Factors of Climate

Organizational climate and job satisfaction

The organizational climate questionnaire was used to measure the climate of the organization. The descriptive statistics, with specific reference to mean, standard deviation, skewness and kurtosis, are presented in Table 4 and were computed for the various dimensions assessed by the questionnaire. In addition, the table provides Cronbach's alpha coefficients for each dimension and the total scale. All the Cronbach alpha scores are above 0.8 with only work environment below 0.6. According to Nunnally, a suitable criterion for instruments in the early stages of development is regarded as between 0.5 and 0.6, although for established scales it would typically be about 0.7. It was thus decided to include all dimensions in the further analysis.

Table 4: Descriptive statistics and reliability of scales of the organizational climate questionnaire

The mean scores were used to summarize the climate in the organization. For the purposes of this study, the recommended cut-off score of 3.2 (on a scale of 1–5, strongly disagree to strongly agree) was used to differentiate between potential positive and negative perceptions, with scores above 3.2 indicating a positive perception and scores below 3.2 indicating a negative perception of that dimension. Research by the HSRC indicates that an average of 3.2 is a good guideline to distinguish between positive and potential negative perceptions.

From Table 3, it is clear that the climate facet means in the organization can be defined as positive, with a mean score of 3.56 across facets. The results indicated that the employees were mostly satisfied with the various aspects of the organization measured by the 11 climate dimensions and the job satisfaction dimension, because all except two dimensions yielded means over the 3.2 cut-off point. The results of the training and development (3.11) and remuneration and reward (2.77) dimensions indicated that employees perceived training and development opportunities in the organization negatively and also considered remuneration and reward practices to be negative (all below the 3.2 cut-off point). Hence, these dimensions could be considered as future areas of development for the organization. The job satisfaction of the organizational members was also measured. This was achieved by employing the global approach, whereby certain questions were asked to elicit affective responses about the employees' jobs. The results indicated that the respondents were satisfied with their jobs (3.36), found their work interesting and challenging and perceived their future in the organization positively.

Table 4 also provides the reliability statistics of the scales of the questionnaire. The alpha coefficients of 11 of the dimensions ranged from 0.81 to 0.89, indicating internal consistencies within the recommended range. The work environment scale, however, appeared to have an unacceptably low reliability (0.59). This alpha value suggests that the items in this scale did not correlate strongly with other items in this scale or with the total work environment scale and consideration should be given to including additional items to or removing some items from this dimension.

A confirmatory factor analysis, using AMOS 7.0, was performed to investigate whether the evidence supported a model of the original 12-factor dimensions of the questionnaire. The results illustrated that most of the indices were wide of their respective recommended values, thereby indicating a lack of fit for the 12-factor model in question. The fit indices used in this analysis included the comparative fit index (CFI), Tucker Lewis index (TLI), parsimony normed fit index (PNFI), parsimony comparative fit index (PCFI) and root mean square of approximation (RMSEA). The chi-square statistic (χ²) (CMIN), that is the comparison between the observed covariance to the hypothesised covariance or model fit computed for the 12-factor model, was 6197.624. The goodness of fit indices for the CFI and TLI were 0.603 and 0.58, respectively, which were both far below 0.9, the acceptable level of fit (Hu & Bentler, 1999). The fit indices of the PNFI and PCFI, 0.5 and 0.57, respectively, were also much lower than the acceptable 0.9. Steiger's RMSEA of 0.095 fell well beyond the acceptable level of 0.06. Hypothesis 1 was therefore not supported.

Because of the lack of fit obtained with the 12 dimensions of the original organizational climate questionnaire, an exploratory factor analysis (EFA) was conducted to investigate the underlying factor structure of the questionnaire. It was expected that these 12 dimensions would be distinct from one another and that the job satisfaction scale would be loaded independently from the other 11 dimensions of the questionnaire. The EFA was conducted using the principal axis factoring technique with an oblique promax rotation. The Kaiser criterion, which specifies that only factors with eigenvalues of 1.00 or greater should be retained and the so-called 'scree test' were used as a guide to determine the number of factors to extract. Using Kaiser's criterion, 12 factors were extracted, explaining 60.23% of the total variance. However, since the 12th factor accounted for only 1.35% of the total scale variance and had only one item with a factor loading higher than 0.3, it was decided to retain 11 factors only. The scree test identified three factors, explaining 44.21% of the total variance.

On the basis of the strength of the Kaiser criterion and the scree test results, it was decided to conduct a confirmatory factor analysis (CFA) to compare these models with the fit indices revealed by the original assessment to indicate the best fit. The following fit indices were used: normed chi-square adjustment (χ2/df), CFI, TLI, PNFI, PCFI and RMSEA.

Table 5 indicates that the three-factor models (hierarchical and oblique) and the original 12-factor model, although a better fit, did not fit the data well. The normed chi-square adjustment for both models was above the recommended ratio of 2.0 for good fit (Tabachnick & Fiddell, 2001). The CFI and TLI yielded poor fit with values lying below the acceptable level of 0.95 (Hu & Bentler, 1999) and the PNFI and PCFI also reported values lower than the acceptable level of 0.90. The RMSEA values for both models were higher than the 0.05, value but within the acceptable level of less than 0.08. The revised 12-factor model (after the CFA) ( 11 dimensions) fitted the data best: χ² = 5264.65, CFI = 0.82, TLI = 0.80 and RMSEA = 0.06.

Table 5: Fit indices for the comparative models of the organizational climate questionnaire

Table 6: Correlation between organizational climate dimensions and job satisfaction

The researchers therefore decided to proceed with the revised 12-factor model (11 dimensions) for further analysis. After inspecting the items that loaded meaningfully, the 11 dimensions were labelled (with alpha values in brackets): leadership of immediate manager (0.97), transformation and diversity (0.85), personal growth and development (0.87), interpersonal belonging and fit (0.87), general feeling of job satisfaction (0.89), employee wellness (0.85), image (0.84), pay (0.85), challenging and interesting work (0.85), physical work environment (0.59) and recognition and acknowledgment (0.82). The Cronbach alpha coefficient for all 11 dimensions was found to be satisfactory, all above 0.8.

One of the objectives of this study was to determine whether there was a strong positive relationship between organizational climate and job satisfaction. Pearson's product–moment correlation was used for this analysis. The findings of this research indicated that there was a strong positive correlation between the two variables (r = 0.813, n = 696, p = 0.000). In terms of the practical significance guidelines provided by Cohen (1988), the relationship between the total organizational climate variable and total satisfaction variable is a large effect size. Hypothesis 2 was therefore supported.

The Pearson product–moment correlation coefficients were also computed among the 11 dimensions. During the analysis of the items and factors, it was found that some dimensions had a personal or direct influence on the individual and others an indirect influence on the individual. The dimensions considered to have a personal influence on the individual included the following:

• personal growth and development
• interpersonal belonging and fit
• employee wellness
• challenging and interesting work.

The dimensions considered to have an indirect influence on the individual included the following:

• leadership of immediate manager
• transformation and diversity
• image
• pay
• physical work environment
• recognition and acknowledgment.

The results reported in Table 7 indicate that organizational climate dimensions that were grouped together as personal factors or factors with a direct influence on the individual as well as organizational climate dimensions that were grouped together as being external to or influencing the individual indirectly were positively and statistically related (at the 0.01 level) to a general feeling of job satisfaction. In terms of the practical significance guidelines, these variables yielded effect sizes ranging from medium to large.

Table 7: Model summary+ of explained variance in job satisfaction

Table 8: Coefficients for the independent variables of model 9

Table 9: Total variance explained for the 12-factor model after extraction (excluding factor with eigenvalues lower than 1)

The hypothesis that organizational climate dimensions that are perceived to be personal to or have a direct impact on the individual will have a greater influence on job satisfaction than organizational climate factors that are perceived as being external to or influencing the individual indirectly was investigated by means of the calculation of a step wise linear regression. As indicated in Table 9, using the step wise estimation technique, nine variables (model 9) predicted 70.9% of variance in job satisfaction. All of the models, including model 9, were statistically significant. The reasons for using the step wise regression estimation technique were that as an exploratory technique it is seen as a model-building rather than a model-testing procedure and it was also used to support the previous analysis such as the correlation and CFA. Standard multiple regression analyses were conducted because our interest was not in finding prediction equations for predicting the coping styles of our sample. Rather, our interest was in assessing the magnitude of the correlations between the dependent and independent variables and in assessing the magnitude of the overall relationship between the dimensions and the independent variables. Standard multiple regression also enabled us to assess how much each independent variable uniquely contributed to the overall relationship because the independent variables were evaluated in terms of how much they added to the prediction of the coping styles, which differed from the percentage of variance accounted for by the combination of all the other independent variables.

Table 9 presents all the variables included in model 9 and their contribution to predicting job satisfaction. The beta value provides information on the contribution of each independent variable. The largest value contributes the most.

From Table 8, it is clear that an organizational climate dimension perceived as personal to the individual (personal growth and development) has the largest beta coefficient of 0.224, therefore contributing the most to explaining job satisfaction when the variance explained by all the other variables in the model is controlled. organizational climate factors that are perceived as external to the individual, for example leadership of the immediate manager, image and pay, also contribute strongly to explaining job satisfaction. On the basis of the results in Table 8, hypothesis 3 was therefore partially supported.