Measurement of the perceptions of human resource practices in a seemingly collectivistic culture

The human resource (HR) function has a direct impact not only on organisational performance but also on employee behaviour and attitudes (Choudhury & Mishra, 2010). This relationship is complex and one should guard against oversimplifying the nature thereof (Bowen & Ostroff, 2004; Nishii & Wright, 2008). Human resource practices build and sustain organisational effectiveness (Enz & Sigaw, 2000) and are especially important in attracting, retaining and developing the skills and knowledge of employees (Choudhury & Mishra, 2010). Human resource practices must be deliberately chosen and used strategically to maintain strong organisational boundaries, to promote high levels of organisational and professional identity; more specifically, it encourages the retention of staff in a highly competitive industry (Choudhury & Mishra, 2010). This study is not focused on the actual effectiveness and success of HR practices or HR practices as presented in policy documents (Wright & Nishii, 2007) but rather the perceptions of employees of HR practices. Orientation: Human resources (HR) practices and specifically the perceptions thereof are not only important for organisational strategy and performance but have a direct impact on employee attitudes and behaviour. The accurate measurement of these perceptions is therefore important.


Introduction
No organization can have exceptional performance without exceptional employees; and that requires exceptional human resource (HR) practices.(Enz & Sigaw, 2000:48) The human resource (HR) function has a direct impact not only on organisational performance but also on employee behaviour and attitudes (Choudhury & Mishra, 2010).This relationship is complex and one should guard against oversimplifying the nature thereof (Bowen & Ostroff, 2004;Nishii & Wright, 2008).Human resource practices build and sustain organisational effectiveness (Enz & Sigaw, 2000) and are especially important in attracting, retaining and developing the skills and knowledge of employees (Choudhury & Mishra, 2010).Human resource practices must be deliberately chosen and used strategically to maintain strong organisational boundaries, to promote high levels of organisational and professional identity; more specifically, it encourages the retention of staff in a highly competitive industry (Choudhury & Mishra, 2010).This study is not focused on the actual effectiveness and success of HR practices or HR practices as presented in policy documents (Wright & Nishii, 2007) but rather the perceptions of employees of HR practices.Boon, Hartog, Boselie, and Paauwe (2011), the developers of the instrument investigated in this study, based their work on the premise that the perceptions of HR practices are essential for employees' experiences in an organisation.These perceptions are mainly formed through their personal interpretations and social constructions of the HR practices.
It is based on the extent to which employees feel that the HR practices are indeed offered to them.
Because of the importance of perceptions of HR practices, an accurate, reliable and valid measurement is essential, not only for academic purposes but also for HR management and HR practitioners.Various instruments were considered for the South African multicultural context, resulting in the selection of the Human Resource Practices Perceptions Questionnaire (HRPPQ), developed by Boon et al. (2011).This instrument was developed based on research by scholars such as Cable and Edwards (2004), Guest and Conway (2002), Ryan and Schmit (1996) and Ten Brink (2004).It is used to measure perceived HR practices in the Netherlands.The researchers reported strong relationships between employee perceptions of HR practices and person-organisation and person-job fit, as well as other employee outcomes.
Employees within the Dutch context were the focus of all these studies.
The purpose of this study was subsequently to build on the extensive work of Boon et al. (2011), Cable and Edwards (2004), Guest and Conway (2002), Ryan and Schmit (1996) and Ten Brink (2004) but within the South African context.This study aimed to validate the HRPPQ for the South African (African) context while considering contextual realities such as collectivism (embedded in, for instance, ubuntu) and Afrocentrism.Previous research added evidence assuming that most black South Africans are representative of the collectivistic culture (Eaton & Louw, 2000;Thomas & Bendixen, 2000) and that white South Africans are mostly individualistic.This research was approached from the position that the majority of the working population in South Africa is from the black South African race group, followed by 6% white people and approximately 5% mixed race and 2% Indian.Therefore, the culture is regarded to be leaning towards collectivism, although it is acknowledged that inter-and intra-race group comparisons are much more complex.The HRPPQ was predominately tested from a Western individualistic standpoint by Boon et al. (2011).This poses contextual questions about the transferability of the instrument to a collectivistic culture, subscribing to an ubuntu relational philosophy, that has an impact on organisational leadership and organisational behaviour in general (Grobler & Singh, in press), as well as on how HR practices are perceived by employees.
According to Mbigi (2000) and Mulemfo (2000), the ubuntu philosophy asserts a sense of community whereas Western individualism promotes differentiation, independence and self-identification (Bochner, 1994).Hofstede (2011) highlighted collectivism to be the promotion of individual integration into society and a sense of belonging, which can also be brought into the context of organisational practices and functioning.Newenham-Kahindi (2009) referred to ubuntu and collectivism interchangeably, as collectivism is regarded to be the underlying foundation of ubuntu.
The objectives of this study were as follows: firstly, to determine the construct (factorial) validity of HRPPQ (Boon et al., 2011) within the South African context; secondly, to develop a reconfigured factor structure should the original configuration yield unsatisfactory results.Thirdly, the final measurement instrument was assessed for convergent and discriminant validity, and lastly a cross-validation analysis was performed to establish whether the instrument is suitable for use in both the private and public sectors.
Recommendations for future research are also made.

Literature and theoretical concepts
Collectivist culture Hofstede (2011) defined culture as the programming of the mind, distinguishing one group from another in varying contexts of nations, ethnicities and/or organisations.The national cultural elements are often also found in business and organisational cultures.Hofstede's model and paradigm for national cultural comparisons is regarded as relevant in the classification of the South Africa national cultural context, with specific reference to individualism versus collectivism.Hofstede's (2011) complete model entails other dimensions such as power distance, uncertainty avoidance, masculinity versus femininity, long-term versus short-term orientation and indulgence versus restraint.
Hofstede's individualism versus collectivism dimension offers the most appropriate framing and understanding of the South African cultural and organisational context (Hofstede, 1998).According to various scholars this cultural variable has an impact on organisational leadership, processes and practices as well as the attitude of employees (Bochner, 1994;Grobler & Singh, in press;Li, Zhang, Yang, & Li, 2015;Wiley, 2000;Yilmaz & Hunt, 2001).
Individualism, according to Hofstede (2011), is characterised by loose ties between individuals; Bochner (1994) iterates the prominence of individualism, autonomy and separateness to be predominately Western.In contrast, the collectivism approach entails an integration of people (individuals) into a group or society.It is characterised by values of 'we-consciousness', belonging, loyalty, relationships, collaboration and group decision-making.The collectivism paradigm is closely aligned to the Afrocentric philosophy articulated in the act of ubuntu (Gathogo, 2008) prevalent in South African cultural settings.
Ubuntu (humanness) means that a person is viewed in the context of the collective within a society and not as an individual (Newenham-Kahindi, 2009).This view is contrary to the Western contexts in which people are seen as autonomous, independent beings.Mabovula (2011) contended that life, including work activities, is carried out in socio-ethical settings.Largely, problem identification, solution generation and idea sharing are conducted collectively within the community settings.Ubuntu encompasses the values of solidarity, a strong sense of community, harmonious relationships, respect, hospitality and care for others in society.

Human resources practice perceptions and collectivism
Human resource practitioners need to take the employee value proposition and exchange, socio-ethical codes of the practices and the cultural context into account.Human resource practices and how they are perceived influence the activities of attracting, recruiting, engaging, retaining, developing and rewarding employees.Human resource practices by nature have ethical, legal and financial implications that largely influence the perceptions that employees have of these practices in an organisation.A study conducted by Browning (2006) added evidence that HR practices have a direct influence on performance, output quality and the behaviour of employees.Valentine, Hollingworth, and Francis (2013) view HR practitioners as the custodians and promoters of ethics in organisations.It is recommended that ethical approaches be maintained in all HR practices and programmes, influencing the organisation, community, relevant stakeholders and employees.Wiley (2000) advocates the five HR ethical codes of integrity, legality, proficiency, professionalism and confidentiality.Ubuntu prescribes social ethics and rule of conduct on how to relate to and interact with others (Gathogo, 2008).Human resource practices within an organisation need to have a strong socio-ethical fibre, aligned to the ubuntu/collectivist values of generosity, protection of human dignity, respect, benevolence, community rootedness and adherence to communal norms (Mabovula, 2011).Human resources promotes ethics through fairness, transparency, and organisational justice practices that promote employee wellbeing (Valentine et al., 2013).
Human resource practices, policies and strategies must largely refer to the national and organisational cultural contexts as culture influences application and the outcomes of HR interventions.The knowledge and understanding of the applicable cultural context leads to the appropriate application of HR practices.Based on Li et al. (2015), collectivistic HR management enables practices such as team-based performance, incentive schemes, rewards, training and development through policies and procedures that promote joint responsibility, cohesion and collaboration.
The HR practices in an organisation need to be comprehensive, integrated and aligned to policies, procedures and the organisational culture.The contents of the South Africaspecific HRPPQ are briefly outlined: Job design: Job characteristics, namely skill variety, task identity, task significance, autonomy and feedback, are closely associated with intrinsic motivation, job satisfaction, engagement and performance (Abbott, Boyd, & Miles, 2006).
In designing job content aspects such as meaningfulness, a sense of responsibility and knowledge of the performance outcomes need to be considered in order to engage, motivate and enhance the quality of performance.

Participation and autonomy:
In a study conducted by Han, Chiang, and Chang (2010) employee participation in decisionmaking was found to be significantly related to psychological ownership, positive work attitudes and behaviours.Swamy (2013) supported employee autonomy as the meeting of employees' motivational needs, to contribute, coordinate and control work inputs.In a study conducted by Bochner (1994) on cross-cultural difference in the self-concept, Bochner acknowledged the self (individual) to be complex, irrespective of the cultural background.Bochner (1994) further proposed that people contain qualitatively varying cognitions, in that they operate and live as private individuals, in a collective and public society.Therefore, people will place varying emphasis on cultural application for each situation.For example, each one has his or her own personal beliefs, identity, talents, skills, interests, qualities, attitudes, traits and psychological states that do not relate to other people.Being in a collectivist culture does not imply total absence of individuality and self-determination (Bochner, 1994).The aforementioned literature advocates for the incorporation of the elements related to autonomy and decision-making in job design in order to achieve job enrichment and meaningful work where employees feel that they can contribute.
Recruitment and selection: Ofori (2011) highlighted the importance of recruitment and selection practices as an entry point for skills, knowledge and competence into the business to ultimately determine success, performance and business sustainability.Recruitment and selection practices in a collectivistic organisation are likely to seek to attract potential candidates sharing similar values, with strong teamwork skills and strong cooperative work behaviours (Yilmaz, & Hunt, 2001).Recruitment decisions should foster collectivist norms and values that are supportive of the organisational culture (Yilmaz & Hunt, 2001).The collectivist approach to the recruitment of like-minded individuals may potentially limit individuality and diverse thinking in some instances.

Training and development:
Training entails systematic activities aimed at improving individual, team and organisational performance, through the transfer of skills, knowledge and competence (Aguinis & Kraiger, 2009).These activities include training needs assessments, training design and the delivery of training.Team-based and collectivist cultures will promote team-based training interventions, as well as those aimed at improving team membership skills.However, Aguinis and Kraiger (2009) differentiate development from training as the acquisition of skills and knowledge for purposes of personal and professional development.Development for career and succession planning equips individuals with adequate skills, experience and competence for future, senior and complex roles (Hills, 2009).Building internal talent capability allows the organisation to improve promotability and career mobility of employees from within, thus improving employee engagement.Collectivistic cultures will promote and progress the careers of those who work well with others (Li et al., 2015).According to Presbitero, Roxas, and Chadee (2016) organisations offering career progression and development opportunities mitigate against the risk of employee turnover.
Employment security: Perceived employment security is strongly associated with overall employee well-being (Virtanen, Vahtera, Kivimäki, Pentti, & Ferrie, 2002).Employment security is a major concern for employees; permanence thereof improves the quality of work and life because of the provision of income stability and the meeting of socio-economic needs of employees and their families (Swamy, 2013).The South African trade unions play a role in negotiating employment security for employees.
Teamwork: Teamwork is critical in collectivist cultures as the individual is absorbed and attached to a group (Bochner, 1994).Teamwork and performance are influenced by variables at both team and individual levels (Rosen et al. 2008).The multilevel approach in instilling team performance considers individual team members' skills, competence and the ability to perform the task within the organisational culture.Organisational culture must be facilitative of team cohesion, collectivist orientation, team membership skills (like dynamic interaction and communication with individuals with heterogeneous skill sets, knowledge and competence).Teamwork requires clear team performance measures and outcomes -for example, productivity, quality, sales, indicators (Scott & Einstein, 2001).Such measures hold the team accountable for monitoring performance results, identification of deficiencies and taking corrective action.Li et al. (2015) asserted that collectivistic cultures stress subordination of personal interests to the greater welfare and goals of the team and organisation.
Rewards: Literature confirms that fair remuneration inclusive of pay and benefit packages strengthens intentions to remain in organisations for longer (Presbitero et al., 2016).Interestingly, collectivistic HR practices are geared towards a team-based reward system rather than the individual (Li et al., 2015).Team-based rewards are tied to team-based performance targets.
Work-life balance: Work arrangements are found to influence the time and psychological energy shared with family, friends and participation in extramural activities (Albertsen, Rafnsdóttir, Grimsmo, Tómasson, & Kauppinen, 2008).Extended overtime, increased work hours and shift work are associated with lower work-life balance.Haar, Russo, Sune, and Ollier-Malaterre (2014) found work-life balance to be instrumental to promoting overall employee well-being, job satisfaction and satisfaction with life in employees across various cultures.
As mentioned before, the rationale of the instrument under investigation in this study was not to measure the effectiveness and success of the HR practices as discussed above but rather to measure whether the HR practices are indeed offered to them.

Research approach
This study employed a typical empirical paradigm using a cross-sectional design and quantitative analysis.Surveys were used as data generation techniques, which enabled the researchers to collect the required data at the same time.

Research participants
The participants consisted of employees of 30 organisations in South Africa, with 15 organisations being from the private sector, including the medical, engineering, retail, construction, financial, telecommunication, pharmaceutical and information technology industries.The public sector was equally represented, with 15 organisations, consisting of national and provincial departments, local government as well as state-owned enterprises.In each organisation, 60 employees were randomly selected to participate in the study.The pooled data could therefore be considered a convenience sample (because of the convenience of accessibility and availability of the respondents within the participating organisations).The fieldwork was conducted by 30 co-researchers working on a larger project; ethical clearance was granted by the institution.
The total sample consists of 1676 participants.In terms of the racial distribution, the majority of the participants were black (73%), followed by white (15.4%), mixed race (6.6%) and Indian (4.8%).The representation of the gender groups was higher for males at 51.5% compared to 48.5% for females.The racial and gender distribution of the sample seems to be relatively representative of the South African workforce in general, taking into consideration that the distribution of the workforce as indicated in Statistics South Africa (2017) was 87% black people, 6% white people (slightly overrepresented in the sample), 2% Indians and 5% mixed race people.According to the same source, the proportion of men in employment is 56% while the proportion for females stands at 44%.
The mean age of the respondents was 36.60 years (SD = 8.90), and the mean tenure in the specific organisation was 7.58 years (SD = 7.33).From these mean scores it is assumed that the respondents were on average mature and that they had the necessary exposure to the organisation and leadership to http://www.sajhrm.co.zaOpen Access be able to respond to the items related to leadership behaviour in their respective organisations.

Measuring instrument
The HRPPQ developed by Boon et al. (2011) was used to measure the perceptions of the HR practices and was the instrument under investigation in this study.In other words, the focus of the measurement was the extent to which employees perceived that specific HR practices were indeed offered to them.
The instrument consists of 38 items.A typical example of an item included in the instrument reads as follows: 'The organisation offers (me) … the opportunity to develop new skills and knowledge for my current job or for possible jobs in the future'.Participants were requested to rate the degree to which they agreed with the statements made on a fivepoint Likert scale.The scale ranged from 1 indicating that they strongly disagreed with the statement to 5 indicating that they strongly agreed with the statement.The maximum score that could be recorded for the instrument was 190 and the minimum was 38.
The 38 items in the instrument are practices that form part of seven high performance factors: (1) training and development, (2) participation, autonomy and job design, (3) appraisal and rewards, (4) teamwork and autonomy, (5) work-life balance, (6) recruitment and selection and (7) employment security of the Cronbach's alphas reported by Boon et al. (2011) for theses, seven factors were all above 0.75.The instrument was developed and standardised using a sample from two large organisations in the Netherlands, one in retail and one in healthcare, and was administered in Dutch.It was therefore deemed necessary to investigate its transferability into the South African context.

Statistical analysis
The statistical analysis was performed by using the Statistical Package for the Social Sciences (SPSS version 24), supported by SPSS AMOS (Analysis of Moment Structures).
The dataset was cleaned via case screening followed by variable screening in order to explain why there was variation in the data.It was deemed necessary to follow this process to ensure that there were no missing values in the data set and get a feel for the data set.The dataset was further inspected for unengaged responses by running a standard deviation and inspecting cases with SD < 0.50.The variables were further screened by means of kurtosis and the central limit theorem, to gather information about the distribution of the data.From the data cleaning process, it was deduced that the missing values were very sparse, and it was therefore not considered to be a possible contributor to any bias.
The first step of the exploratory factor analysis (EFA) was to evaluate the appropriateness of the sample size.The item-torespondent ratio of ±1:20 is regarded to be acceptable according to Meyers, Gamst, and Guarino (2013) and Tabachnick and Fidell (2007).Secondly, the inter-correlations between items were inspected using Bartlett's test of sphericity (Hair, Black, Babin, & Anderson, 2010).With this test, the statistic generated should be significant (p < 0.05) for an EFA to be considered an appropriate technique (Hair et al., 2010).Thirdly, the Kaiser-Meyer-Olkin measure was used to quantify whether the items correlated sufficiently to determine whether a factor analysis could be performed.The minimum level set for this statistic is 0.60 (Tabachnick & Fidell, 2007).
To aid in the interpretation of the initial results, orthogonal rotation, specifically the Promax rotation, was used.The decision regarding the number of variables (factors) to be retained was based on the Guttman-Kaiser eigenvalue greater-than-one rule (K1 rule), together with the scree plot (specifically the shape of the curve) and, lastly, the Monte Carlo PCA (principal component analysis) for parallel analysis.The purpose of the Monte Carlo parallel analysis is to determine the number of factors that account for more variance than the components derived from random data.The eigenvalues obtained from the actual data are compared to the eigenvalues obtained from the random data.If the actual eigenvalues from the principal component analysis of the actual data are greater than the eigenvalues from the random data, the component (factor) is retained.Meyers et al. (2013) indicate that a guide for variance accounted for by the factors needs to meet the lower limit of 50%.The Cronbach's alpha coefficient was determined for factors of the instrument, taking into consideration that the general rule according to Nunnally and Bernstein (1994) is α > 0.70.
To operationalise this construct definition, a higher (or secondary) order, multidimensional model of the HRPPQ was conducted by means of a confirmatory factor analysis (CFA).Confirmatory factor analysis is generally intended to examine whether a second-order factor exists and whether it explains the relationships among the lower-order factors (as identified by the EFA) with the Analysis of Moment Structures (AMOS) maximum likelihood procedure (Byrne, 2010).To assess the model fit, several fit indexes were used including the comparative fit index (CFI), the root-mean-square error of approximation (RMSEA), chi-square (χ 2 ) and the ratio of the differences in chi-square to the differences in degrees of freedom (χ 2 /df).Given that there is no one acceptable cut-off value for what constitutes adequate fit, it was elected to evaluate each model and to recommend the model closest to the CFI value of 0.90, an RMSEA value of 0.05 and χ 2 /df, a ratio of less than 5.00 or lower (Byrne, 2010).
The first model was a one-factor solution (unidimensional) in which all the items identified through the EFA were indicative of one larger human resource practice perception (HRPP) factor.The second was a first-order factor model in which items were allowed to load onto their respective factors.The third was a second-order factor model in which items were loaded onto their respective factors and the factors loaded on a second-order latent HRPP factor.
An elementary cross-validation assessment to determine invariance between the private and public sectors was also performed, as measurement invariance is regarded to be a prerequisite for meaningful interpretations and valid crossgroup comparisons.It is important that items and constructs are understood and interpreted in the same way across different samples; the variance in the observed score differences between groups (in this case the private and public sector) should not be a result of group membership but the construct being measured.The indexes of the CFA were used to assess the measurement invariance.
Finally, convergent validity was assessed by the composite reliability (CR) and the average variance extracted (AVE), with critical values of >0.70 and <0.50, respectively.The discriminant validity was determined by comparing the AVE with the maximum shared variance (MSV).Proof of discriminant validity would be when MSV < AVE and where the average shared variance (ASV) is less than the AVE (Hair et al., 2010).

Results
The EFA conducted was based on an adequate variableto-case ratio (44:1).The Kaiser-Meyer-Olkin measure of sampling adequacy was performed and the value was 0.94, exceeding the critical value of 0.60 (Tabachnick & Fidell, 2007), thus supporting the strategy to perform an EFA.The Bartlett's test of sphericity was also conducted and an approximate chi-square value of 25 930.71 was reported (degrees of freedom = 708).The chi-square value was further significant at a level smaller than 0.001, indicating sufficient correlation between the items to proceed with the EFA.
The Guttman-Kaiser eigenvalue greater-than-one rule (K1 rule) was used in conjunction with the scree plot to determine the number of factors.The results of Kaiser's criterion are reported in Table 1.
Ten factors reported eigenvalues larger than one, with the first factor accounting for 34% of the variance in the HRPP.The total variance accounted for by all 10 factors was close to 71%, thus exceeding the 50% guide by Meyers et al. (2013).
Cattell's scree test, which is aimed at retaining the components (factors) before the break (elbow rule), was performed on all three studies independently.The interpretation of the scree plot depicted is problematic, as it is not clear where the elbow flattens off.The interpretation of the scree plot as well as the K1 rule are regarded by Pallant (2013) as being too conservative as measures to determine the exact number of factors.As a result, a further, more stringent technique, namely the Monte Carlo parallel analysis simulation technique, was performed.Ten components were included in the Monte Carlo parallel analysis to test the assumption of a 10-component (factor) solution, as suggested by the interpretation of the K1 rule (see Table 1).The results are reported in Table 2.
The results reported in Table 2 indicate that the first nine components have actual eigenvalues larger than the criterion value from the parallel analysis.The results of the Monte Carlo parallel analysis suggest, thus, a nine-factor model.The nine factors accounted for 68% of the total variance (see Table 1).
The descriptive statistics on item level as well as the factor loadings, communalities (h) and structure coefficients from the Promax rotation are reported in Table 3.Each item starts with 'The organisation offers (me) …'.
The results of the EFA conducted on the HRPPQ are summarised in Table 3.A factor loading cut-off point of 0.40 for inclusion in the interpretation of a factor was used.Further, 37 of the original 38 items loaded on nine factors.The communalities are relatively high, ranging between 0.52 and 0.78.After inspection of Table 3 it is evident that the basis   rewards was therefore renamed rewards, and it consists of six items.
The results of the correlational analysis (Pearson correlation) between the nine retained components, from here on referred to as factors, are reported in Table 4.
The correlations between the pairs of factors ranged between 0.17 and 0.63.Because of these high correlations, multicollinearity was tested using the variance inflation factor (VIF) and tolerance.The analysis yielded results with the highest VIF value of 2.41 and the lowest tolerance value of 0.42.These values fall within the parameters of less than 5 for the VIF value and larger than 0.10 for the tolerance value.No evidence of multicollinearity was thus found.
The descriptive statistics as well as the internal consistency of each of the adapted factors as assessed by Cronbach's alpha coefficient are reported in Table 5.
The descriptive statistics in Table 5 show that five factors, namely job design, participation and autonomy, training, employment security, teamwork as well as the HRPPQ total score, reported relatively high mean scores (on a five-point Likert scale).Two factors reported below-average mean scores, namely recruitment and selection (M = 2.58, SD = 1.15) and work-life balance (M = 2.72, SD = 0.95).The skewness and kurtosis values for all factors did not exceed the critical values of 2.00 and 7.00, respectively, which is an indication that the data was normally distributed (West, Finch, & Curran, 1995).All factors, except recruitment and selection (which is not an HR practice that is generally utilised by employees) and work-life balance, reported negative values of skewness, with the skewness values ranging between -0.67 and 0.10.This is an indication that the distribution in general has relatively few small values and tails off to the left.The kurtosis value ranged from -0.89 (recruitment and selection) and 0.89 for teamwork.The Cronbach's alpha coefficients of the factors are acceptable if the guideline of a > 0.70 (Nunnally & Bernstein, 1994) is applied.It would thus appear that the factors possess acceptable levels of internal consistency with Cronbach's alpha coefficients ranging from 0.76 to 0.88.
It was further deemed necessary to inspect the factor structure of the HRPPQ for construct and specifically factorial validity by conducting a CFA.The results of the three models tested are reported in Table 6 in accordance with their respective fit indexes, more specifically the CFI, the RMSEA, chi-square (χ 2 ) and the ratio of the differences in chi-square to the differences in degrees of freedom (χ 2/df ).
An assessment of the best-fitting of the three models was conducted through the application of CFA.The second-order factor model (all 37 items from the original 38 items) was identified as the worst-fitting model (χ 2/df [598] = 3.82, CFI = 0.931, RMSEA = 0.048).By analysing the chi-square test values, it further appears that the first-order factor model is slightly better than the one-factor model.The difference in chi-square between the first-order factor and the one-factor models is 36 (i.e.It was further deemed necessary to conduct an elementary cross-validation analysis to assess the possibility of invariance between two sample groups, namely the private and public sectors.The sample was split into the two sectors, using 500 cases randomly selected from each sector.The results reported for the two sample groups were χ 2/df (571) = 2.24, CFI = 0.92, TLI = 0.90, RMSEA = 0.05, expected crossvalidation index (ECVI) = 3.33 and χ 2/df (571) = 2.35, CFI = 0.93, TLI = 0.92, RMSEA = 0.05, ECVI = 2.59 for the private   Convergent validity of the nine HRPPQ was assessed by the CR and the AVE, while the discriminant validity was determined by comparing the AVE with the MSV as well as the ASV.The results are reported in Table 7.
The HRPPQ with its nine factors met the conditions of convergent validity, with all CR values higher than the criterion of > 0.70 and the AVE value > 0.50.Evidence of discriminant validity was also found as all the AVE values were larger than the MSV value of 0.40.The ASV was also less than the AVE, which supports the notion of discriminant validity.

Discussion
The original instrument, the HRPPQ, was developed and standardised by Boon et al. (2011) but in the Netherlands, for a Dutch-speaking population.It was deemed necessary to validate the instrument within the South African context, which has a more collectivistic culture (compared to the West), which is closely related to the African relational philosophy of ubuntu.This was done from an etic perspective, with the rationale of making sure that accurate measurements could be made to enable HR managers and practitioners to design and implement interventions to improve and positively change the perceptions of the HR practices.This would have a positive impact on employee attitudes and behaviour in general.
The value of this study lies in the fact that the original conceptualisation of the HRPPQ -which originally consisted of seven factors (HR practices) -was adjusted to meet the contextual realities.The employment security as well as work-life balance factors remained the same as postulated by Boon et al. (2011).
The impact of the collectivistic work culture that is found in South Africa significantly influenced the terms of the original teamwork factor as well as the performance appraisal and reward factors.One significant item was added to the teamwork factor, which read, 'The organisation offers (me) periodic evaluation of my performance'.This item was initially part of the performance appraisal and rewards factor together with an item that read, 'The organisation offers (me) fair appraisal of my performance'.The latter was the only item that was excluded in the new structural configuration of the HRPPQ -performance in a more collectivistic culture is often seen in terms of a group effort and not as individual performance, as is the case in an individualistic culture.The initial factor of Boon et al. (2011) called performance appraisal and rewards was therefore renamed to rewards as it did not include any performance management items.
Another factor where the collectivistic and relational nature of ubuntu has impacted on the original factor structure, is the split of the factor related to participation, autonomy and job design into job design, which is considered to be more characteristic of the mechanistic factor, and the more relational side of the original factor, namely participation and autonomy.This factor is related to trust and respect (to allow employees to work on their own) but also to participate in decision-making and other organisational processes -which bears some resemblance to the ubuntu philosophy.
Based on the results obtained, it seems that the HRPPQ is a suitable (valid and reliable) instrument for measuring the perceptions of employees of the HR practices in their organisation.This study serves as a reference for the accurate measurement of perceptions of HR practices, as it was found to be invariant for both the private and public sectors during cross-validation analysis.

Limitations and recommendations
The limitations identified for this research are mainly in terms of the methodology.The HRPPQ is based on selfreporting, a method that may lead to method bias, and this may still be a reality, even with the assurance provided to participants during the briefing regarding anonymity as well as confidentiality.Social desirability and subsequent response bias will always remain a concern and a limitation in studies such as this one, while self-reporting may be seen as a onesided report from the respondents' side.An additional possible limitation is that the wording of the initial scale was used 'as is', without adapting it to the South African (multilingual) context.There is also a limitation associated with the use of a cross-sectional design, which might have increased the relationship between the nine components artificially.
It is recommended that a qualitative approach from an emic perspective be followed to further explore the contextual realities.A qualitative study can explore how the HRPPQ constructs can inductively be integrated with collectivism and ubuntu, thereby validating and complementing the current study through triangulation.The integration of both etic (outside) and emic (inside) approaches will result in richer findings (Morris, Leung, Ames, & Lickel, 1999), as the two methods are complementary and contradictory and will raise more questions while stimulating further inquiry.The inquiry will further explore whether South Africa is predominately a collectivistic culture or perhaps a hybrid of both individualistic and collectivistic cultures.

Conclusion
This study suggested that the collectivistic culture in the South African work context has an impact on the perceptions of HR practices, specifically in terms of performance management.A further interesting finding is the separation of the training and career development practices, which might resonate with the notion of talent management (attraction, retention, engagement and the development of employees).The study further contributed in terms of a validated factor structure, which can be used for future academic research to determine the relationship between perceived HR practices and employee and behavioural variables.It provides an accurate measurement for HR managers and practitioners to measure perceptions and to develop strategies and interventions to enhance them to the benefit, not only of the organisation in terms of performance, but of employee satisfaction, well-being and ultimately productivity.
1865-1829), which is distributed as chisquare with 571-514 = 57 degrees of freedom.The best-fitting model is thus the first-order model (model b ) (χ 2/df [571] = 3.20, CFI = 0.949, RMSEA = 0.042), with the 37 items loading directly on their respective factors, which are in this case the factors (i.e.job design, participation and autonomy, recruitment and selection, training, development, employment security, teamwork, rewards and work-life balance).

TABLE 2 :
Results of the Monte Carlo parallel analysis.

TABLE 1 :
Eigenvalues larger than one and explanation of variance.
Boon et al. (2011)ethod: principal component analysis.Their first factor, originally called participation, autonomy and job design (consisting of nine items) was divided into two factors, namely job design (three items) and participation/ autonomy (consisting of six items).A further adaptation is that the factor concerning training and development was also split into two separate factors, namely training (consisting of two items) and development with seven items.The employment security as well as work-life balance factors remained the same periodic evaluation of my performance'.This item was initially part of the performance appraisal and rewards factor together with Item 28, which reads, 'The organisation offers (me) fair appraisal of my performance'.Item 28 was the only item that was excluded from the new structural configuration of the HRPPQ.The initial factor ofBoon et al. (2011)called performance appraisal and

TABLE 3 :
Boon et al. (2011)romax rotation) and descriptive statistics of items on the Human Resource Practices Perceptions Questionnaire.# , The item numbers in the original instrument byBoon et al. (2011).HRPPQ, Human Resource Practices Perceptions Questionnaire; SD, standard deviation; h, communalities.

TABLE 4 :
Correlations between the extracted factors (adapted structural configuration of the Human Resource Practices Perceptions Questionnaire).

TABLE 6 :
Comparison of an a priori nine-factor structure (the adapted factor structure of the Human Resource Practices Perceptions Questionnaire).All chi-square values are significant at p < 0.001; the Δχ 2 is in relation to one-factor model.CFI, comparative fit index; RMSEA, root-mean-square error of approximation.†, Unidimensional model.‡, Model with separate factors.§, Model with factors that load onto a higher order construct.¶, Thirty-seven items as determined by the exploratory factor analysis; Note: *, p < 0.01 (two-tailed).

TABLE 5 :
Descriptive statistics, Cronbach's alpha coefficient of the factors of the Human Resource Practices Perceptions Questionnaire (adapted structural configuration).
and development was split into two separate factors: training and development, with the latter being more concerned with the typical career development practices such as coaching for development, career planning, job rotation, job enlargement and job enrichment, and vertical promotion.
The adjusted factor structure consists of nine factors or HR practices with 37 of the original 38 items.The factor originally called participation, autonomy and job design was divided into two factors: job design and participation and autonomy.A further adaptation is that the factor concerning training

TABLE 7 :
Convergent and discriminant validity of the HRPPQ (adapted structural configuration)