes to a safer society by implementing custodial measures and prison sentences and by giving the people in its care the opportunity to build up a socially acceptable life” (DJI, n.d.). This statement reveals safety and rehabilitation as two organizational values. Although retribution is not mentioned in mission statements or policies of the Agency itself, the Dutch government expressed that the enforcement of prison sentences should be primarily focused on retribution (Struijk, 2020). Nevertheless, the humane treatment of detained individuals remains an important value of the Agency, reflected by its vow to promote a positive prison climate (van Ginneken et al., 2018).

Method

Results

Descriptive statistics of the included variables are provided in Table 1. In general, prison workers felt more affective commitment to their team (M = 3.22) than to their prison facility (M = 3.06) or the overarching prison agency (M = 3.01). Results of a repeated measures ANOVA and post hoc pairwise comparison confirmed that affective organizational commitment significantly differed across all three foci (team, prison facility, and prison agency), F(1.481, 3,948.776) = 134.73, p < .001. The scores on affective commitment of the included job roles were largely in line with the scores of all prison workers. Furthermore, results revealed that respondents scored significantly higher on the value of safety (M = 3.79, p < .001) than all other values, followed by the value of rehabilitation (M = 3.70, p < .001). Scores on the predictor variable of role clarity showed that respondents, on average, had a clear understanding of the responsibilities and tasks related to their work role (M = 4.19), felt supported by their supervisor (3.91), and were quite satisfied with the provided training (3.48). Notably, staff members reported an average level of work pressure (2.45).

OPEN IN VIEWER

Table 1:

Descriptive Statistics

Variable	Sample	n	M	SD	Items
Affective commitment
Agency	All prison workers	2,667	3.01	0.73	8
	Correctional officers	1,074	2.98	0.75	8
	Security officers	878	3.06	0.72	8
	Labor supervisors	374	3.00	0.74	8
	Case managers	276	2.96	0.65	8
	Intake officers	65	2.87	0.62	8
Facility	All prison workers	2,667	3.06	0.76	8
	Correctional officers	1,074	3.03	0.76	8
	Security officers	878	3.13	0.75	8
	Labor supervisors	374	3.05	0.80	8
	Case managers	276	3.01	0.69	8
	Intake officers	65	2.99	0.65	8
Team	All prison workers	2,667	3.22	0.78	8
	Correctional officers	1,074	3.30	0.79	8
	Security officers	878	3.11	0.74	8
	Labor supervisors	374	3.14	0.85	8
	Case managers	276	3.26	0.62	8
	Intake officers	65	3.48	0.71	8
Predictors
Role clarity	All prison workers	2,649	4.19	0.55	5
Work pressure	All prison workers	2,633	2.45	0.98	3
Quality of training	All prison workers	2,631	3.48	0.85	3
Supervisor support	All prison workers	2,645	3.91	0.98	2
Rehabilitation	All prison workers	2,622	3.70	0.73	3
Retribution	All prison workers	2,651	3.37	0.79	3
Humane treatment	All prison workers	2,648	3.38	0.69	3
Safety	All prison workers	2,655	3.79	0.64	3

Chi-square tests and an independent t-test were used to assess whether respondents and non-respondents significantly differed across a number of personal and job-related characteristics. Chi-square test results revealed significant (but small) differences between respondents and non-respondents for job role. Security officers were slightly underrepresented χ²(1, N = 5,828) = 127.42, p < .001. In contrast, case managers and labor supervisors were more likely to respond. Furthermore, results revealed a significant difference in age between survey respondents (M = 45.27, SD = 12.88) and non-respondents (M = 42.84, SD = 13.60), t(5,748.09) = 7.0, p < .001. On average, survey respondents were 2.44 years older (95% CI [1.75, 3.12]). Nevertheless, the effect size measured by Cohen’s d was small (d = 0.18).

Reliability

The internal consistency of the AOCS was examined using Cronbach’s alpha and split-half reliability. Although studies tend to differ in their description of acceptable alpha scores, alpha scores between 0.71 and 0.80 are usually deemed as satisfactory, and scores between 0.81 and 0.90 are seen as excellent (Tavakol & Dennick, 2011). A Spearman Brown score of 0.70 or higher is usually considered acceptable reliability (de Vet et al., 2017). As can be seen in Table 2, the Cronbach’s alpha values for all three scales were excellent, ranging from 0.83 to 0.92 across all prison workers as well as the different job roles. The second method to measure internal consistency was split-half reliability. Table 2 shows that Spearman Brown scores for all three affective organizational commitment foci were excellent across all prison workers, as well as the different job roles. Only among intake officers, the reliability of the scales that measure affective commitment to the prison agency (r = 0.76) and facility (r = 0.75) was somewhat lower, though still within an acceptable range.

OPEN IN VIEWER

Table 2:

Reliability Scores

Commitment	All workers		Correctional		Security		Labor		Case manager		Intake
	α	r	α	r	α	r	α	r	α	r	α	r
Agency	0.86	0.82	0.86	0.82	0.86	0.82	0.87	0.84	0.85	0.82	0.83	0.76
Facility	0.87	0.85	0.87	0.85	0.87	0.85	0.89	0.89	0.87	0.83	0.84	0.75
Team	0.90	0.86	0.89	0.85	0.89	0.85	0.92	0.88	0.89	0.85	0.91	0.88

Factor Structure

To assess the factor structure of the AOCS, we performed a confirmatory factor analysis (Brown, 2015). We tested two models. The first (one-factor) model represented one construct, including all indicators of the affective commitment measures. The second (three-factor) model considered three separate constructs (i.e., commitment to the prison agency, facility, and team), each measured by their own set of indicators. Because of the repeated use of identically worded items across the three affective commitment foci, the error terms of these items were likely to exhibit a shared variance. We therefore allowed the error terms between identically worded items to covary (Vandenberghe et al., 2004). In addition, to account for potential links among the three scales, the three foci were modeled as separate but related factors by allowing covariances between latent factors. We rely on the following fit statistics to assess model quality: the Comparative Fit Index (CFI), Tucker–Lewis Index (TLI), and Root Mean Square Error of Approximation (RMSEA) (Schumacker & Lomax, 2016). We used a likelihood-ratio test to compare the two nested models and inspect whether the more complex model provided a significantly better fit to the data than the one-factor model.

Table 3 shows that the fit of the AOCS substantially improved from the one-factor model to the three-factor model. Most importantly, the log likelihood ratio test confirmed that the three-factor model fit the data significantly better than the one-factor model χ²(3), 8,787.12, p < .001. The fit statistics presented in Table 3 furthermore show that the three-factor model provided a similarly good fit across the different job roles. ¹ A series of log likelihood ratio tests verified that the three-factor model fit the data significantly better than the one-factor model across all job roles. This result demonstrates that the three affective organizational commitment foci are factorially distinct for prison workers regardless of job role. The confirmatory factor analysis further indicated a high covariance between affective commitment to the prison agency and the facility (0.87). In comparison, the covariances between affective commitment to the team and the facility (0.53), as well as the agency (0.46), were lower. While affective commitment to the prison agency, facility and team are factorially distinct, the high covariances between affective commitment to the agency and facility indicate that these foci are more similar than affective commitment to the team.

OPEN IN VIEWER

Table 3:

Confirmatory Factor Analysis Fit Statistics

Subsample	χ²	df	P	CFI	TLI	RMSEA
All prison workers
One-factor model	10,710.77	228	< .001	0.72	0.66	0.15
Three-factor model	1,923.64	225	< .001	0.96	0.95	0.06
Correctional officers
One-factor model	5,070.76	228	< .001	0.68	0.61	0.15
Three-factor model	1,063.38	225	< .001	0.95	0.93	0.06
Security officers
One-factor model	2,764.70	228	< .001	0.79	0.75	0.12
Three-factor model	795.16	225	< .001	0.95	0.94	0.06
Labor supervisors
One-factor model	1,673.18	228	< .001	0.76	0.71	0.15
Three-factor model	537.32	225	< .001	0.95	0.94	0.07
Case managers
One-factor model	1,416.72	228	< .001	0.69	0.62	0.15
Three-factor model	369.08	225	< .001	0.96	0.95	0.05

Note. CFI = comparative fit index; TLI = Tucker–Lewis index; RMSEA = root mean square error of approximation.

Construct Validity

Construct validity was assessed by evaluating the underlying theoretical and empirical assumptions of affective commitment and its correlates. Since the three affective commitment scales were found to be factorially distinct, we inspected associations between predictors and all three foci separately, using multilevel linear regression analyses. The data were structured hierarchically across three levels: individuals (Level 1) nested within teams (Level 2) nested within prison facilities (Level 3). To inspect the proportion of variance attributable to each level, we calculated ICCs based on empty models for each of the three commitment foci (see Table 4). The ICCs represent the ratio of between-group variance to the total variance. The results reveal that a substantial proportion of the variance for all three commitment foci is attributable to the team level, especially regarding affective commitment to the team. Only a small proportion of the variance is attributable to the facility level. Nevertheless, regression analyses were run with 3 levels to account for the nested structure of the data. Based on methodological recommendations and sample size, higher order levels were required to have a minimum of five observations (Snijders & Bosker, 2011). Therefore, out of 2,667 participants in 446 teams, 2,208 participants in 275 teams were included in multilevel linear regression models. The fixed effects coefficients and standard errors are reported in Table 5. The multilevel linear regression of correctional officers, security officers, labor supervisors, and case managers is provided as supplementary materials (see Tables S2-S5).

OPEN IN VIEWER

Table 4:

Multilevel Linear Regression Analysis Null Models With Random Intercepts and ICCs Per Level

Level	Agency		Facility		Team
	Variance	ICC%	Variance	ICC%	Variance	ICC%
1. Individual	0.483	–	0.518	–	0.481	–
2. Prison team	0.024	4.9%	0.028	5.1%	0.105	17.7%
3. Prison facility	0.009	1.7%	0.022	3.8%	0.009	1.6%

Note. ICC = intraclass correlation coefficients.

OPEN IN VIEWER

Table 5:

Multilevel Linear Regression Analysis of Affective Organizational Commitment (N = 2,208 Across 275 Teams Within 23 Prison Facilities

Variable	Agency		Facility		Team
	B	SE	B	SE	B	SE
Job and organization
Role clarity	0.08**	0.03	0.13***	0.03	0.15***	0.03
Quality of training	0.15***	0.02	0.14***	0.02	0.07***	0.02
Work pressure	0.03*	0.02	0.03*	0.02	0.04*	0.02
Supervisor support	0.15***	0.02	0.15***	0.02	0.17***	0.02
Organizational values
Rehabilitation	0.05*	0.02	0.03	0.02	0.01	0.02
Retribution	-0.003	0.02	-0.01	0.002	0.03	0.02
Humane	0.16***	0.02	0.18***	0.02	0.11***	0.02
Safety	0.05*	0.03	0.04	0.03	0.04	0.03
Control variables
Male	0.10***	0.03	0.11***	0.03	0.11***	0.03
Age	0.003*	0.001	0.003*	0.001	-0.003**	0.001
Correctional officers	-0.16***	0.04	-0.16***	0.04	0.15**	0.05
Labor supervisors	-0.17***	0.05	-0.17**	0.06	0.03	0.08
Case managers	-0.15**	0.06	-0.13*	0.07	0.20*	0.09
Intake officers	-0.25***	0.13	-0.26	0.14	0.33*	0.16

Note. Sex is dummy coded with female as the reference category. The job role is dummy-coded with security staff as the reference category.

*

p < .05. **p < .01. ***p < .001.

All job and organizational factors are significantly related to the affective organizational commitment foci (see Table 5). Notably, work pressure had positive associations with affective organizational commitment in the multilevel regression analyses, while Pearson correlations demonstrated negative associations between work pressure and affective commitment to the prison agency (r = -0.05, p < .003) and facility (r = -0.06, p < .001). Standardized regression coefficients revealed experienced support provided by the supervisor as the strongest linkage to affective commitment to the prison agency (β = 0.14, p < .001), facility (β = 0.14, p < .001), and team (β = 0.16, p < .001). While perceived quality of received training also provided a strong association to affective commitment to the prison agency (β = 0.14, p < .001) and facility (β = 0.13, p < .001), the predictor was less important regarding affective commitment to the team (β = .05, p < .001). It appears that the strength of associations between affective organizational commitment and job and organizational variables depended on the commitment foci being measured.

The multilevel linear regression analyses among the different job roles generally mirrored those of all prison workers taken together, although some results did not reach statistical significance. For correctional officers, only the organizational variable of supervisor support was significantly associated with affective organizational commitment across all foci (see Supplementary Table S2). In line with the sample including all prison workers, supervisor support most strongly related to affective commitment to the prison agency (β = 0.13, p < .001), facility (β = 0.12, p < .001), and team (β = 0.19, p < .001). For security officers (Supplementary Table S3) as well as labor supervisors (Supplementary Table S4), supervisor support and quality of training were associated with all affective organizational commitment foci. As for case managers (Supplementary Table S5), the smallest subsample, the results were either inconsistent or not significant.

Next, as shown in Table 5, the values of rehabilitation, humane treatment, and safety were significantly associated with affective commitment to the entire organization (agency). Notably, only the value of humane treatment was significantly associated with affective commitment to the other two foci. The value of retribution was not significantly associated with any of the affective organizational commitment foci. Standardized coefficients showed that humane treatment provided the strongest associations with affective commitment to the prison agency (β = 0.11, p < .001), facility (β = 0.13, p < .001), and team (β = .09, p < .001). The results of the multilevel linear regression analyses among the different job roles show that only the value of humane treatment was significantly associated with affective organizational commitment across the three foci for both correctional and security officers (see Supplementary Tables S2–S3). Also, in line with the sample including all prison workers, humane treatment most strongly related to affective commitment to the prison agency, facility, and team for correctional (β = 0.15, β = 0.16, β = 0.10, p < .001) and security (β = 0.14, β = 0.16, β = .09, p < .001) officers. In regard to the two smaller samples of labor supervisors and case managers, results were not statistically significant (see Supplementary Tables S4–S5).

Discussion

This study aimed to assess the suitability of the AOCS (Allen & Meyer, 1990) in a prison context by testing its reliability, factor structure, and construct validity in a large and nationwide sample of Dutch prison workers. While research indicates the relevance of affective organizational commitment to prison management, employees, and detained individuals (Lambert et al., 2021), most prison studies use a less refined measurement of affective commitment and have not tested the psychometric properties of included scales in the context of prison work. The current study addressed these issues through a multilevel psychometric assessment of the acclaimed AOCS in regard to staff members from a variety of different job roles employed in 23 correctional facilities spread throughout the Netherlands.

The findings confirm the scale’s reliability when applied to the full sample as well as across different job roles. The analyses furthermore show that affective commitment to the prison agency, facility, and team is factorially distinct, although commitment to the facility and team appear more closely related than to the more distant prison agency. Multilevel null models furthermore show that a substantial proportion of the variance lies at the team level across all three commitment foci. This indicates that team-level factors (in addition to individual factors) best explain prison workers’ affective organizational commitment. In terms of construct validity, the concepts are linked to theoretical and empirical assumptions regarding job and organizational characteristics (Lambert et al., 2010, 2016; Meyer et al., 2002). Value congruence is mostly associated with affective commitment to the prison agency, although the value of humane treatment aligns strongly across all foci as anticipated (Abbott et al., 2005; DJI, n.d.; Finegan, 2000). While patterns of associations between predictors and affective organizational commitment across the job roles largely mirror those of all prison workers taken together, some results fail to reach statistical significance. This can be explained by the substantial differences in sample sizes and, consequently, statistical power. Nevertheless, the strong reliability and consistent factor structure of the scale across all five job roles confirm its robustness and suitability for use in a prison setting.

In addition to confirming the relevance of the AOCS in a prison context, the results of the current study advance the conceptual understanding of affective organizational commitment in this unique environment. First, the established factor structure of the scale demonstrates that affective organizational commitment operates at multiple levels of the prison organization, emphasizing the need to consider the target of affective commitment in theoretical models. The finding that prison workers are differently committed to various parts of the prison organization aligns with research conducted outside the prison context (Becker, 1992; Vandenberghe et al., 2004; Wombacher & Felfe, 2017). The results additionally illuminate how the work team as an organizational focus may be particularly relevant for understanding affective commitment in a penitentiary environment.

Second, the substantial team-level variance indicates that affective organizational commitment is not only shaped by individual characteristics but also by the immediate work environment. In correctional settings, such team-level differences may arise from shared experiences within teams, such as collective perceptions of safety or consideration. These collective experiences can foster feelings of support and belonging, and are key in strengthening affective attachments. As the first study to examine affective commitment of prison workers using a multilevel approach, these findings shed light on how affective commitment develops within correctional organizations. Third, the importance of positive work experiences in relation to affective organizational commitment as found in this study aligns with theoretical expectations derived from social exchange theory, as well as empirical findings as established in other studies (Lambert et al., 2010, 2016; Meyer et al., 2002). These last two findings collectively underline the theoretical notion that affective commitment is, to a large extent, the result of the interplay between a worker and their work environment.

Last, the links between affective commitment and organizational values support the theoretical proposition that value congruence is imperative to the development of the affective bond, and are in line with the person-environment fit theory. The discovery that the value of humanity related most strongly to all commitment foci is in line with previous findings (Abbott et al., 2005; Finegan, 2000). Value congruence appears most likely to lead to affective commitment if the organization’s values reflect prosocial clusters (Abbott et al., 2005).

Conclusion

This article highlights the strong psychometric properties of the AOCS for Dutch prison workers and recommends its use for exploring the links between affective organizational commitment and outcomes such as well-being, absenteeism, turnover, and perceptions of prison climate. The insights provided by the current study advance the conceptual understanding of affective commitment in a prison context by emphasizing its multifocal, multilevel, and value-driven nature. These results underline the need to refine the current knowledge of affective organizational commitment by acknowledging the existence of different organizational foci. Researchers should be aware that prison workers feel differently committed to various parts of their organization and that predictors are differently associated with these foci. Furthermore, the findings suggest that prison management would do well to consider team-level strategies to reinforce affective commitment, such as peer mentorship programs or shared goal setting, in addition to implementing individual-level interventions, such as providing regular constructive feedback. More nuanced research on predictors and outcomes of affective organizational commitment can help prison management where they can make changes in their organization to foster affective commitment of prison workers, and potentially improve living conditions for detained individuals.

Supplemental Material