Institutional autonomy, academic quality, and research productivity in African public universities: examining the mediating role of institutional operations

Structural Path Coefficients and Hypothesis Test Results
Hypothesis Path Beta SE t-value p-value 95% CI Result
H1 Autonomy >> Research Productivity 0.341 0.053 6.43 < .001 [0.24, 0.45] Supported
H2a Autonomy >> Inst. Operations 0.749 0.025 29.96 < .001 [0.70, 0.80] Supported
H2b Inst. Operations >> Research Productivity 0.243 0.063 3.86 < .001 [0.12, 0.37] Supported
H3a Autonomy >> Academic Programmes 0.748 0.029 25.79 < .001 [0.69, 0.80] Supported
H3b Acad. Programmes >> Grad. Employability 0.179 0.059 3.03 .002 [0.06, 0.30] Supported
H4 EF x Autonomy >> Research Productivity -0.144 0.058 2.48 .013 [-0.26, -0.03] Supported
Control Governance >> Autonomy 0.748 0.027 27.70 < .001 [0.70, 0.80] Control path
Note. SE = standard error. CIs are bias-corrected bootstrap intervals (10,000 resamples). EF x Autonomy = interaction term between external factors composites and autonomy latent variable scores.
Source: Field Data

Table 4 presents the structural path estimates and hypothesis testing outcomes from the PLS-SEM analysis. The results indicate a robust and statistically significant model, with all hypothesised relationships supported at conventional significance levels.
Institutional Autonomy has a significant positive effect on Research Productivity (β = 0.341, p < .001), confirming H1. It also strongly influences Institutional Operations (β = 0.749, p < .001) and Academic Programmes (β = 0.748, p < .001), supporting H2a and H3a and underscoring its central role in shaping internal institutional processes.
Institutional Operations significantly predict Research Productivity (β = 0.243, p < .001), supporting H2b and indicating an important indirect pathway linking autonomy to research outcomes. Similarly, Academic Programmes significantly affect Graduate Employability (β = 0.179, p = .002), supporting H3b and highlighting the relevance of programme design to labour market outcomes.
The moderating effect of External Factors on the autonomy–research productivity relationship is negative and significant (β = −0.144, p = .013), supporting H4 and suggesting that external constraints weaken the beneficial impact of autonomy on research performance.
Institutional Governance significantly predicts Autonomy (β = 0.748, p < .001), confirming its role as a structural antecedent.