**4.3 Reliability and validity of approach to improving OHS**

**Table 5** illustrates that the Cronbach coefficient is greater than 0.6, therefore it is necessary to measure the adequacy level of the reliability to test the causal effect of the relation. According to Ursachi et al. and Nunnally and Bernstein [42, 43], the *Approaches to Improving Occupational Health and Safety of the Nigerian Construction Industry DOI: http://dx.doi.org/10.5772/intechopen.113011*

*Initial model of the approaches to improving OHS.*

**Figure 3.** *Final model of the approaches to improving OHS.*

generally accepted standard is Cronbach's alpha of 0.6–0.7 which shows a suitable level of reliability. Also, 0.8 or higher is a good level. But, values greater than 0.95 are not certainly good, as they can be seen as a sign of redundancy [44]. Similarly, the composite reliability in all cases exceeded the least level of 0.6 as suggested by Bagozzi and Yi [45]. While the average variance explained was equal to 0.5 according to Awang [46], however, Bagozzi [47] stated that a value of 0.3 and above could be


#### **Table 3.**

*Goodness of fit of the model of approaches to improving OHS.*


#### **Table 4.**

*Standardized regression weight of the factors of approaches to improving OHS.*

considered. **Figure 2** indicates that the model fit is acceptable. Also, the standardized regression weights of the success factors were higher than 0.5 and significant at the 95% confidence level, confirming the convergent validity of the constructs as illustrated in **Table 5**. The validity and reliability of the approaches to improving occupational health and safety of the Nigerian construction industry had been demonstrated and structured in 31 variables at eight dimensions.
