Psychology and Psychiatry AFFECTIVE AND CONTINUANCE COMMITMENT IN CALL CENTRES: VALIDATION OF MEYER AND ALLEN QUESTIONNAIRE Assis. Prof. Dr. Leonor Pais 1 Dr. Cristina Souza de Castro 1 Assis. Prof. Dr. Lisete dos Santos Mendes Mónico 1 1 Faculty of Psychology and Educational Sciences, University of Coimbra, Portugal ABSTRACT The objective of this work is to contribute to validation of the Meyer and Allen Commitment Questionnaire for the call centre population in Portugal. According to the authors, organizational commitment is a psychological state, a mindset, a dependent variable that explains the link between individual and organization. Eighteen hundred employees (32.1% male; 56.9% female; M = 31.12 years, DP = 8.91) answered the Affective and Continuance scales of the tool. We carried out exploratory factor analysis with 30% of team members, randomly selected from the sample. The bi-dimensional solution was responsible for 51.1% of the total variability and reproduced the original structure proposed by the authors (item 13 was excluded due to loading in both factors). Confirmatory factor analysis was performed with the remaining 70% of the sample. According to the reference values, goodness of fit was evaluated as good /41 = 7.26; SRMR =.066, NFI =.922, CFI =.932, TLI =.909 and acceptable according to RMSEA =.071 (90% CI,.063 to.078), p <.001, sustaining the Affective (α =.807) and Continuance (α =.717) dimensions. Keywords: Call Centres, Organizations, Organizational Commitment, Affective Commitment, Continuance Commitment. INTRODUCTION Organizational commitment is a concept extensively explored in the literature for decades, and defined as a psychological state, a mindset, a dependent variable that explains the link between individual and organization [1]. In the 90s, Meyer and Allen made the first attempt to develop an integrative model for the concept, with three components, incorporating attitudinal and behavioural perspectives. Affective commitment, Normative commitment and Continuance commitment are three dimensions which indicate the desire to be part of an organization, the obligation to remain in the organisation and the need to stay due to the perceived benefits [2]. Commitment is a central concept to explain different attitudes and behaviours in the workplace, and a key to understanding how to attract and retain talents in an organization [1]. In other words, employees commitment to the company where they work influences the intention to leave, and consequently turnover rates. Call centres are new workplaces associated with particularly high turnover rates, which makes commitment especially critical for these organizations. And because service delivery occurs through human interaction, the relationships established between Customer
SGEM 2014 International Multidisciplinary Scientific Conferences on Social Sciences and Arts Service Representatives (CSRs) and customers are also influenced by employees commitment to guaranteeing high quality service [3]. Research aiming to understand the influence of organizational commitment on different attitudes, behaviours and even performance indicators in customer service operations is frequent, with contributions from several authors. Peccei and Rosenthal [4] made the first attempt to conceptualize and operationalize organizational commitment to customer services as part of a broader research to identify determinants of important aspects of service such as quality and individual performance. Authors such as Houlihan [5] and Hutchinson, Purcell and Kinnie [6] also explored the application of high commitment management practices in call centres to foster discretionary efforts and involvement among employees. The objective of the present work is to contribute with validation of the Meyer and Allen Commitment Questionnaire [2] for a call centre population in Portugal, to allow effective use of two scales of this tool in this workplace. The proposed model [2] suggests the construct is composed of three dimensions: Affective commitment, Continuance commitment and Normative commitment. The first refers to employees emotional attachment, identification and involvement with the organization. Continuance commitment is related to the need to remain in the organization due to the perceived costs associated with leaving. Normative commitment translates the obligation to stay in the organization. These dimensions are independent of each other, although the employee could experience each one of them according to antecedents of his personal and professional history. That is to say, an employee with strong Affective commitment will work enthusiastically, making any effort for the success of the organization. On the other hand, a strong Continuance commitment means the employee will work just to achieve their personal goals, because there are costs associated with leaving or no other option of work. The last dimension is related to a moral responsibility to the organization, implying that the work is correctly done because there is a moral obligation to do so. Nascimento, Lopes and Salgueiro [7] translated and tested the validity of the Meyer and Allen Commitment Questionnaire with a Portuguese population. The theoretical model initially proposed was not validated, but it was possible to confirm and validate three independent scales for Affective, Continuance and Normative Commitment. They also explain that different studies have shown relationships between these dimensions [8]. For example, a positive relationship between Affective and Normative commitment, in some cases so strong that the independency of these two scales has already been questioned. A relationship between Affective and Continuance commitment is apparently less frequent, although the authors also referred to correlations found in more than one study. Continuance and Normative commitment are weakly correlated. As there is some evidence that Affective and Normative commitment are strongly correlated, and as in call centres emotional work and precarious labour are strong features intensively discussed by several authors [9], we opted to focus only on the scales that undoubtedly could specifically measure identification and involvement with the work and the instrumental relationship with the organization. That is the reason for applying and testing only the Affective and Continuance scales. Therefore, we considered it appropriate to perform exploratory factor analysis before the confirmatory
Psychology and Psychiatry one, to analyse how the items would load in the two emergent factors. The bidimensional solution achieved was confirmed in Confirmatory Factor Analysis. MATERIALS AND METHODS Sample The sample is composed of 1814 employees (CSRs) working in a call centre company in Portugal. 582 are male (32.1%, M age =31.7 years, SD=7.60) and 1033 female (56.9%, M age =31.12 years, SD=8.91), with 199 (11.0%) missing values considering gender. These CSRs were distributed in 184 teams, belonging to different call centre operations. In a large call centre provider, each operation represents a type of service outsourced from another company (a client organization ), and tailored to serve this company s customers according to its products, process and procedures. Measure The Meyer and Allen Commitment Questionnaire [2] is originally composed of three scales, to measure the Affective, Normative and Continuance dimensions of the construct. We used the scales validated for the Portuguese population by Nascimento, Lopes and Salgueiro [7], although applying only the Affective and Continuance ones. As already mentioned, this decision was motivated by certain features of this workplace, such as the strongly emotional connotation of this type of activity being also frequently associated with precarious labour [9]. The Affective scale has 6 items (three of them reversed) and the Continuance, 7 items. The questions were all targeted to the Operation to which the employee belongs. Questions were presented to the respondents with a Likert scale where 1 = totally disagree and 7 = totally agree. The inverted items are 1, 3 and 6. Procedures and data analysis The Commitment Questionnaire was made available to all CSRs through online survey software, Questionpro, which allowed rapid and efficient distribution. The responses were anonymous. Exploratory factor analysis and descriptive statistics were performed using the Statistical Package for Social Sciences (SPSS, version 21.0; SPSS Inc., Chicago, IL). We eliminated 23 outliers. Confirmatory factorial analysis was executed with AMOS. Reliability was calculated by Cronbach's alpha [10]. The normality of the variables was assessed by the coefficients of skewness (Sk) and kurtosis (Ku) uni and multivariate, showing that no variable presented values violating normal distribution, Sk <3 and Ku <10. RESULTS Exploratory factor analysis We carried out exploratory factor analysis (EFA) with 30% of team members (n = 537), randomly selected from the sample, with factors extracted using the Principal Component method followed by Varimax rotation with Kaiser normalization. Factors
SGEM 2014 International Multidisciplinary Scientific Conferences on Social Sciences and Arts with an eigenvalue above 1 were retained, according to the Scree Plot, serving the percentage of extracted variance. EFA validity was evaluated through the Kaiser- Meyer-Olkin (KMO) measure and also applying Bartlett s test of sphericity. Results indicate a good factorability of the correlation matrix: the KMO was higher than.70 (KMO =.819) and the Bartlett s test of sphericity showed a X 2 (78) = 2300.6, p <.001. The bi-dimensional solution obtained from PCA was responsible for 51.14% of the total variability, and reproduced the original structure proposed by Meyer and Allen [2]. Factor 1 explained 27.7%, and factor 2, 23.4% of the total variance. Item 13 (since I have already put so much of myself into this operation I do not consider working elsewhere) presented factorial loadings higher than.30 in both factors, which does not contribute to their orthogonality. For this reason, the item was excluded. Table 1 presents the factorial loadings (s), communalities (h 2 ) and descriptive statistics of each of the 12 retained items. All factorial loadings are higher than.50, loading only in one dimension. Communalities are higher than.29. Items grouped in the first dimension correspond to Affective commitment; items loading in the second dimension correspond to Continuance commitment. The Cronbach s Alpha of the instrument was α = 0.783 for the global scale; α =.814 for Affective dimension; and α =.765 for Continuance dimension, which shows acceptable reliability [10]. Table 1: Principal Component Analysis of Affective and Continuance dimensions of the Commitment Questionnaire: Mean (M), standard deviations (SD), factorial loadings (s) and communalities (h2) of the rotated component matrix Items 2-This operation has a great deal of personal meaning for me. Affective (s) Dimensions Continuanc e(s) h 2 M SD.797.052.637 4.49 1.74 1(r)- I do not feel emotionally attached to this operation..784 -.032.616 4.98 1.90 6(r)- I do not feel a strong sense of belonging to my operation..724 -.069.529 5.78 1.54 5- I would be very happy to spend the rest of my career with this operation..706.155.523 3.54 1.94 4- I really feel as if this operation s problems are my own..668.088.454 4.19 1.79 3(r)- I do not feel like part of the family at my operation..610 -.061.376 5.50 1.68 8- It would be very hard for me to leave my operation right now even if I wanted to..106.777.615 4.44 1.86 12- Too much in my life would be disrupted if I decided I wanted to leave my operation now..083.767.596 4.93 1.88 11- One of the few serious consequences of leaving this operation would he the scarcity of available alternatives. -.264.715.581 4.87 1.81 9- One of the major reasons I continue to work for this operation is that leaving would require considerable personal sacrifice, because another operation could not.056.709.506 4.07 1.81 cover the benefits I have here. 7- I feel that I have too few options to consider leaving this operation. -.054.542.296 3.92 1.72 10- Right now staying with my operation is a matter of necessity as much as desire..378.515.408 4.78 1.72 Eigenvalues 3.44 2.70 % of Variance 27.73 23.41 (r): reversed items
Psychology and Psychiatry Confirmatory factor analysis Confirmatory factor analysis was performed with the remaining 70% of the sample (n = 1254). Goodness of fit was analyzed by the indexes of NFI (Normed fit index; good fit>.80) [11], SRMR (Standardized Root Mean Square Residual; appropriate fit<.08) [12], TLI (Tucker-Lewis Index; appropriate fit>.90) [12], CFI (Comparative fit index; good fit>.90) [11], RMSEA (Root Mean Square Error of Approximation; good fit<.05, acceptable fit<.08) [14] [11], and X 2 (p>.05, but irrelevant if N>500) [13] [11]. Figure 1 shows the diagram of the proposed model, and in Table 2 we can see the goodness of fit indexes of this model (Model 1). Figure 1: Proposed model of Affective and Continuance dimensions of the Commitment Questionnaire. The goodness of fit indexes of the proposed model (Model 1) are shown in Table 1. Model 1was not adjusted. Observation of the results of the first analysis showed that item 11 has high covariance between its error term and the Affective latent construct, indicated by the modification indices [15] (MI = 148.4), and also high covariance with item 12 (MI = 90.39). We decided to exclude it in Model 2 (see goodness of fit indexes for Model 2 in Table 2). Despite the acceptable fit values for NFI and SRMR [12], we found that the fit of this model can be improved by correlation of the residual variability between variables 3 and 6 (MI = 204.3), and 1 and 3 (MI = 71.7), which was done in Model 3 (see Table 2). This procedure increased the fit of the model of Affective and Continuance dimensions of the Commitment Questionnaire (see goodness of fit indices for Model 3). Table 2: Goodness of fit indexes obtained in factorial validation of the Affective and Continuance dimensions of the Commitment Questionnaire Model NFI SRMR TLI CFI χ 2 /df RMSEA RMSEA 90% Confidence Interval 1.788.096.747.797 17.82*** (df=53).116.109 -.122*** 2.841.078.808.850 14.18*** (df=43).103.095 -.110***
SGEM 2014 International Multidisciplinary Scientific Conferences on Social Sciences and Arts 3.922.065.909.932 7.26*** (df=41).071.063 -.078** ** p =.009 *** p <.001 According to the reference values of Model 3, we conclude that goodness of fit was evaluated as good, sustaining the Affective and Continuance dimensions of the Commitment Questionnaire. Table 3 presents the estimates, standard errors, critical ratios, and standardized regression weights for the confirmatory structural analysis of Model 3 (see diagram of the estimated model in Figure 2). All estimated parameters are statistically significant (see critical ratios), and items have mostly high standardized regression weights, ranging from.35 to.82. The correlation between the two factors is moderate. In Table 3 we also show the descriptive statistics for items, revealing that the highest score is for item 6, and the lowest for item 5. Table 3: Estimates, Standard-errors (SE), Critical ratios (CR), and Standardized regression weights (SRW): Confirmatory structural analysis of Model 3, Means (M) and standard deviations (SD) Factors Items Estimate SE CR SRW M SD Affective Continuance *** p <.001 Commit1 1.000.612 5.08 1.83 Commit2 1.216.061 20.098 ***.815 4.57 1.67 Commit3 0.647.045 14.405 ***.430 5.45 1.72 Commit4 1.019.057 17.783 ***.648 4.09 1.76 Commit5 1.196.064 18.609 ***.693 3.50 1.93 Commit6 0.728.050 14.665 ***.504 5.67 1.62 Commit7 1.000.346 3.88 1.75 Commit8 2.307.219 10.522 ***.752 4.38 1.85 Commit9 2.014.195 10.339 ***.669 3.98 1.82 Commit10 1.473.152 9.683 ***.529 4.76 1.68 Commit12 1.807.179 10.092 ***.605 4.96 1.80 Figure 2: Estimated model of Affective and Continuance dimensions of the Commitment Questionnaire in call centres
Psychology and Psychiatry The descriptive statistics and reliability coefficients of the Affective and Continuance dimensions of the Commitment Questionnaire are shown in Table 4. The Cronbach s alpha values are classified as good or acceptable [10]. The Affective dimension of the construct was statistically higher than the Continuance one, t (1253) = 7.70, p<.001. Table 4: Descriptive statistics and reliability coefficients (Cronbach s alpha) of Affective and Continuance dimensions of the Commitment Questionnaire N Cronbach s α Minimum Maximum M SD Affective 1254.807 1.00 7.00 4.73 1.25 Continuance 1254.717 1.00 7.00 4.39 1.22 Commitment (global scale) 1254.767 1.50 7.00 4.57 0.97 DISCUSSION The statistical procedures applied resulted in a contribution to validation of the Affective and Continuance scales of the Meyer and Allen Commitment Questionnaire for a population of CSRs working in a large call centre company in Portugal. Both scales will allow us to measure effectively the Affective and Continuance dimensions of the construct that represents the link between an individual and the organization, a psychological state which influences behaviours, attitudes and even performance. Affective commitment refers to employees emotional attachment, identification and involvement with the organization; Continuance commitment is related to the need to remain in the organization due to the costs associated with leaving [2]. A moderate correlation between the two factors was found, but a similar result had also been reported in previous studies, as mentioned in Nascimento, Lopes and Salgueiro [8], without compromising the independence of the factors. Considering the goodness of fit of the model, this was obtained in Model 2, and enhanced in Model 3. Overall, we consider the fit was good, but it was only obtained after excluding one item and for all indices due to correlation of the residual variability of a set of items as indicated in Figure 2. This means there is a percentage of variability that is not explained by the model in our sample, and this is exactly what should be studied in future research in the same population, with its very particular features. REFERENCES [1] Cardoso, L., Castro, C.S., Gomes, D. Organizações, comprometimento e identificação: semelhanças e diferenças entre modelos e uma perspectiva de integração. In Gomes, A. D. Psicologia das organizações, do trabalho e dos recursos humanos. Coimbra: Imprensa da Universidade de Coimbra, pp 353-375, 2011. [2] Meyer, J. P., & Allen, N. J. Commitment in the workplace: Theory, research, and application. Thousand Oaks, CA: Sage Publications, 1997. [3] Malhotra, N., Budhwar, P., & Prowse, P. Linking rewards to commitment: an empirical investigation of four UK call centres. The International Journal of Human Resource Management, vol 18/12, pp. 2095-2128, 2007.
SGEM 2014 International Multidisciplinary Scientific Conferences on Social Sciences and Arts [4] Peccei, R., & Rosenthal, P. The antecedents of employee commitment to customer service: evidence from a UK. International Journal of Human Resource Management, vol 8/issue 1, pp 66-86, 1997. [5] Houlihan, M. Tensions and variations in call centre management strategies. In Human Resource Management Journal, vol 12/issue 4, pp 67-85, 2002. [6] Hutchinson, S., Purcell, J., & Kinnie, N. Evolving high commitment management and the experience of the RAC call centre. In Human Resource Management Journal, vol 10/issue 1, pp 63-78, 2000. [7] Nascimento, J. L., Lopes, A., & Salgueiro, M. D. F. Estudo sobre a validação do Modelo de Comportamento Organizacional de Meyer e Allen para o contexto português. Comportamento organizacional e gestão, vol 14/issue 1, pp 115-133, 2008. [8] Meyer, J. P., Stanley, D. J., Herscovich, L., & Topolnytsky, L. Affective, continuance, and normative commitment to the organization: A meta-analysis of antecedents, correlates, and consequences. Journal of Vocational Behavior, vol 61, pp 20-52, 2002. [9] Matos, P.R.M.A. Precarious Labour in Portuguese Call Centres: An Anthropological Study, Doctoral thesis Goldsmiths, University of London. [Thesis]: Goldsmiths Research Online, 2010. [10] Nunnaly, J.C. Psychometric theory. New York: McGraw-Hill. 1978 [11] Schumacker, R.E., & Lomax, R.G. A beginner s guide to structural equation modeling. Mahwah, NJ: Lawrence Erlbaum Associates, 1996. [12] Brown, T. Confirmatory factor analysis for applied research. New York: The Guilford Press, 2006. [13] Bentler, P. Quantitative methods in psychology: Comparative fit indexes in structural models. Psychological Bulletin, vol. 107,pp 238-246, 1990. [14] Kline, R.B. Principles and practice of structural equation modeling (3rd ed.). New York: The Guilford Press, 2011 [15] Bollen, K. A. Structural equations with latent variables, New York: Wiley, 1989.