Adaptation and Validation of Indonesian Version of the Commitment to Change Scale

Introduction

An organization is always changing. These changes are often partly unplanned and gradual.¹ Therefore, organizations must find solutions to these challenges and problems if they want to survive, prosper, and perform effectively.² Organizations that are able to adapt to these changes will be able to develop over a long period of time and overcome threats caused by the internal and external environment.³ With no exception for higher educational institutions.

Organizational change in a higher educational institution is affected mostly by external pressures, such as government rules and regulations that occur continuously. This is done to continue to adapt to the times and increase competitiveness with other countries. For this reason, higher educational institutions in Indonesia are required to always be ready to face and anticipate changes, so that they can continue to compete and contribute to improving the quality of human resources. Mangundjaya explains that in order to survive and compete, every organization must change, and this requires commitment from its employees to change.⁴

Herscovitch & Meyer⁵ stated that commitment to change was the key point to implementing transformation. If each member of the organization has a commitment to change, the transformation will be successfully applied.⁶ Previous studies have indicated that positive employee attitudes such as commitment to change play a vital role in employee acceptance of organizational change and its long-term success.^6–8 Employee commitment to change is also one of the most important antecedents to avoid failure in change implementation.^5,9

The word commitment can be defined as an employee’s attachment to various foci such as the organization as a whole, units within the organization, supervisor or even a change.^5,10 In 1991, Meyer & Allen¹¹ proposed a model of organizational commitment. Organizational commitment is defined as a psychological state that leads an employee to maintain his or her organizational membership.^11,12 Furthermore, Meyer and Herscovitch¹³ made some adjustments from the original model and proposed a general model of workplace commitment so that it could be applied to other workplace commitments. With this change, commitment has changed its definition as “a force (mindset) that binds an individual to a course of action of relevance to one or more targets”.⁵

In contrast to previous research which typically defined commitment to change as a uni-dimensional construct,^14,15 Armenakis, Harris and Field¹⁶ and Herscovitch and Meyer⁵ were the first researchers to describe commitment to change as a multi-dimensional construct. Based on the general model of workplace commitment they had presented previously,¹³ Herscovitch and Meyer⁵ further introduced a three-component model of commitment to change. They defines a commitment to change as a force (mind-set) that binds an individual to a course of action deemed necessary for the successful implementation of a change initiative.⁵ This mindset can take different forms: (a) a desire to provide support for change based on a belief in its inherent benefits (affective commitment to change), (b) a recognition that there are costs associated with failing to provide support for change (continuous commitment to change), and (c) a sense of obligation to provide support for change (normative commitment to change). Herscovitch and Meyer⁵ proved using the EFA test that commitment to change consisted of three factors where the correlation test between factors showed different magnitudes and directions, this indicates that commitment to change is a multidimensional variable.

Until now, there are not many studies that focus on adaptation and validation commitment to change scale. However, this scale becomes very important because change occurs continuously, and employee commitment to change has a very important role in providing success for these changes. Some studies on adaptation and validation of commitment to change were carried out in Canada,¹⁷ Pakistan,^3,17 India,¹⁸ and Turkey.¹⁹ In Indonesia itself, although there have been many studies discussing commitment to change,^20–22 no studies have been found on the adaptation and validation of commitment to change scale.

Based on previous studies regarding the validation of commitment to change, several differences were found in the results of their studies. For example, the results of a study by Meyer et al¹⁷ on an Indian sample showed that the correlation between affective and normative commitment to change was greater than continuance and normative commitment to change. This is different from the two subsequent studies where kalyal et al³ who used Pakistani samples found that the relationship between continuance and normative commitment to change was more significant than affective and normative commitment to change, as well as the study conducted by Soumdjaya et al,¹⁸ although both used samples of Indian society, the results obtained were closer study conducted by Kalyal et al compared to Meyer et al. This suggests that even within the same culture, the relationship of normative commitment to change to the other two components of commitment to change varies and more similar studies are needed to gain further insight into them.¹⁸

Providing an original scale to people with different languages and cultures certainly has its own obstacles. Some items may not be appropriate due to cultural differences. Thus, careful and in-depth translation and development of culturally appropriate items may be required to address comprehension issues.³ This cultural incompatibility sometimes requires researchers to adjust or eliminate the items to be used as happened in the previous studies.^3,18 Beaton²³ said that in conducting research on subjects that have different languages and cultures, a poor translation process can lead to instruments that are not equivalent to the original questionnaire. Generic scales are not always culturally sensitive, and thus may need to be adapted for certain contexts. For this reason, it is not permissible to directly use existing or validated scales in other countries that have different languages and cultures. This raises concerns about the misinterpretation of the question.

On the other hand, changes in various organizations that take place continuously make the commitment to change scale a tool that will be needed by many authors and practitioners to determine employee commitment to the changes that occur. Thus, the authors sees that a valid and reliable commitment to change scale will be very important to be used in Indonesia, especially to obtain data and evidence more objectively and scientifically.

Based on these, this study aims to adapt and validate the commitment to change scale (affective, continuance, and normative commitment to change) in the Indonesian Version. The final Indonesian version of the Commitment to Change Scale was tested using the following: (1) to examine its content validity by using Content Validity Index (CVI), (2) to examine its construct validity by using Confirmatory Factor Analysis (CFA), (3) to examine its Convergent and Discriminant validity by investigating the relationship between commitment to change with job satisfaction and perceived Stress. Authors correlated commitment to change with job satisfaction and job stress because in previous studies, job stress^24,25 and job satisfaction^20–22,26 were shown to have a significant relationship with commitment to change, (4) reliability was verified using Cronbach’s alpha.

Methods

Procedure

This research was conducted through several stages in accordance recommended by Beaton et al.²³ The first step was to ask permission from the people who developed the Commitment to Change Scale measuring instrument.⁵ The next stage is to carry out a cross-cultural adaptation, starting from (1) forward translation by two people, who had English education and translation study backgrounds, (2) translation synthesis, which is discussing the results of the translation among the authors, translator 1 and translator 2, (3) backward translation by natives with applied linguistic backgrounds and another person with English education backgrounds, (4) expert committee review, where the minimum composition consists of methodologist, professional in the research field, language professionals, and translators (forward and backward translators) involved in the process to date.²³ In this study, the expert committee was reviewed by 2 translators, 3 people who mastered psychological concepts, 1 methodologist, and 1 linguist. (5) pre-testing with 30 subjects. Furthermore, the authors carried out the validity testing phase by testing content validity, construct validity, and reliability.

In the process of collecting data, authors go to several higher educational institution to ask for permission to conduct research in their place. Higher educational institution that provide support will provide access to be able to distribute online questionnaires to faculty members. Before filling out the scale they were asked to read and fill out a consent form regarding their willingness to fill out the scale.

Participants

The participants of this study were purposively sampled based on a study conducted by Beaton et al²³ regarding the process of cross-cultural adaptation with self-reported measures. Participants in this study were 204 faculty members from six universities in Indonesia who had worked for more than 6 years, this was done to ensure they had known and experienced two work situations, before and after the change. Researchers have suggested that the sample used should be no less than 200, whereas in terms of the ratio of observations to variables, the general rule is to have a minimum value from five times as many observations as the number of variables to be analyzed, and a more acceptable sample size has a ratio of 10:1.²⁷ In this study there are 18 variables to be analyzed so that it meets the recommended minimum number.

Based on the demographic data obtained in this study, it can be described that most of the participants were female 57.4% (N = 117). Participants were aged 25 years and over with the majority being in the age range of 25–34 years, 46.1% (N = 94) and the others are 35–44 years, 41.7% (N = 85); 45–54 years, 10.8% (22); and over than 55 1.4% (3). Based on work tenure, most have worked for 7–12 years 77.4% (N=158), the others is 13–18 years 10.3% (N=21) and more than 18 years 12.3% (N = 25). Based on academic rank there were 93.1% (N=190) assistant professors and 6.9% (N=14) were associate professors. Based on the educational level there were 89.7% (N=183) participants with master’s degree and 10.3% (N=21) were doctoral degree.

Study Measures

To measure commitment to change we used a measuring tool developed by Herscovitch & Meyer.⁵ This tool consists of 3 components; Affective Commitment to Scale (ACTC), Continuance Commitment to Scale (CCTC), and Normative Commitment to Scale (NCTC). Examples of items that have been converted into Indonesian are: ACTC, eg item 1, “Saya percaya adanya manfaat pada perubahan ini” (I believe there is a benefit in this change). CCTC, eg item 7, “Saya tidak memiliki pilihan lain selain mengikuti perubahan ini” (I have no other choice but to follow this change). NCTC, eg item 13, “Saya merasa bertanggung jawab untuk mewujudkan perubahan ini” (I feel responsible for making this change happen). Each component consists of 6 questions with a total of 18 items on a scale of 1–7 (1 = strongly disagree to 7 = strongly agree). The use of this scale refers to the scale developed by Herscovitch & Meyer.⁵

Job Satisfaction Survey (JSS) developed by Spector²⁴ was used to measure job satisfaction. JSS has 36 items to measure employee perceptions and attitudes including; salary, promotions, supervision, additional benefits, contingent rewards, operating conditions, co-workers, nature of work, and communication. Respondents were asked to indicate their views by filling in one of the six-point Likert-type scales from 1 (strongly disagree) to 6 (strongly agree). The research used JSS which was translated and adapted by Azra et al²⁵ with a reliability coefficient of 0.93. The Cronbach’s alpha for this study was 0.92.

Perceived Stress Scale (PSS) by Cohen & Janicki-Deverts,²⁶ was used to measure job stress. Respondents were asked to indicate the frequency by filling in one of the 5-point Likert-type scales starting from never to very often. This study uses PSS which has been translated by Saraswati,²⁸ with a reliability coefficient of 0.90, by eliminating 5 items. The reliability coefficient in this study was 0.88, but none of the items were omitted.

Data Analyses

To see the socio-demographic picture such as gender, age, marital status, grade, education, and length of work in this study, descriptive statistics were used. Standard deviation and Mean were used to describe the variable. Data analysis used Content Validity Index (CVI) consisting of Item Content Validity Index (I-CVI) and Scale Content Validity Index (S-CVI). Then the authors also used Confirmatory Factor Analysis (CFA), convergent and discriminant correlation analysis, and reliability analysis.

In conducting content validity using CVI, Lynn²⁹ recommends a minimum of three experts, but preferably no more than 10. In this study, there were seven experts. A four-point ordinal scale was used to be rated by experts on each item according to relevance, clarity, simplicity and ambiguity,³⁰ which included the following: (1) not relevant; (2) somewhat relevant; (3) quite relevant; and (4) highly relevant. The I-CVI was calculated as the number of experts giving a rating of 3 or 4 divided by the total number of experts for each item. The I-CVI should be 1.00 when there are five or fewer experts. Meanwhile, if there are 6 or more experts, the standard can still be lowered, but not less than 0.78.³¹

CFA was tested using Lisrel 8.80. Convergent and discriminant correlations were tested using the Pearson correlation coefficient by correlating ACTC, CCTC, NCTC with JSS and PSS. Reliability was assessed by Cronbach’s alpha and Construct Reliability (CR).

Ethical Consideration

The ethics enforcement of this research is supplemented by Universitas Padjadjaran Research Ethics Committee License no. 824/UN6.KEP/ EC/2020 and informed consents.

Results

Demographic Characteristics of the Study Participants

Based on data analysis using t-test and ANOVA, demographic variables do not have implications for CCTC. The difference was found that age, tenure, academic rank, and education affect ACTC. Age and tenure also affect CCTC. Thus, it can be said that faculty members with an age range of 45-54 years have the highest affective commitment to change than any other age ranges. The higher their tenure, academic rank, and education, the higher their affective commitment to change. On the other side, the higher their age and working period will increase their normative commitment to change, but this will decrease when they reach the age of 55 and over. The results of the t-test and ANOVA are presented in Table 1.

Table 1 Demographic Characteristics of the Study Participants

Content Validity

In this study, authors used CVI to measure content validity. As noted by Lynn,²⁹ authors compute two types of CVIs. The first type involves the content validity of individual items and the second involves the content validity of the overall scale. Based on the results of the I-CVI measurement, it is known that there is one item less than 0.78 (I-CVI = 0.71). The authors revised this item before the CFA. At first the item reads “It would be too costly for me to resist this change”, and after the authors revised it to “I risked a lot of things if I resisted this change”. This is in accordance with Polit & Beck³¹ stating that information from the I-CVI is used by authors as a basis for revising, removing, or replacing items that are below the standard. S-CVI/Ave = 0.97 indicates that this scale has excellent content validity because of higher than 0.90.³¹

Confirmatory Factor Analysis

The results of the model fit test showed good fit results. This is because the results of the 6 indicators that there are 7 of them have met. To get good fit results, P value must be > 0.05, RMSEA Should be less than 0.80,³² while GFI, AGFI should be above or equal to 0.90,³³ so do NFI, NNFI and CFI.³⁴ The data in this study shows that χ2 = 240.73 (P = 0.000), RMSEA = 0.07, GFI = 0.90, AGFI = 0.93, NFI = 0.90, NNFI = 0.90, and CFI = 0.92.

Cronbach’s alpha obtained good results, with ACTC (α = 0.71); CCTC (α = 0.83); and NCTC (α = 0.77). Whereas in testing the Construct Reliability (CR) values obtained are also quite good with ACTC = 0.85; CCTC = 0.86; NCTC = 0.86.

Based on the CFA for the model, each item has a t value greater than 1.96 and a loading factor in the range of 0.50 to 0.86. If we break it down into each dimension, then the ACTC loading factor ranges from 0.50–0.74; CCTC = 0.51–0.83; and NCTC 0.50–0.86, we can see this in more detail in Table 2.

Table 2 Loading in Particular Items in ACTC, CCTC, and NCTC

Convergent and Discriminant Correlation

The convergent and discriminant validity of the Indonesian version of Commitment to Change was examined by using Pearson’s correlation to correlate commitment to change with job satisfaction and job stress, each component of commitment to change was correlated with job satisfaction and job stress to obtain a more specific picture. The correlation coefficient of each variable can be seen in Table 3.

Table 3 Inter Correlation Between ACTC, CCTC, NCTC, Job Satisfaction, and Job Stress

The table above shows that ACTC was positive and significantly related to Job Satisfaction with coefficient correlation of 0.271 and negative and significantly related to Job Stress with coefficient correlation of −0.282. CCTC and NCTC were not related both to Job Satisfaction and Job Stress. The highest correlation is ACTC with Job Stress, and the lowest is NCTC with Job Stress.

Discussion

The main aim of this research is to adapt and validate the commitment to change scale (affective, continuance, and normative commitment to change) in the Indonesian Version. This instrument has been tested using CVI, and CFA, as well as convergent and discriminant validation. All three components of the commitment to change scale were also found to have acceptable levels of reliability.

Based on the content validity test, it was found that the S-CVI was acceptable As for the I-CVI measurement, there is one item that is below the set standard. This happens because the word “mahal” which is a translation of the word “costly” is deemed inappropriate by some experts on the existing sentence For that reason, the authors revised it with more appropriate words, before being tested using CFA.

CFA test results showed that all items have a factor loading that exceeds the minimum requirement. This shows that the Commitment to Changing Scale is a measurable construct that can be distinguished from the others. Furthermore, in this study it was found that the highest loading factor was found in items that are part of the NCTC (0.86). This is not surprising considering that Indonesia is one of the countries that tend to have a collective culture compared to individual ones.³⁵

Another finding of this study is the higher correlation between ACTC and NCTC compared to others. This is in line with expectations from collective culture and reinforces other studies using Indian samples,¹⁷ although other studies have also found a greater correlation between ACTC and CCTC in Pakistan and Indian samples.^3,18 In this study, this may occur because Indonesian people like to socialize, love to return favors, want to be accepted by groups, and take benefit from their involvement in the group.

The results of the convergent and discriminant correlation tests show similarities with previous studies, where commitment to change is positively correlated with job satisfaction^20–22,36 and negatively with job stress.^3,37 Although in this study only ACTC has a relationship with these two variables, this is sufficient to indicate a commitment to change behavior. This is in line with Herscovitch & Meyer⁵ saying that the emergence of one form of commitment is enough to achieve a focal behavior.

This explains that when people are satisfied with their jobs they will be committed to the change because they perceive that change is important for the organization and believe in the inherent benefits. On the other hand, the more a person feels stressed, the lower the level of ACTC. This finding is in line with Kalyal et al³ who said that since ACTC has shown a willingness to support change initiatives, variables such as ambiguity or work-related stress would be negatively correlated with this particular type of commitment.

Another interesting finding from this study is that there is no correlation between CCTC and NCTC on both job satisfaction and job stress. There is no correlation between CCTC and Job Satisfaction. This study supports the study from Hinduan et al,³⁶ but is different from other studies which say that CCTC has a negative correlation to job satisfaction²² and positive to stressors.³ However, NCTC is still in line with two previous studies saying that NCTC correlates with neither Job Satisfaction²² nor Stressors,³ although there are differences in other studies saying that NCTC has a positive correlation with job satisfaction.³⁶ With these findings it can be said that if people are satisfied with their work, they will prefer to develop an attitude of commitment to change because it is based on a belief in its inherent benefits rather than because of fear of costs or just a sense of obligation in providing support for change. On the other hand, the authors also argue that no correlation occurs because the stressor is a stimulus that creates ambiguity and uncertainty for those affected so it will be difficult to decide whether they will develop CCTC, NCTC, or choose not to commit.

Thus, research conducted using Indonesian samples can add broader insights and further complement previous studies that have been conducted in several countries with eastern cultures such as Pakistan and India, and provide some evidence that commitment to change scale from Herscovitch and Meyer’s (2002) can be generalized to eastern world countries. Besides this, the findings in this study can also help management to be able to increase employee commitment in dealing with change. This study provides more insight that by providing satisfaction to employees and minimizing employee stress levels can increase employee commitment to change, especially from an affective perspective.

Study Limitations

The main limitation of this study is the limited respondent. Due to the COVID-19 pandemic, the authors were unable to collect data directly and the potential respondents were not comfortable in filling out the electronic questionnaire. The limitation of online survey might also affected the quality of the data although the authors had already taken several steps to anticipate the effects.

Conclusions

The process of adapting Commitment to change scale to the Indonesian version has been carried out in accordance with the existing steps and procedures. Based on the empirical test of the validity and reliability of the Indonesian version of commitment to change scale, the results meet the criteria. Thus commitment to change scale has satisfactory psychometric properties and can be used in Indonesian population. This scale can be compared with other international research and can be used for other studies in organizational settings.

Data collection in this study was carried out during the Covid 19 pandemic and it was not possible to collect data directly, so it was difficult to ascertain the condition of the participants in filling out the questionnaire. It is hoped that with the decline in the Covid-19 pandemic, further research can collect data directly to further ascertain the condition of the participants in filling out the questionnaire provided.