A Note on the Factorial Structure of the Maslach Burnout Inventory-General Services (MBI-GS) for Trinidad and Tobago Management Level Employees.
Prior studies investigating the factorial structure of the MBI-GS have produced fairly consistent results. Most studies confirm the basic three-factor structure across various countries and occupations with either no or minor modifications (see Schutte et al. 2000; Richardsen and Martinussen 2005; Langballe et al. 2006; Chirkowska-Smolak and Kleka 2011; Bria et al. 2014). In general, these studies support the idea that the three-factor model provides superior fit over alternative one and two-factor models. The majority of these studies, however, have been conducted in developed countries such as Germany, Italy, Norway, France and Spain, with high levels of literacy. Only a few dimensional studies of the MBI-GS have been conducted in less-developed countries, particularly in Latin America and the Caribbean. These studies appear to also support the three-factor model. For example, using a Spanish version of the MBI-GS, Fernandez-Arata et al. (2015) confirmed the three-dimensional latent MBI-GS with a Peruvian sample. In the Caribbean, using the same Spanish version of the MBI-GS in a Dominican Republic sample, Tomas et al. (2016) found support for the three-factor structure with the elimination of one item. Similarly, Oramas, Gonzalez, and Vergara (2007) supported the three-factor structure for the Spanish version in Cuba. No other studies were found investigating the factorial structure of the MBI-GS within the Caribbean, whether English- or Spanish-speaking.
The present research is conducted in Trinidad and Tobago, an island nation located in the southern part of the Caribbean, close to the northern edge of South America. While originally colonised by the Spanish, the islands came under British control in the early 19th century, and gained their independence in 1962. The population of Trinidad and Tobago is approximately 1.2 million, with a diverse economy. According to the CIA World Factbook, the official language is English with 100% of the population speaking English as the primary language; however other languages are also spoken including Spanish, French, and Caribbean Hindustani (a dialect of Hindi). The legal system operates under English common law. While Trinidad and Tobago is an English-speaking nation, it is considered a less economically developed, primarily rural country that ranks 70th in the world in GDP per capita and 138th in the world in life expectancy (CIA 2016).In addition, while the official literacy rate is above 98%, several independent studies have indicated the true Trinidad and Tobago full literacy rate may actually be closer to 50% (see Allen-Agosta 2012). Trinidad and Tobago ranks 54th on the 2009 PISA test scores.
The English version of the MBI-GS scale was administered at several major organisations in Trinidad and Tobago, with a combined sample size of one hundred-sixty-eight (168). The sampled organisations all had either a technology or professional orientation. Given the low technology nature of the commercial sector in Trinidad and Tobago, the sample size of 168, while smaller than many studies of the MBI-GS in other more developed countries, is a representative sample of the larger technical organisations and related staff within the Trinidad and Tobago population.
Methodology and Results
An exploratory factor analysis (EFA) was conducted on the full Trinidad and Tobago sample (n=168) using PASW Statistics 18, with extraction based on eigenvalues>1.0 using a Varimax rotation, and a four-factor solution was achieved. The Kaiser-Meyer-Olkin (KMO) index (0.82) and Bartlett's spherical test ([chi square]=937.1, p<0.00) supported the use of EFA. The four-factor solution, however, was not consistent with the expected 3-factor MBI-GS; one factor combined two exhaustion variables with all the cynicism variables, and another factor combined one exhaustion variable with one efficacy variable. The other two factors reflected the remaining efficacy and exhaustion variables respectively. Given the issues related to literacy and other cultural issues in less developed communities, it was then decided to run the analysis excluding respondents with office positions, such as secretaries, or field workers. Although we did not specifically measure education levels, generally these individuals would have had a high school education at best. The remaining sample (N=104) consisted of supervisors, middle, and senior managers, all with some college education, and most holding college degrees. This sub-sample is referred to as the "management" sample.
An EFA was then run on the management sub-sample, and this time the expected three-factor solution was obtained. Again the Kaiser-Meyer-Olkin (KMO) index (0.81) and Bartlett's spherical test ([chi square] = 1, p<0.00 supported the use of EFA. The resulting Scree Plot also suggested a three-factor solution. The three-factor solution indicated the expected loadings, with the exhaustion, efficacy, and cynicism variables loading primarily together on separate factors. Only two variables (one exhaustion variable and one cynicism variable) did show cross loadings (>0.40) between two factors. Total variance explained by the three-factor solution was 70.1%.
In order to test the fit of the three-factor solution for the management sub-sample we ran a confirmatory factor analysis (CFA) with the maximum likelihood procedure. We performed the test for the three-factor solution resulting from the EFA, and also examined whether a modification would result in a better fit. For the non-modified three-factor solution, the results indicated a fit problem. The goodness-of-fit index (GFI) was 0.683, the comparative fit index (CFI) was 0.737, and the root-mean-square error of approximation was 0.176, all indicating fit problems.
To obtain a better fit using the three-factor solution, modifications were made. First, four errors were allowed to correlate, which resulted in improvement in model fit. However, this full 16-item model still had problems with respect to the various fit indices. A final model was developed that reduced the MBI-GS to four variables per dimension. This eliminated the exhaustion variable of 'feeling tired when getting up in the morning and facing another day', two efficacy variables reflecting 'contribution to the organisation' and 'confidence about getting things done', and the cynicism variable reflecting 'interest in work since starting this job'. Table 1 summarises the various models and goodness of fit statistics.
While the final four-variable, three-factor model represented substantial improvement over the unmodified model, the goodness of fit measures indicated only a marginally acceptable fit. Since our management sub-sample is relatively small compared to many factorial structure studies, greater weight should probably be given to the CFI (Tabachinick and Fidell 2007; Kline 2005; Hu and Bentler 1999). The CFI is only slightly above the oftentimes acceptable 0.90 level, but below the universally recognised 0.95 level for a good fit. The RMSEA is above the commonly accepted < 0.05 level, however sometimes < 0.08 is often considered an upper acceptable limit for the RMSEA (MacCallum, Browne, and Sugawara 1996). The GFI is near 0.90, but several researchers have noted problems with the GFI as an overall measure of fit (Sharma et al 2005).
The key contribution of this research is to suggest that in less-developed countries with lower levels of literacy and different cultural contexts between classes and groups, researchers must be cautious using standardised instruments across a general population. This is true even for scales developed in the U.S. that are being used in other English-speaking countries. Examining the MBI-GS instrument with a broad sample in Trinidad and Tobago clearly resulted in a poor fit, and results were inconsistent with the hypothesised three MBI-GS dimensions of exhaustion, cynicism, and efficacy. Restricting the analysis to a higher educated supervisory and managerial level sub-sample did result in the expected three-dimension model. Given that some researchers have noted that the true literacy rate in Trinidad and Tobago may be closer to 50%, it is not surprising that some of the words used in the MBI-GS such as "enthusiastic", "exhilarated", and "accomplish" may be subject to confusion among respondents with potentially lower levels of reading ability.
Although the expected three-dimension model was achieved with the managerial sub-sample, the model fit was, at best, only marginally acceptable even with a modified four-variable per dimension model. This may point to another problem of using scales in different cultures, even when the language is the same, that is: the issues derived from different cultural contexts and variations in idiom usage. The MBI-GS, for example, uses idiomatic expressions such as 'burned out', 'face another day' and 'emotionally drained'--and while these terms are generally understood in the U.S., they may not have the same meaning in other English-speaking countries. Scale developers need to be careful about using these types of idiomatic phrases in cultural contexts other than the ones where the expressions originated.
Alarcon, G. M. 2011. "A meta-analysis of burnout with job demands, resources, and attitudes." journal of Vocational Behaviour 79: 549-62.
Allen-Agosta, L. 2012. "Paula Lucie-Smith: The Right to Read." Caribbean Beat 117 http://caribbean-beat.eom/issue-117/right-read#axzz4aHjr MiBn
Baba, V.V., B.L. Galperin, and T.R. Lituchy. 1999. "Occupational mental health: A study of work-related depression among nurses in the Caribbean." International journal of Nursing Studies 36 (2): 163-9.
Bakker, A. B., E. Demerouti, and W.B. Schaufeli. 2002. Validation of the Maslach Burnout Inventory--General Survey: An Internet study." Anxiety, Stress, and Coping: An International journal 15: 245-60.
Bria, M., F. Spanu, A. Baban, and D. Dumirascu. 2014. "Maslach Burnout Inventory--General Survey: Factorial validity and in variance among Romanian healthcare professionals." Burnout Research open access
Central Intelligence Agency (CIA). 2016. The World Factbook. Washington, DC: Central Intelligence Agency, https://www.cia.gov/library/publications/the-world-factbook/index.html
Chirkowska-Smolak, T., and P. Kleka. 2011. "The Maslach Burnout Inventory--General Survey: validation across different occupational groups in Poland." Polish Psychological Bulletin 42: 86-94.
Demerouti, E., A.B. Bakker, F. Nachreiner, and W.B. Schaufeli. 2001. << The job demands resources model of burnout." Journal of Applied Psychology 86: 499-512.
Edwards, G., B. Schyns, R. Gill, and M. Higgs. 2012. "The MLQ factor structure in a UK context." Leadership and Organisation Development Journal 33 (4): 369-82.
Fernandez-Arata, M., A. Garcia, and C. Soto. 2015. "Analisis Estructural e Invarianza de Medicion del MBI-GS en Trabajadores Peruanos." Liberabit 21 (1): 9-20.
Hu, L., and P.M. Bentler. 1999. "Cutoff criteria for fit indexes in covariance structure analysis: Conventional criteria versus new alternatives." Structural Equation Modeling 6:1-55.
Kline, R.B. 2005. Principles and Practice of Structural Equation Modelling (2nd ed.). New York: The Guilford Press.
Langballe, E., E. Falkum, S. Innstrant, and O. Aasland. 2006. "The Factorial Validity of the Maslach Burnout Inventory--General Survey in Representative Samples of Eight Different Occupational Groups." Journal of Career Assessment 14 (3): 370-84.
Leiter, M., and C. Maslach. 2009. "Nurse turnover: The mediating role of burnout." Journal of Nursing Management 17: 331-39.
MacCallum, R.C., M.W. Browne, and H.M. Sugawara. 1996. "Power analysis and determination of sample size for covariance structure modelling." Psychological Methods 1 (2): 130-49.
Oramas, A., A. Gonzales, and A. Vergara. 2007. "El desgaste profesional y factorializacion del MBI-GS." Revista Cubana de Salud y Trabajo 8 (1): 37-45.
Richardsen, A. M., and M. Martinussen. 2005. "Factorial validity and consistency of the MBI-GS across occupational groups in Norway." International Journal of Stress Management 12: 289-97.
Schaufeli, W.B., and A.B. Bakker. 2004. "Job demands, job resources, and their relationship with burnout and engagement." Journal of Organisational Behavior 25: 293-315.
Schaufeli, W. B., M. Salanova, V. Gonzalez-Roma, and A.B. Bakker. 2002. "The measurement of engagement and burnout: A two sample confirmatory factor analytic approach." Journal of Happiness Studies 3: 71-92.
Schaufeli, W.B., and T.W. Taris. 2005. "The conceptualisation and measurement of burnout: Common ground and worlds apart." Work and Stress 19: 256-62.
Schutte, N., S. Toppinen, R. Kalimo, and W. Schaufeli. 2000. "The factorial validity of the Maslach Burnout Inventory-General Survey (MBI-GS) across occupational groups and nations." Journal of Occupational and Organisational Psychology 73: 53-66.
Sharma, S., S. Mukherjee, A. Kumar, and W.R. Dillon. 2005. "A simulation study to investigate the use of cutoff values for assessing model fit in covariance structure models." Journal of Business Research 58 (1): 935-43.
Tabachnick, B.G., and L.S. Fidell. 2007. Using Multivariate Statistics (5th ed.). New York: Allyn and Bacon.
Tomas, M., S. de Los Santos, A. Alonso-Andres, and I. Fernandez. 2016. "Validation of the Maslach Burnout Inventory-General Survey on a Representative Sample of Dominican Teachers: Normative Data." Spanish Journal of Psychology https://www.ncbi.nlm.nih.gov/pubmed/27873562
Wu S., W. Zhu, Z. Wang, M. Wang, and Y. Lan. 2007. "Relationship between burnout and occupational stress among nurses in China." Journal of Advanced Nursing 59 (3): 233-39.
Table 1: Goodness of Fit for Various 3-Factor MBI-GS Models: Management Sub-Sample Model [chi square] GFI CFI RMSEA Un-modified 3-factor model 365.2 0.683 0.737 0.176 Modified (4 errors correlated) 259.4 0.766 0.833 0.144 Modified (4 variables per factor) 93.5 0.885 0.902 0.077
|Printer friendly Cite/link Email Feedback|
|Title Annotation:||NOTES and COMMENTS|
|Author:||Phillips-Hall, Cheryl Ann; Rodriguez, Carlos L.; Galbraith, Craig S.|
|Publication:||Social and Economic Studies|
|Date:||Sep 1, 2017|
|Previous Article:||A Case Study of the Influence of Garveyism among the African Diaspora.|
|Next Article:||Puerto Rico: Suffering the "Dutch Disease" in Reverse.|