PhD Thesis - Chapter 4

for a PhD thesis with quantitative mediation analysis, Findings and Data Analysis should present the results in a logical sequence, moving from descriptive statistics to measurement assessment, direct effects, mediation effects, and hypothesis testing... the writing should be analytical and objective, focusing on what the data show rather than interpreting the theoretical implications in depth (which is usually reserved for chapter 5)... a typical paragraph structure is as follows:

Introduction to the Findings Chapter

begin the chapter by briefly explaining what will be presented...

Example:

this chapter presents the findings of the study and the results of the quantitative data analysis... the analysis was conducted using structural equation modelling (SEM-PLS)... the chapter begins with respondents' demographic profiles, followed by descriptive statistics, assessment of the measurement model, evaluation of the structural model, and testing of the mediation hypotheses... the findings are presented according to the research objectives and hypotheses formulated in Chapter 1...

Descriptive Statistics

present the mean, standard deviation, minimum and maximum values...

Example:

table 4.5 presents the descriptive statistics for all study variables... the results indicate that Employee Engagement recorded the highest mean score (M = 4.12, SD = 0.65), suggesting that respondents generally reported a high level of engagement... Organizational Support had a mean score of 3.89 (SD = 0.71), while Job Satisfaction recorded a mean score of 3.95 (SD = 0.68)... the relatively low standard deviation values indicate limited variability among respondents' responses...

Measurement Model Assessment

discuss reliability and validity...

Example of Reliability

internal consistency reliability was assessed using Cronbach's alpha and composite reliability (CR)... as shown in Table 4.6, all constructs exceeded the recommended threshold of 0.70... Organizational Support achieved a Cronbach's alpha of 0.891 and a composite reliability value of 0.918... similarly, Employee Engagement and Job Satisfaction reported composite reliability values of 0.925 and 0.913, respectively... these findings indicate satisfactory internal consistency reliability for all constructs...

Example of Convergent Validity:

convergent validity was evaluated using average variance extracted (AVE)... The AVE values ranged from 0.623 to 0.748, exceeding the recommended value of 0.50... this result confirms that the indicators adequately represent their respective latent constructs and explains more than 50% of the variance in the observed measures...

Example of Discriminant Validity

discriminant validity was assessed using the Heterotrait-Monotrait (HTMT) ratio... all HTMT values were below the threshold of 0.85, indicating satisfactory discriminant validity... therefore, the constructs were empirically distinct from one another and suitable for subsequent structural model evaluation...

Structural Model Assessment

report collinearity, R², effect sizes, and path coefficients...

Example:

prior to hypothesis testing, collinearity issues were assessed using variance inflation factor (VIF) values... the results indicated that all VIF values were below 3.3, suggesting that multicollinearity was not a concern... the explanatory power of the model was evaluated using the coefficient of determination (R²)... the results revealed that Job Satisfaction explained 58.4% of the variance in Employee Performance (R² = 0.584), indicating moderate explanatory power...

Direct Effects Analysis

present path coefficients and hypothesis results...

Example:

the direct relationship between Organizational Support and Employee Performance was examined through bootstrapping analysis... the findings indicate a significant positive relationship (β = 0.372, t = 5.412, p < 0.001)... therefore, H1 is supported, suggesting that higher levels of Organizational Support are associated with increased Employee Performance...

Mediation Analysis (Most Important Section)

for mediation, report:

• Path a (IV → Mediator)
• Path b (Mediator → DV)
• Indirect effect (a × b)
• Direct effect after mediation
• Type of mediation

Example of Full Mediation Paragraph

the mediating effect of Employee Engagement on the relationship between Organizational Support and Employee Performance was examined using the bootstrapping procedure with 5,000 resamples... the results show that Organizational Support significantly predicted Employee Engagement (β = 0.614, t = 11.276, p < 0.001), while Employee Engagement significantly predicted Employee Performance (β = 0.421, t = 6.835, p < 0.001)... the indirect effect was significant (β = 0.259, t = 5.742, p < 0.001), with the 95% confidence interval not containing zero (LL = 0.173; UL = 0.347)... however, the direct effect of Organizational Support on Employee Performance became non-significant after the inclusion of Employee Engagement (β = 0.086, t = 1.412, p = 0.158)... these findings indicate that Employee Engagement fully mediates the relationship between Organizational Support and Employee Performance... therefore, H4 is supported...

Example Partial Mediation Paragraph

the mediation analysis revealed a significant indirect effect of Organizational Support on Employee Performance through Employee Engagement (β = 0.213, t = 4.987, p < 0.001)... the direct effect remained significant after the mediator was included in the model (β = 0.242, t = 3.724, p < 0.001)... since both the direct and indirect effects were significant, Employee Engagement partially mediates the relationship between Organizational Support and Employee Performance... thus, the mediation hypothesis is supported...

Summary of Hypothesis Testing

Example:

table 4.15 summarizes the hypothesis testing results... of the six hypotheses proposed, five were supported and one was not supported... Organizational Support significantly influenced Employee Engagement and Employee Performance... furthermore, Employee Engagement significantly mediated the relationship between Organizational Support and Employee Performance, providing empirical support for the proposed mediation model...

Recommended PhD Writing Formula for Mediation Results

for every mediation hypothesis, follow this paragraph structure:

the mediating effect of M on the relationship between X and Y was assessed using bootstrapping... the results showed that X significantly affected M (β = ..., t = ..., p = ...)... in addition, M significantly affected Y (β = ..., t = ..., p = ...)... the indirect effect was significant (β = ..., t = ..., p = ..., 95% CI [...] )... the direct effect of X on Y remained significant/non-significant (β = ..., t = ..., p = ...)... therefore, M partially/fully mediates the relationship between X and Y... hence, Hypothesis Hx is supported...

this is the standard paragraph style commonly used in PhD theses employing SEM-PLS, AMOS, SmartPLS, PROCESS Macro, or CB-SEM mediation analysis...