Analysis of Ability to Deal with Work Demands - Paper Example

Introduction

For analysis purposes, SPSS was used. For purposes of confirming the proposed structural model, partial least square variance-based structural equation modeling (PLS-SEM), an exploratory technique, which uses data in testing the relationship between latent variables to establish path relationship within models, was used. PLS-SEM is generally renowned for its significant contribution in social studies, as it assumes a vital role in the comprehension of the relationship among a set of observed variables. It additionally works effectively with relatively smaller sample sizes and complex models (Hair, Sarstedt, Ringle & Gudergan, 2017). PLS-SEM is additionally more suitable in cases where theory is less established. It is for the most part used in proposing theories and hypotheses in exploratory research. In structural equation modeling, reflective or formative are two forms of measurement scales. When the predictors result in a latent variable and are not tradable, they are classed as formative. However, when the markers are profoundly correlated and tradable, they are considered as reflective, and their validity and reliability must be carefully investigated. In a reflective scale of measurement, the causality direction shifts to the yellow color indicator from the blue-color latent variable. Because the majority of the indicators in this investigation are of a reflective type, the thoughtful analysis was applied.

Is your time best spent reading someone else’s essay? Get a 100% original essay FROM A CERTIFIED WRITER!

Save 25%

A comparison of methodologies for the investigation of interaction effects among latent variables based on PLS-SEM was conducted. In the assessment of SEM, the first phase involves the checking of the outcomes of the measurement model followed by those of the structural model for purposes of validation. Suitability of estimation models was estimated through the loadings of the items with regards to the latent variables, the Average Variance Extracted (AVE), the Discriminant Validity and the Composite Reliability Score. From the process, loading of over 0.70 was thought to be high, while loadings of a range between 0.40 and 0.70 were acceptable when elimination of the indicators did not bring about an increment in the model's reliability.

The AVE and Composite Reliability Score were viewed as optimum when above 0.50 and 0.70, respectively. The discriminant validity was considered when the loading of the items of the indicator variables is relatively above than the loading of similar indicators on other latent variables. Considering the above, all variables in the current study were incorporated into the PLS-SEM examination as explanatory variables. Only latent variables, indicators, and the path that achieved a 0.05 level of significance were engaged in the measurement model following Bootstrapping analysis. Additionally, pointers were considered only when the loading of the indicators was above 0.7, had internal reliability of over 0.7 and AVE of over 0.5. In the analysis, the discriminant validity was confirmed as well.

The model investigated the interrelationships among variables on work demands (within three constructs of "workability", "physical demands," "mental demands") through the deployment of the modeling approach based on PLS structural equation using SmartPLS modeling software. Moreover, the assessment of the structural model incorporated the significance and level of the path coefficients through the performance of bootstrapping procedure. Bootstrapping is a measure or test or measure that is founded on random sampling but with replacement. The strategy allows the assessment of the sample distribution using random sampling approaches. It is a comparatively simple approach to extraction of estimates of confidence intervals and standard errors of multiple parameters of the sample distribution. It is additionally an appropriate method for checking and controlling and checking the reliability of outcomes.

Measurement Model

Before the analysis of the structural model, validity and reliability were assessed and established. Later, discriminant validity and average variance extract were evaluated to gauge the outer loadings of the reflective measurement models. Figures 1, 2 and 3, shows the assessment criteria for the model. Analysis of the latent variables' outer loadings indicates that loadings were in the range of 0.48 and 0.93. Since the removal of predictors below 0.7 did not change general reliability, the indicators were retained in the model. Scores of composite reliability and AVE indicated an excellent validity and reliability (Hair, 2014). Ultimately, Table 4 illustrates the discriminant validity that was established from the model. In the matrix within the table, the off-diagonal values represent the correlations between constructs and latent variables. Accordingly, the results indicate that there is discriminant validity among all constructs in the evaluation based on the criterion of cross loading. The results from PLS-SEM to establish the correlation between "work ability" and "physical demands" variables.

Structural Model

After the confirmation of the construct measurements as valid and reliable, the structural model was considered appropriate for the assessment of the predictive abilities of the model, as well as the relationship between its components. The outcomes affirmed that the structural model, as well as all beta paths, was statistically significant with the p-value of 0.05.

Target Endogenous Variance

In the circles, the numbers illustrate the degree of variance of the latent variable is explained through the latent variables. In this analysis, R2, the coefficient of determination was 0.677 for the "workability:, an endogenous latent variable that demonstrated that the two latent variables, "mental demands" and "physical demand," explain moderately 67.7% of the fluctuation. Also, the outcomes demonstrate a sufficient prescient legitimacy of the examination's model.

Significance and Coefficient Sizes of Inner Model Path

The path coefficients showed the strength of a variable on other variables. Besides, the weight of path coefficients enables the ranking of the statistical importance of variables. Based on such, the internal model of the study explained that "physical demands" has a powerful effect on "workability". Moreover, "physical demands" and "mental demands" have a strong effect (0.513 and 0.469, respectively) on collaborating variable that impacts indirectly on "work abilities." The internal model additionally implied that "physical demands" has a significant impact on "mental demands" component. Besides, the path relationships, initially hypothesized between variables, were considered to be statistically significant, since all the coefficients of standard path appeared to be above 0.1. Accordingly, it can be concluded that "physical demands" directly and mental demands" indirectly are reasonably strong predictors of "workability" variable.

References

Hair, J. F. (2014). A primer on partial least squares structural equation modeling (PLS-SEM). Thousand Oaks: SAGE Publications.

Hair, J. F., Sarstedt, M., Ringle, C. M., & Gudergan, S. (2017). Advanced issues in partial least squares structural equation modeling. Los Angeles: SAGE.