Measuring high performance work practice systems: the training, information, participation and autonomy (tipa) scale

The primary aim of this research was to develop a scale for measuring HPWPs and to validate it using structural equation modelling. The study sample comprised 3046 employees of 33 organizations in the industry and services sectors of the Basque Country (northern Spain).

Available online at http://www.iaeme.com/ijm/issues.asp?JType=IJM&VType=11&IType=2

Journal Impact Factor (2020): 10.1471 (Calculated by GISI) www.jifactor.com

ISSN Print: 0976-6502 and ISSN Online: 0976-6510

© IAEME Publication Scopus Indexed

MEASURING HIGH PERFORMANCE WORK PRACTICE SYSTEMS: THE TRAINING, INFORMATION, PARTICIPATION AND

AUTONOMY (TIPA) SCALE Nekane Balluerka, Aitor Aritzeta*, Arantxa Gorostiaga

Faculty of Psychology, University of the Basque Country UPV/EHU, Spain

Unai Elorza and Damian Madinabeitia

Higher Polytechnic School, University of Mondragon, Spain

*Correspondence Author Email: aitor.aritzeta@ehu.eus

ABSTRACT

High performance work practices (HPWPs) have shown positive effects on organizational performance outcomes However, valid and reliable measures of HPWPs are lacking, as are tools that examine employee-level perceptions of HPWPs

in non-English speaking countries Thus, the primary aim of this research was to develop a scale for measuring HPWPs and to validate it using structural equation modelling The study sample comprised 3046 employees of 33 organizations in the industry and services sectors of the Basque Country (northern Spain) Development and validation of a measure to examine a HPWP ‘system’ was carried out over two studies In the first, and based on previous studies, we selected variables related to training, information, participation and autonomy (TIPA) Items for each variable were then developed, and their psychometric properties were analysed In the second study, we tested a structural equation model which hypothesized that TIPA scale scores would predict perceived organizational support and organizational commitment; relationships with firm size and employees´ educational level were also analysed with the aim of providing additional validity evidence Based on the results

of the two studies we conclude that the TIPA scale shows appropriate validity and reliability indexes and can therefore be considered an adequate tool for assessing a HPWP system encompassing training, information, participation and autonomy.

Keywords: High performance work practices; Training; Information; Participation; Autonomy; TIPA Scale; organizational commitment

Cite this Article: Nekane Balluerka, Aitor Aritzeta, Arantxa Gorostiaga, Unai Elorza and

Damian Madinabeitia, Measuring High Performance Work Practice Systems: The Training,

Information, Participation and Autonomy (TIPA) Scale, International Journal of Management

(IJM), 11 (2), 2020, pp 248–263

http://www.iaeme.com/IJM/issues.asp?JType=IJM&VType=11&IType=2

1 INTRODUCTION

Theories of strategic human resource management (SHRM) systems sustain that a ‗system‘ consisting of various human resource practices can have considerable positive effects on an organization‘s performance outcomes (Beckmann & Kuhn, 2010) However, inadequate implementation or the use of a single SHRM activity or high performance work practice (HPWP) may show no positive effect or even have a negative influence on the organization‘s economic performance and employee outcomes (Ichniowski, Shaw, & Prennushi, 1997) Some HPWPs are synergistic, since the overall effect of the HPWP system is greater than the sum of the effects of the individual practices by themselves However, there is currently no consensus about which practices and techniques are the most appropriate for building HPWP activities or bundle In fact, researchers have demonstrated that there is no such thing as the best bundle, but rather a set of practices that are the most suitable in each case (Ones, Denis,

& Schmidt, 2017)

Most of the published research in this field has been conducted in the USA and the UK (Guest, Michie, Conway, & Sheehan, 2003), and consequently there is a strong need for additional evidence to support the strategic HPWP-performance relationship from different contexts (Gerhart, 2007) In Spain, research on strategic HPWPs (i.e HPWP systems) is still scarce A review of 67 studies covering HPWPs in Spain between 2001 and 2010 indicated that although the increase in the number of publications during this period was larger in Spain than in any other European country, no studies were conducted with the aim of creating and validating measures of a HPWP system (Bayo-Moriones & Larraza-Kintana, 2012) Furthermore, the majority of studies conducted in this field have focused on managers‘ perceptions or on actual or implemented HPWPs (organizational level), rather than on employees´ perceptions of these HPWPs In this context, some studies have shown that the relationship between employees‘ perceptions of HPWPs and their actual behaviour is far from constant across organizations This means that it is the perception of the system of practices, rather than the system itself, that is the key to understanding workers‘ behaviour (Elorza, Harris, Aritzeta, & Balluerka, 2016) In order to address these gaps in the literature and to further examine the process through which HPWPs impact employees´ behaviour and perceptions, as well as organizational performance, it is important to conduct analyses in a non-US/UK context

Given the gap in the SHRM literature in Spain and the need for reliable and valid measurement tools in native languages, the primary aim of the present research was to develop and validate a scale to measure employees´ perceptions of a HPWP system defined

by four organizational practices: training, information, participation and autonomy (TIPA) The development and psychometric validation of the TIPA scale are presented over two empirical studies Study 1 involved the process of developing the scale and analysing the psychometric properties of its constituent items In study 2, and with the aim of giving more support to the validation process, we used a structural equation model to test the hypothesis that TIPA scale scores would predict employees´ perceived organizational support and organizational commitment

2 TRAINING, INFORMATION, PARTICIPATION AND AUTONOMY

AS PART OF A HPWP SYSTEM

The SHRM literature assumes that a bundle of combined variables — rather than single practices — affects employees' behaviour and attitudes, both of which will, in turn, contribute

to organizational performance This suggests that it is more important to consider people‘s perceptions of HPWPs than managers‘ HPWP reports (Arthur & Boyles, 2007) There is, in fact, evidence that rather than the actual or implemented HPWP it is employees‘ perceptions

of HPWPs that determine their behaviour and, therefore, outcomes (Elorza, Aritzeta, & Ayestaran, 2011) The intended HPWP system is mostly defined by the leader‘s knowledge and perception of the practices that will maximize employees´ positive responses to the organization However, not all intended HPWPs are implemented, and those that are often differ from the leader`s initial intention, due to divergent perceptions of the leader‘s strategy and the actual strategy implemented (Mintzberg, 1983), or simply because of non-uniform input from the middle managers responsible for implementing HPWPs (Zohar & Luria, 2005)

In addition to these likely differences, actual HPWPs may be perceived and interpreted differently by employees Since employees´ perceptions of HPWPs vary depending on their values, personality and other individual variables, employees will show different attitudinal and behavioural reactions, which may affect organizational performance outcomes (Nishii & Wright, 2008)

From the perspective of SHRM theory, the HPWP system should aim to enhance employees' ability, motivation and opportunity to perform, which, in turn, would boost performance in manufacturing organizations (Appelbaum, Bailey, Berg, & Kalleberg, 2000)

In the SHRM field this is known as AMO theory and it is the most widely used SHRM theory among those that include individual employee-level processes (Paauwe, 2009) It evokes three different employee characteristics: (A) the ability to perform (selection, skills improvement); (M) stimulating employee motivation by offering outstanding work conditions (career paths, information provision, excellent rewards); and (O) providing the opportunity for performance (stimulating teamwork, voice and employee involvement) Since its inception, this model has become dominant in SHRM research

The set of practices that together look to stimulate employees´ performance is also referred to as a high-involvement work system (HIWS), with the most widely researched set

of HIWS practices being the so-called 'PIRK' bundle (based on Lawler, 1986) The acronym

PIRK refers to power (participation and autonomy of the individual to make decisions within the company), information (different relevant data about outputs, costs, profitability, etc for employees), rewards (focusing their power, information and knowledge in a way that benefits the organization) and knowledge (training and development in relation to skills and abilities)

(Vandenberg, Richardson, & Eastman, 1999) Although the value of the PIRK model for employee and organizational performance alike has been widely tested in a variety of contexts (Butts, Vandenberg, DeJoy, Schaffer, & Wilson, 2009; Riordan, Vandenberg, & Richardson, 2005; Vandenberg et al., 1999), such practices are firmly grounded in the work enrichment and worker empowerment literature (Maynard, Gilson, & Mathieu, 2012; Wood, Van Veldhoven, Croon, & de Menezes, 2012)

Since the organizational sample used in the present research does not involve any (or barely any) job rotation the ‗abilities policy‘ cannot be improved by selecting and hiring qualified people Therefore, it is important to train and develop the skills and knowledge of existing employees in order to improve this aspect Aligning employees with the organization's objectives will enhance motivation, which could be achieved through participation, information, global performance-related retribution or employment security Finally, opportunity policies can be developed through autonomy and customer orientation (internal or external) The main objective of these policies is to empower employees to perform freely in pursuit of the organization's common goals While we acknowledge that a larger number of different practices could be implemented, the objective of this research is not

to identify all these practices but, rather, to examine employees‘ perceptions of the specific training, information, participation and autonomy practices which are central to most theories

in SHRM (i.e the AMO theory) Although these practices are also central to the HIWS approach (the PIRK model) we have not considered another core practice, that is, the

contingent reward The reason for this was that most of the participants in this study were members of cooperatives In cooperatives the contingent reward policy is implicitly

considered, so no variability was expected in this HPWP

It seems obvious that AMO policies should be implemented together or should enable other HPWP policies to grow through the development of some of them In this regard, Garman, McAlearney, Harrison, Song, and McHugh (2011) proposed a conceptual model of HPWPs with four practice bundles, comprising 14 management practices and nine factors influencing the adoption and perceived sustainability of these practices, and in which autonomy, participation, information and training practices were considered central In a similar vein, Boxall and Macky (2009) argued that giving employees autonomy and allowing them to participate in decision making about their work will make them feel more motivated Thus, they will try to perform better and will use all the information the organization provides them with to the benefit of common goals In addition, receiving information about organizational results may make them feel more attached to the organization, in which case they will try to take advantage of all the training they receive in order to reach common personal and organizational objectives If they lack training they may feel frustrated at not being able to perform at their maximum, and if they do not receive information about the results they are obtaining through their work they may not feel involved in the project In the event that they lack autonomy as regards using their training in new (innovative) ways, they may feel as if they have wasted time on training, leading them to be less motivated If the different HR policies are working together the outcome may be additive or even synergistic (Macky & Boxall, 2007) However, if some policies are encouraging employees while others are limiting their experience, the outcome may become negative rather than simply of no benefit This line of reasoning leads us to expect that all SHRM policies will be subsumable under a higher-order factor which we will call the ‗system‘

3 THE EFFECT OF THE ‘SYSTEM’ ON INDIVIDUALS

Social exchange theory is based on the idea that when employees perceive they are supported

by the organization in relation to their job performance activities they will reciprocate with an exchange behaviour that is regarded as fair to the other party This reciprocal exchange will take the form of attitudes and behaviours that will benefit the organization (Blau, 1964)

In this context, research has shown a positive association between the implemented HPWPs of AMO theory and perceived organizational support (POS) Specifically, Nishii and Wright (2008) and Whitener (2001) suggested that employees showed high POS through the help of HR practices In particular, HR practices combined as a system will result in employees having a positive perception of organizational support (Kehoe & Wright, 2013) For their part, Eisenberger, Armeli, Rexwinkel, Lynch and Rhoades (2001) noted that employees consider the organization to have a positive or negative orientation towards their contribution and well-being based on the organization‘s SHRM practices This means that the behaviour of an employee is influenced by his/her perceived organizational support A positive perception of organizational support leads to commitment from employees and a desire to repay the organization by attaining better outcomes and increased performance (Eisenberger, Huntington, Hutchison, and Sowa, 1986) This relationship between HPWPs and POS is explained by employees´ perceptions of implemented practices, because what really affects employees´ behaviours are their perceptions of HPWPs rather than their literal exposure to such practices (Elorza et al., 2011) In order to elicit the desired attitudes and behaviours from employees, it is therefore more important to meet their perceptions (as far as possible) than it is to implement more HPWPs or to seek to implement them perfectly (Lepak, Jiang, Han, Castellano, & Hu, 2012; Van de Voorde, Paauwe, & Van Veldhoven, 2012) In

light of the above, we expect that the selected practices will strengthen employees‘ perceptions about organizational support and, therefore, we expect that the HPWPs, measured and defined as a system, will show a positive effect on workers‘ POS

Commitment is generally defined as ―a force that binds an individual to a course of action

of relevance to one or more targets‖ (Meyer & Herscovitch, 2001, p 301) According to these authors, commitment in the workplace can take a number of forms, such as commitment to the organization, the job, the occupation or profession, the supervisor and co-workers The association between HPWPs and organizational commitment has been demonstrated in several studies (e.g Ang, Bartram, McNeil, Leggat, & Stanton, 2013; Kehoe & Wright, 2013; Takeuchi & Takeuchi, 2013), the results of which have shown a positive relationship between SHRM practices and organizational commitment For example, a study in New Zealand by Edgar and Geare (2005) found that HPWPs had a significant positive relationship with employee´s organizational commitment, while Innocenti, Pilati, and Peluso (2011) showed, using structural equation modelling, that such practices had a significant positive effect on employees´ organizational commitment in Italy Ang et al (2013) found that when employees perceived solid high-performance work systems, this had a positive and significant impact on affective commitment Boon, Den Hartog, Boselie, and Paauwe (2011), using regression analyses with cross-sectional data, found that perceived HPWPs had a significant positive relationship with employees´ organizational commitment in The Netherlands Based on the aforementioned evidence, we expected to find a positive association between our HPWP

system and employees‘ organizational commitment

4 STUDY 1

4.1 Overview

The purpose of Study 1 was to generate and refine items for measuring HPWPs and to test the psychometric properties of the items that would then constitute the TIPA scale (i.e those covering training, information, participation and autonomy) The latter was done by examining the dimensionality and convergent validity of the TIPA scale using confirmatory factor analysis Reliability was analysed by means of Cronbach‘s alpha, composite reliability (CR) and the average variance extracted (AVE) Comparisons between the values of the squared inter-dimensional correlations and the AVE values for these dimensions were used to provide evidence of the discriminant validity of the TIPA scale

4.2 Content Background

The initial version of the scale was developed by drawing up a set of self-rating items for four core aspects of a HPWP system: training, information, participation and autonomy We examined the review conducted by Garman et al (2011) in order to consider the key elements associated with HPWPs Items were written by a panel of two doctoral-level specialists, taking into account the HPWPs and policies that are mainly used in Spanish firms First drafts

of the items were discussed with managers and supervisors from several organizations and were improved prior to inclusion in the final battery The initial TIPA scale comprised 12 items, three for each of the abovementioned dimensions (see Appendix 1) All items were designed to be answered on a Likert scale ranging from 1 (Totally disagree) to 6 (Totally agree)

5 METHOD

5.1 Participants

Participants for this study were 3046 employees of 33 organizations from the industry (70.7

%) and services (29.3%) sectors Overall, 87.6% of employees belonged to organizations that were cooperatives, 11.9% to private limited companies, and 0.6% of employees did not report any information on the type of organization; 10.8% of them worked in small organizations (fewer than 50 employees), 50% in medium-size organizations (between 50 and 250 employees), 32.9% in large organizations (more than 250 employees), and 6.3% did not report the size of their organization With respect to age, 14.2% of employees were under 30, 39.9% were aged between 30 and 45, 6.1% between 45 and 60, 6.1 % were over 60, and 18.2% did not report their age

5.2 Procedure

Employees were invited to complete a structured questionnaire in order to provide information about the perceived system of HPWPs and discretionary behaviour Employees‘ regular meetings were used to inform and invite them to complete the questionnaire A few days after the study was announced, employees who wished to participate reported to a specified room in the organization, where they completed the survey Before administering the questionnaire, the general objectives of the study were explained and standard instructions were given, emphasizing the confidentiality of all information provided and the voluntary nature of participation in the study

5.3 Data Analysis

To test the four-dimensional structure and the convergent validity of the TIPA scale, a robust least squares factor analysis was conducted over the polychoric correlation matrix, using the EQS v6.3 software (Bentler, 2006) This estimation method was used for two reasons: 1) the indicators were not measured on a continuous scale, and 2) this method produces accurate test statistics, parameter estimates and standard errors under a great variety of conditions The adequacy of model fit was assessed by means of the following indices: Satorra-Bentler (SB) chi-square statistic (Satorra & Bentler, 1994); Bentler-Bonnet normed fit index (BBNFI) and non-normed fit index (BBNNFI) (Bentler & Bonnet, 1980); comparative fit index (CFI) (Bentler, 1990); incremental fit index (IFI) (Bollen, 1989); and the root mean square error of approximation (RMSEA) (Browne & Cudeck, 1993; Steiger, 2000) Values for the BBNFI, BBNNFI, CFI and IFI range from zero to 1.00, and a value above 95 indicates a good fit (Hu

& Bentler, 1999) For the RMSEA, values of less than 05 are considered a close fit, and those

of less than 08 an adequate fit (Finch & West, 1997)

Cronbach‘s alpha coefficients, composite reliability (CR) and average variance extracted (AVE) values were used to estimate the reliability of each dimension of the scale As Cronbach‘s alpha assumes that the items have been measured without error, it sometimes overestimates reliability (Bollen, 1989) Hence, CR and the AVE constitute good complements to this index The conventional cut-off criterion for an acceptable alpha statistic

is 70 and above (Nunally & Bernstein, 1994) The same cut-off value can be considered acceptable for CR (Bacon, Sauer, & Young, 1995) The AVE is generally considered acceptable at or above 50%

6 RESULTS

6.1 Convergent Validity and Dimensionality

The factor loadings of the items ranged between 72 and 94, except for the item ―Siento que

me falta información para hacer bien mi trabajo (I feel I need more information to do my job properly)‖, part of the Information dimension or latent construct, whose value was 28

Furthermore, the Lagrange multiplier suggested that this item should also be related to the other three dimensions Given that convergent validity is only guaranteed when all the items have factor loadings above 60 (Bagozzi & Baumgartner, 1994) and show clear relationships with their corresponding dimensions (Hatcher, 1994), the abovementioned item was removed from the TIPA scale After removing this item, all the obtained factor loadings were statistically significant and ranged between 74 and 94, indicating that all the items were important to the definition of their corresponding dimensions and that the scale had convergent validity

The four dimensions or latent constructs assessed by the TIPA scale, with indicator items, item loadings and their statistical significance, are shown in the first four columns of Table 1

The value of the Satorra-Bentler chi-square statistic (df = 38, n = 3,040) equal to 487.13, p =

.0001 indicated a statistically significant lack of fit of the model However, the sensitivity of the chi-square statistic to the violation of the assumptions on which it is based (Bollen, 1989) and, specifically, its dependence on sample size (Floyd & Widaman, 1995) means that the assessment of fit should be based mainly on alternative indices Those measures of fit which are less sensitive to sample size showed a good fit, with values of the BBNFI (.98), BBNNFI (.97), CFI (.98) and IFI (.98) close to 1.00 and the RMSEA equal to 06

Table 1 Reliability and convergent validity of the TIPA scale Dimensions or

latent

constructs

Indicator Standardized

loading

t robust Cronbach’s α CR AVE

Training

Participation

Autonomy

S-B (df = 38, n = 3,040) = 487.13 (p < 0001); BBNFI = 98; BBNNFI = 97; CFI = 98; IFI = 98; RMSEA

= 06

* p < 05

6.2 Reliability

Cronbach‘s alpha coefficients for each dimension or latent construct are shown in the fifth column of Table 1 The indices ranged between 74 and 93, exceeding the recommended value of 70 Composite reliability (CR) values were above 70 for all the dimensions, ranging between 75 and 93 The AVE values ranged between 60 and 83, exceeding the recommended level of 50 (see the last two columns of Table 1)

6.3 Discriminant Validity

The AVE values for the four dimensions were higher than the squared inter-construct correlations (the latter values ranged between 22 and 43), which can be regarded as evidence for the discriminant validity of the TIPA scale

7 STUDY 2

7.1 Overview

The overall results obtained in Study 1 led us to propose a scale comprising 11 items corresponding to the four dimensions defined in our HPWP system: training, information, participation and autonomy These items provide the basis for turning these four dimensions into observed variables by averaging their indicators, thus enabling us to analyse their predictive capacity by means of a structural equation model Specifically, the hypothesis is that the four dimensions of the TIPA scale, as a HPWP system, have a direct effect on employees‘ perceived organizational support and organizational commitment As a prior step,

we test the reliability, convergent validity and discriminant validity of the measurement model Evidence of validity based on relationships between the TIPA scale and other relevant variables is also provided

The specific purpose of Study 2 was to test the psychometric properties of the TIPA scale,

in other words, to examine the components of a structural equation model in which the HPWP system is considered as an exogenous latent construct and perceived organizational support and organizational commitment are treated as endogenous unobserved variables In addition

to providing evidence of validity based on the predictive capacity of this model, relationships between HPWPs and relevant organizational variables such as size of the organization, employee position and employees‘ level of education are analysed in order to complete the validation process

8 METHOD

8.1 Participants and Procedure

The sample was the same as in Study 1, and the same procedure was used for data collection

8.2 Measures

In addition to the aforementioned TIPA scale, three additional tools were included in the pool

of instruments The first was a questionnaire that gathered employees‘ sociodemographic data and characteristics of the organization, and which included questions about age, department, level of education and size of the organization, among others Perceived organizational support (POS) (Eisenberger, Cummings, Armeli, & Lynch, 1997) was measured by means of three questions answered on a Likert scale ranging from 1 (Totally disagree) to 6 (Totally agree) A sample item is ―My organization really cares about my wellbeing.‖ The internal consistency of this tool, measured by Cronbach‘s alpha, was 87 Finally, affective commitment was measured using a shortened version of the instrument proposed by Meyer, Allen, and Smith (1993) The items used were those which showed the highest factor loadings

on this dimension Items (an example of which is ‗I feel a strong sense of belonging to my organization‘) were answered on a Likert scale ranging from 1 (Totally disagree) to 6 (Totally agree) The alpha coefficient value for this variable was 91

8.3 Data analysis

The convergent validity and the overall fit of the measurement model were analysed by means

of a robust least squares factor analysis conducted over the polychoric correlation matrix,

using the EQS v6.3 software (Bentler, 2006) Cronbach‘s alpha coefficients, CR and AVE were again used to estimate the reliability of the people-based management system, perceived organizational support and organizational commitment latent constructs

The value of the squared correlation between two latent constructs was compared with the AVE values of these latent constructs in order to obtain evidence of the measurement model‘s discriminant validity The difference between the fit of the overall measurement model and the fit of the structural model was taken as nomological validity evidence The hypotheses about the direct effects of a people-based management system on perceived emotional support

and organizational commitment were tested by means of the standardized B coefficient, unstandardized t coefficient and R square statistic

Finally, the Mann-Whitney U test was used to analyse whether the implementation of a

people-based management system in an organization was associated with the size of the organization or his/her educational level

9 RESULTS

9.1 Convergent Validity and Overall Fit of the Measurement Model

The latent constructs included in the structural model, with indicator items, item loadings and their statistical significance are shown in the first four columns of Table 2

Table 2 Reliability and convergent validity of the measurement model Latent

construct

Indicator Standardized

loading

t robust Cronbach’s α CR AVE

TIPA

POS

OC

S-B (df = 42, n = 3,040) = 466.15 (p < 0001); BBNFI = 95; BBNNFI = 94; CFI = 96; IFI = 96; RMSEA =

.08

* p < 05 POS: Perceived organizational support; OC: Organizational commitment

The factor loadings of the items ranged between 69 and 75 for the TIPA latent construct, between 80 and 87 for perceived organizational support, and between 77 and 87 for organizational commitment, indicating that the items were important for the definition of their corresponding constructs and that the model has convergent validity

The value of the Satorra-Bentler chi-square statistic (df = 41, n = 3,040) equal to 466.15, p

= 0001 indicated a lack of fit of the model However, as already noted, the sensitivity of the chi-square statistic to sample size means that the fit assessment should be based on alternative indices Measures of fit less sensitive to sample size indicated that the hypothesized model showed a good fit to the data, with BBNFI (.95), BBNNFI (.94), CFI (.96) and IFI (.96) values close to 1.00 and a RMSEA value equal to 08

9.2 Reliability

Cronbach‘s alpha coefficient values for TIPA, perceived organizational support and organizational commitment were 80, 89 and 87, respectively The corresponding CR values

were 82, 89 and 88, and the AVE values 53, 72 and 64, respectively These sets of values allow us to conclude that the measurement model shows good reliability

9.3 Discriminant and Nomological Validity

Except for one case, the AVE values for the three latent constructs were higher than the squared inter-construct correlation These results indicate that the discriminant validity of the model is acceptable Furthermore, as the difference between the fit of the overall measurement model and the fit of the structural model was not statistically significant, it can

be concluded that the measurement model has nomological validity

9.4 Predictive Capacity of the Structural Model

Upon analysing the hypothesized association between latent constructs, several significant relationships emerged (see Figure 1) Scores on the TIPA scale had a direct and positive effect

on both perceived organizational support (B = 86, t = 30.76, p < 05, R 2 = 74) and

organizational commitment (B = 75, t = 27, p < 05, R 2 = 56)

9.5 Relationship with other Variables

Figure 1 Predictive capacity of the structural model