97 Key Words: Local public services; citizen’s satisfaction; quality perception; SERVQUAL; PLS. Public sector organizations –and specifically those ones operating at local government level– are not immune to pressure in order to improve citizen’s service on a continuous basis. Whereas some of these pressures arise from own local authorities (from a genuine desire by local public responsibles to improve the quality of those services provided to CITIZENS’ POSITION IN SERVICE QUALITY PERCEPTION. AN APPROACH ANALYSIS IN THE CASE OF SPANISH LOCAL ADMINISTRATIONS José L. VÁZQUEZ Prof. Dr, Faculty of Economic and Entrepreneurial Sciences, University of León (Spain)1 Pablo GUTIÉRREZ Assistant Professor, Faculty of Economic and Entrepreneurial Sciences, University of León (Spain)2 María P. GARCÍA Researcher Scholar, Faculty of Labour Sciences, University of León (Spain)3 Public sector organizations are not immune to pressure in order to improve their services to citizens. At this point, a major problem is that of “citizen-customer” satisfaction surveys being prone to focus on individuals’ perceptions on service delivery: what the individual think about the quality of delivered service. Authors have usually considered tools like the SERVQUAL scale (useful at measuring service quality in a diverse number of organizations and situations, both in private and public sector spheres) as a prior reference when researching in the quality field, but as being the “source”, then changed (modified) into a more or less “new” proposal. Following this research guideline, in this paper a new model for measuring perceived quality level in local government activities is presented. Three dimensions are suggested for citizens to valuate service quality: technical, functional and overall features. Transylvanian Review of Administrative Sciences, 15 E/2005, pp. 97-106 1 Facultad de CC. Económicas y Empresariales. Campus de Vegazana s/n 24071-León (Spain). Tel. (+34) 987291751. 2 Facultad de CC. Económicas y Empresariales. Campus de Vegazana s/n 24071-León (Spain). Tel. (+34) 987291752. 3 Facultad de Ciencias del Trabajo. Campus de Vegazana s/n 24071-León (Spain). Tel. (+34) 987295245. 98 citizens), others come “forced” either through public initiatives or through an increase in consumer activism. At this point, a major problem is that of “citizen-customer” satisfaction surveys – an increasing and expensive phenomenon in the public sector – being prone to focus on individuals’ perceptions on service delivery: what the individual thinks about the quality of delivered service. A well-known potential tool at this purpose is the SERVQUAL model originally developed by Parasuraman, Zeithaml and Berry and subsequently refined as a general methodology for measuring service quality in a diverse number of organizations and situations, both in private and public sector spheres. The model identifies specific criteria for citizens to valuate service quality. The model also invites subjects to allocate weights to each one of five identified dimensions regarding service quality, expected to reflect their relative importance from the individual’s perspective. Adding weights according to size of gaps identified in the other sections of questionnaire allows an assessment on the “focus” of the organization. Authors have ususally considered this scale as a prior reference when researching in the quality field, but as being the “source”, then changed (modified) into a more or less “new” proposal. Following this research guideline, in this paper a new model for measuring perceived quality level in local government activities is presented. Three dimensions are suggested for citizens to valuate service quality: technical, functional and overall features. 1. Introduction: Public sector organizations, and so –and perhaps in particular– those operating at local level, are not inmune to pressure in order to improve customer service on a continuous rhythm and basis. Whereas some of these pressures arise from own local authorities (that is, from a genuine desire by local public responsibles to improve the quality of those services provided to citizens by their institutions), others come “imposed” or “forced” either by means of public initiatives or through an increase in consumer activism. In the public sector context, it is recognized that elected authorities must face more troubles and difficulties –thus coming into a need of making a bigger effort– when trying to improve citizens’ service with respect to those other managers working in the private sector. Even when private-service quality initiatives might have a very little deal to public/social services provision, they are mostly intended to enhance the public organization’s bottom line – its long term profitability–; then justifying all required investments (not only money, but also in terms of working hours, changing minds and rutines...) and getting funds on the basis of the desired purpose of finally getting a financial pay-back. However, real trouble comes when forgeting that in case of private sector organizations definition of target customers’ groups is usually to be quite an easy task (at least from a priori theoretical point of view), either real (existing) and/or potential. These expected consumers are those individuals determined to pay the prevailing market price to obtain the property or the right to use the concrete good or service. Conversely, in a public context, financial costs of such customer-service initiatives often have to be supported through budget reallocation taking funds away from other activities, most public sector organizations having quite a number of diverse potential “customers” asking for their services. Only in a few cases –once more in relative terms– the situation is similar to that above of private sector, “customer-citizens” paying directly or indirectly for provided and received goods and/or services. In some other cases, individuals are to be strictly public-services recipients or users –sometimes unwillingly–, but make little or not at all financial contribution towards their provision (even consciously, i.e. free- riding). Even more, it could be also possible to find a concrete subject paying for a particular service from public administration, but not experiencing its benefits by direct or indirect use. 99 At this point, the real fact is that the concrete role of possible public stakeholders and/or elected representatives could be clearly ambiguous. Summarily, a new concept on public marketing could and must be attempted to correctly categorize citizens (1). 2. Some brief comments on quality, satisfaction and reputation concepts: Even when (perceived) quality and (perceived) satisfaction are both terms widely used in the research literature on services, a clear distinction between the two concepts remains a challenge even nowadays (2). Indeed, most researchers agree that they are two different constructs. Moreover, it uses to be very easy to appreciate differences between these concepts in case of tangible products (goods). Meanwhile if considering services, definition on quality uses to be stated a posteriori as the perceived difference between expected and obtained service (and thus e.g. a service being perceived as achieving a high quality standard by the only reason of prior expectations on it were exceeded). On the other hand, customer satisfaction or dissatisfaction are both well-known and established concepts not only in marketing literature but also in several other disciplines. When related to consumer research, customer satisfaction has been used in order to describe differences between concrete alternatives and/or brands (3). Authors have also referred to this concept as a common denominator to describe differences between product groups and industries (4). Clearly, an individual’s satisfaction cannot be directly assessed by using an objective measure tool (5). However, and even when being considered as an abstract and theoretical phenomenon, it can be measured as a weighted average of multiple indicators (6). Measurement errors in the corresponding obtained index depend on quality and quantity of used variables (7). Doing things in this way becomes a common denominator when trying to make it possible comparisons between industries, companies and individuals. Customer satisfaction is also the accumulated experience of an individual’s purchase and consumption experiences. It is mainly influenced by two factors: expectations and experienced service performance (5). Perceived performance is influenced by subjects’ perception on service quality, marketing-mix, brand name and image of the company/organization. As far as satisfied customers tend to maintain their consumption pattern or increase demand of a same good or service, not only customer orientation, but especially customer satisfaction has become an important indicator about product quality and future revenue. At this point, Fornell (8) claims that individuals’ satisfaction clear and undoubtely influences purchasing behaviour: satisfied customers tend to be loyal customers, but loyal customers are not necessarily satisfied. According to Yi (3), a customer’s satisfaction is conceived as a function on his/her perceived (service) quality and expectations. We propose that not only satisfaction, but also a second concept, that of reputation, are both influenced by quality. In marketing literature, major attention has been focused in the concept of brand (see e.g. 9, 10, 11). Motivations for studying the importance of brands come from financial and strategic reasons. Inside this context, brand awareness, image and reputation regarding a concrete product or supplier influences buyers’ purchasing decision (i.e. a good brand or reputation stimulates purchase by simplifying decision effort and rules). Thus reputation and/or brand become an issue of attitudes and beliefs regarding either to brand awareness and image (12), and so increasing customer satisfaction and loyalty (8). Keller (13) suggests that brand awareness “…relates to the likelihood that a brand name will come to mind and the eases with which it does”, whereas brand image is considered as “… [those] perceptions about a brand as reflected by the brand associations held in consumer memory”. Attitudes and beliefs are influenced by previous experience: individuals with a previous experience background will base their attitudes and beliefs on an experienced good and/or service quality; meanwhile, subjects with little or no experience may base their attitudes and beliefs on reputation. Moreover, even nations, regions, governments etc. enjoy a reputation. 100 As a direct consequence, reputation may be aggregated to a macro level extent by using the concept of country of origin. Citizens living in a concrete region (usually) have previous experience on services offered by governments in such a region; thus –and among other factors–they base their valuation of satisfaction concerning government acts on perceived service quality. On the other hand, residents outside the concrete territory may (usually) have no experience with the policy offered by governments in that region; thus, their decision must be at least partly based on the country’s or region’s reputation and advice from people who may have a previous expertise about. Based on their own and previous research on the country of origin topic, Papadopoulos et al. (14) concluded that there is enough evidence to posit that: (i) a “country of origin” effect does exists; (ii) both final consumers and industrial buyers’ groups are affected by “made-in” images; and (iii) “made-in” stereotypes can be changed. Thus, reputation becomes an important issue both from a marketing and a strategic point of view. We will propose that reputation is an important factor influencing citizens and companies’ satisfaction with local government’s policies. 3. Quality in public services: Increasing –and expensive– “citizen-customer” satisfaction surveys in public sector use to be prone to focus on every individuals’ perceptions on service delivery, i.e. what every individual thinks about the quality of delivered service. At this purpose, a well-known potential tool is the SERVQUAL. This model originally developed by Parasuraman, Zeithaml and Berry (15) has been subsequently refined as a general methodology for measuring service quality in a diverse number of organizations and situations, both in private and public sector spheres (e.g. 16 by own authors; 17 and 18 for local governments). This model identifies specific criteria for citizens to valuate service quality. The SERVQUAL criteria are classified in five major dimensions: 1. Tangibles: regarding the appearance of physical facilities, equipment, personnel, and communications materials. 2. Reliability: the ability to perform the promised service dependably and accurately. 3. Responsiveness: the willingness to help customers and provide them a prompt service. 4. Assurance: the competence of the system and its credibility in providing a courteous and secure service. 5. Empathy: the approachability, ease of access and effort taken to understand customers’ needs. As stated above, authors have used to use and/or refine the SERVQUAL scale when researching on quality issues, but it is also usual for them to go further, even changing or turning it into mostly really “new” ones. In this paper we have developed a new model for valuating quality in local government services, considering three different dimensions for criteria: 1. Technical features: objetive items for quality valuation. 2. Functional features: subjetive items for quality valuation. 3. Overall features: general items about different features in local governments. 4. Research methodology and results: Partial Least Squares (PLS), a Structural Equation Modeling (SEM) tool, was used to make analyses in this research. SEM enables researchers to simultaneously examine the structural component (path model) and measurement component (factor model) in one same model (19). The use of PLS has advantages over other SEM tools, e.g. LISREL, as far as PLS can be applied to explore the underlying theoretical model. Furthermore, PLS can be used when working with relatively smaller sample sizes because it does not require restrictive distributional assumptions about the underlying data (19). 101 McArdle (20) takes this distinction to make several statements contrasting components analysis and factor analysis. Chin (21) suggests these statements should also be considered for PLS and LISREL: – PLS is a tool that tries primarily to estimate the variance of endogenous constructs and in turn their respective manifest variables (if reflective): the focus should be shifted from only assessing the significance of parameter estimates (i.e. loadings and structural paths) to that of predictive validity. – LISREL is superior to PLS on mathematical grounds: this point refers to the fact that LISREL is a population based model for estimating loadings and structural path estimates. Only under the joint condition of large sample size and large number of indicators per factor will the estimate of the factor loadings and structural path estimates approximate that of the LISREL estimate. Otherwise, the loadings in a PLS analysis tend to be overestimated and the structural paths, conversely, underestimated (22, 23). An examination of the component versus common factor distinctions will also suggest that communality represent yet another factor. Thus, superiority of LISREL over PLS refers to the ability to estimate the underlying population parameters. As noted in statement one, this becomes less of a concern if the objective is to account for multivariate variance in a predictive sense. – LISREL is superior to PLS on statistical grounds: this statement is relatively contentious and depends on the perspective of the researcher (the reverse statement, suggesting that PLS has better statistical sampling properties than LISREL, could equally be made). Yet, due to the nature of the PLS algorithm, the construct score estimates are biased and only consistent under the condition of high communality, appropriate number of indicators per construct, and increasing sample size. Nonetheless, because PLS is a limited information estimation procedure, an appropriate sample size tends to be much smaller than that needed for a full information procedure such as LISREL. Even with distributional violation, the Maximum Likelihood estimation procedure for LISREL can be quite robust and may possibly, as mentioned above, produce better estimates of the population parameters. The generalizability issue for multiple group comparisons needs also be considered. LISREL provides a statistical basis using a chi-square test for multiple group comparison. The generalizability of PLS scores for group comparisons has to be determined yet. Predictive relevance, on the other hand, is a different issue that should be further explored. For example, the use of a mean loss function on the holdout data in a sample reuse procedure can be a viable means for choice selection of indicators. – Finally, PLS is superior LISREL on practical grounds: PLS is computationally more efficient than LISREL in the same way that a components analysis is faster than a Maximum Likelihood factor analysis. In summary, an understanding of the issues related to choice of component versus common factor analysis can provide a basis for choosing PLS or LISREL as an analysis technique. We think clearly articulated that the aim of LISREL is to estimate causal model parameters whereas PLS is to maximize variance explained. For further understood on this issue, one only needs to have a look on the component/factor analytic distinctions. In our research the questionnaires for the survey were designed to measure five distinct latent constructs (see Table 1): objective features (ξ 1 , subjective features (ξ 2 , overall features (ξ 3 , perceived quality (ξ 4  and reputation (ξ 5 . At designing the model, there was another element –or explicit variable– that was to be related with these latent variables: citizen satisfaction (ξ 6 . 102 Table 1: Measurement instruments of the quality of services provided by municipalities. Latent variables: Explicit variables: Objective (technical) features (ξ 1 ) a) Office setting in council. b) The forms of the Town Hall are foolproof. c) Telephone (fax, internet) is a good communication with the Town Hall. d) Notifications are interesting for someone. e) Notifications use an easy language with citizen. f) Employees are always prepared for helping citizens. g) Locals are in a good condition. Subjective (functional) features (ξ 2 ) a) If they promise something, they will carry out it in time. b) I can understand how things are running at my Town Hall. c) I know who received my complaints. d) Employees are trained to do their job perfectly. e) Employees inform about services. f) Employees are kind and polite. g) Employees take into account my personal situation. Overall features (ξ 3) a) Services have changed because our ideas/suggestions. b) Our complaints are effectively processed. c) We trust in our Town Hall employees. d) Town Hall employees have resources enough for developing services. Perceived quality (ξ 4 ) a) Global quality is good. b) Quality in this Town Hall is better than in any other one. Reputation (ξ 5 ) a) Public services must be developed for municipalities. b) Our Town Hall reputation is very good. The Local Government Citizen Reputation Index (LGCRI) is an economic indicator developed in order to measure customer satisfaction with local public services and reputation of local administration. In our model, six interrelated variables –five latent and one explicit variable– were introduced. This was done based on well-established theories and approaches in customer behaviour, the result being applicable at a number of different public governments. The intended LGCRI model is depicted in Figure 1. It can be seen that a set of explicit variables is associated to each one of the latent variables. The model as a whole is important at determining the main goal variable. The structural model contains what follows: • A number i of exogenous latent constructs, represented as ξ i (or directly as xi i , depending on used dictionary). • A number j of endogenous latent constructs, represented as η j (eta). • Paths connecting ξ i to η j , statistically represented as γ ji (gamma) coefficients. • Paths connecting one concrete η m to another η n , represented as β nm (beta). • Shared correlation matrix between one concrete ξ h and other ξ k , represented as Ø kh (phi) • A shared correlation matrix among the error terms of the η j , called ψ (psi). • The error terms themselves, represented as ζ j (zeta). On the other hand, in Figure 1 there is also a measurement model, composed of: • A number z of X z variables and a number c of Y c variables, which represent observations or the actual data collected. X z and Y c are the measures of the exogenous and endogenous constructs, respectively. Each X z should load into one ξ i , meanwhile each Y c should load into one η j . 103 • Paths connecting any observed variable X z and its correspondant ξ i (i.e. the item loading on its latent variable), represented as λx z (lambda-x). • Θδ (theta-delta) coefficients, representing the error variance associated with every Xz variable (i.e. the variance not reflecting its latent variable ξ i ), showed as δ z . • Paths connecting any observed variable Y c and its correspondant η j (i.e. the item loading on its latent variable), represented as λy z (lambda-y). • Θε (theta-epsilon) coefficients, representing the error variance associated with every Yc variable (i.e. the variance not reflecting its latent variable η j ), showed as ε c . Both these two models (structural and measurement) can be summarized in a new more simplified figure, showing those relationships found to be most important in the model (Figure 2). The field research providing data for this study was done in 2004 and early 2005, finally obtaining up to 400 required questionnaires for the sample to be representative (at 95.5 % level, e = ± 5.0 %) according to the size of total population (number of town council demarcations). These questionnaires were directly fulfilled at 76 different town councils in Castilla y León, a region in the North of Spain. The 76 different locations were ramdomly selected in every one of nine provinces (Spanish official intermediate geographical demarcation between those of municipality and region), with the only restriction of taking into consideration the weight of each province in the region, according to its population level (citizens older than 16), as it can be seen in Table 2. The causality model of figure 3 summarizes the various structural regressions of our model. The path coefficients are the standardized regressions coefficients. The R2`s are also shown. The X 2 X 3 X 4 X 5 X 6 X 7 X 8 X 9 X 10 X 11 X 12 X 13 X 14 X 15 X 17 X 18 Y 1 Y 2 Y 4 Y 5 Y 3 X 1 δ9 δ10 δ2 δ3 δ4 δ5 δ6 δ7 δ1 δ8 δ11 δ12 δ13 δ14 δ15 δ16 δ17 δ18 ξ1 ξ2 ξ3 η1 η3 η2 ε1 ε2 ε3 ε4 ε5 ζ2 ζ3ζ1 γ12 γ11 γ13 β31 β21 β32 φ23 φ13 φ12 λy1 λy 2 λy 4 λy 5 λy 3 λx1 λx 2 λx 3 λx 4 λx 5 λx 6 λx 7 λx 8 λx 9 λx 10 λx 11 λx 12 Λx 13 λx 14 λx 16 λx 17 λx 18 Figure 1: Network including constructs and measures for LGCRI model. 104 Objetive Features (ξ )1 Subjetive Features (ξ )2 General Features (ξ )3 Perceived Quality (ξ )4 Satisfaction (ξ )6≈Y Reputation (ξ )5 significance levels shown next to the path coefficients in parentheses are coming from PLS-VB. All relations are significant values. Table 2: Universe and sample distributions (%) by province in the Castilla y León region. Province: Citizens older than 16: Universe (%): Sample (%): Avila 84,249 9.83 % 10.00 % Burgos 101,344 11.83 % 11.50 % León 183,622 21.43 % 24.00 % Palencia 66,613 7.78 % 7.75 % Salamanca 121,676 14.20 % 14.00 % Segovia 64,649 7.55 % 7.25 % Soria 40,868 4.77 % 4.50 % Valladolid 95,747 11.18 % 11.25 % Zamora 97,916 11.43 % 9.75 % Total: 856,684 100.00 % 100.00 % Once data were analyzed, it was obtained that those variables in the intended model identified as objective features, subjective features and overall features were to have a significant impact on citizen’s perceived quality of local public services (0.369, 0.373 and 0.346, respectively, as shown in Figure 3). Perceived quality direct impact on reputation was not so important (0.232), meanwhile being a very important factor at influencing (explaining) citizen satisfaction (0.789). Finally, quite a good relationship was found between citizen satisfaction and reputation (0.476). Thus, it can be understood that perceived quality acts as an important indirect effect (through satisfaction) on reputation. Obtained R2 values were 0.417 for reputation, and 0.570 for satisfaction, respectively. These values can be considered to be very satisfactory, taken into account the complexity of the model. Figure 2: Causality model describing causes and consequences. 105 Objetive Features (ξ )1 0.369 (0.000) 0.373 (0.000) 0.346 (0.000) 0.232 (0.000) 0.789 (0.000) 0.476 (0.000) R =0.450 2 R =0.417 2 R =0.570 2 Subjetive Features (ξ )2 General Features (ξ )3 Perceived Quality (ξ )4 Satisfaction (ξ )6≈Y Reputation (ξ )5 Figure 3: Final causality model. 5. Conclusion: Main intended objective was to ascertain whether or not we can suggest a model wherein perceived quality, satisfaction and reputation of local public services (at the level of municipalities) would be interacting and could be considered together by public authorities. Obtained results points clearly at this possibility. What’s more, it has been found a strong and significant relationship between these three constructs. Thus, if those elements and factors determining perceived quality are managed in the right way, either direct and indirect results could be obtained on reputation. A first finding at this point is the fact that direct relationship between perceived quality and reputation is far not so powerful as the indirect influence on it through satisfaction. However, we musn’t forget that perceived quality is understood as determined by three constructs including 23 items/variables acting as “basic” instruments at the end when trying to get or explain effects on satisfaction and reputation. 6. References: 1. Vázquez, J.L., “Past, present and future of public and social dimensions in Marketing conceptual development”, International Review of Public and Non Profit Marketing, vol. 1, nr. 1, 2004, pp. 9-34. 2. Hurley, R.F. and F. Estalami, “Alternative indexes for monitoring customer perceptions of service quality: a comparative evaluation in retail context”, Journal of the Academy of Marketing Science, vol. 26, nr. 3, 1998, pp. 209-21. 3. Yi, Y., “A critical review of customer satisfaction”, in Zeithaml, V.A. (ed.), Review of Marketing, American Marketing Association, Chicago (IL), 1989, pp. 68-123. 4. Meeks, J.G., “Utility in economics: a survey of the literature”, in Turner, C.F. and E. Martin (eds.), Surveying Subjective Phenomena, vol. 2, Russel Sage Foundation, New York (NY), 1984, pp. 41-91. 5. Simon, J.L., “Interpersonal Welfare Comparison Can Be Made and Used for Redistribution Decisions”, Kyklos, 1974, nr. 27, pp. 68-98. 6. Johnson, M.D. and C. Fornell, “A framework for comparing customer satisfaction across individuals and product categories”, Journal of Economic Psychology, vol. 12, 1991, pp. 267-286. 7. Fornell, C. and B. Wernerfelt, “Defensive marketing strategy by customer complaint management: a theoretical analysis”, Journal of Marketing Research, vol. 24, nr. 4, 1987, pp. 337-346. 8. Fornell, C., “A national customer satisfaction barometer: the Swedish experience”, Journal of Marketing, vol. 56, nr. 1, 1992, pp. 6-21. 9. Aaker, D., Brand equity, The Free Press, New York (NY), 1991. 10. Leutheser, L., “Defining, measuring and managing brand equity: a conference summary”, Report Nr. 88-104, Marketing Science Institute, Cambridge (MA), 1991. 106 11. Aaker, D. and A. Biel (eds.), Building strong brands, Lawrence Erlbaum Ass., Hillsdale (NJ), 1992. 12. Maltz, E., “Managing brand equity: a conference summary”, Report Nr. 99-110, Marketing Science Institute, Cambridge (MA), 1991. 13. Keller, K.L., “Conceptualizing, measuring and managing customer-based equity”, Journal of Marketing, vol. 57, nr. 1, 1993, pp. 1-22. 14. Papadopoulos, N.G., L.A. Heslop, F. Graby and G. Avlontis, “Does ‘country-of-origin’ matter? Some findings from a cross-cultural study of consumer views about foreign products”, Report Nr. 87-104, Marketing Science Institute, Cambridge (MA), 1987. 15. Parasuraman, A., V.A. Zeithaml and L.L. Berry, “A conceptual model of service quality and its implications for future research”, Journal of Marketing, vol. 49, nr. 4, 1985, pp. 41-50. 16. Parasuraman, A., V.A. Zeithaml and L.L. Berry, “Alternative scales for measuring service quality: a comparative assessment based on psychometric and diagnostic criteria”, Journal of Marketing, vol. 70, nr. 3, 1994, pp. 201-230. 17. Donnelly, M., M. Wisniewski, J.F. Dalrymple and A.C. Curry, “Measuring service quality in local government: the SERVQUAL approach”, International Journal of Public Sector Management, vol. 8, nr. 7, 1995, pp. 15- 20. 18. Donnelly, M. and J.F. Dalrymple, “The portability and validity of the SERVQUAL scale in measuring the quality of local public service provision”, in Proceedings of the ICQ-1996 International Conference on Quality, October 15th-18th, 1996, Yokohama (Japan). 19. Gefen, D., D. W. Straub, and M.C. Boudreau, “Structural equation modeling and regression: guidelines for research practice”, Communications of the Association for Information Systems, vol. 4, nr. 7, 2000, pp. 1- 70. 20. McArdle, J.J., “Principles versus principals of structural factor analyses”, Multivariate Behavioral Research, vol. 25, nr. 1, 1990, pp. 81-87. 21. Chin, W.W., PLS-graph user’s guide version 3.0, 2001, (user’s manual provided by Wynn Chin together with PLS-GRAPH version 3.00 build 279 software). 22. Dijkstra, T., “Some comments on maximum likelihood and partial least squares methods”, Journal of Econometrics, vol. 22, 1983, pp. 67-90. 23. Dijkstra, T., Latent variables in linear stochastic models: reflections on maximum likelihood and partial least squares methods, 2nd ed., Sociometric Research Foundation, Amsterdam (The Netherlands), 1985.