Konsep Variabel Intervening (Mediating)
📌 Variabel intervening (mediating) acts as an intermediary in the relationship between an independent variable (IV) and a dependent variable (DV).
👩‍🏫 It is crucial not to confuse intervening variables with moderating variables.
📝 Researchers introduce mediating variables when established theories predict an IV influences a DV, but recent empirical research shows a weak or non-significant direct effect (a research gap).

Syarat dan Model Penelitian
📌 A conceptual requirement for an intervening variable is that it must theoretically function as both a dependent variable (influenced by the IV) and an independent variable (influencing the DV).
📊 Common models include one IV, one mediator (M), and one DV (Model 1), or multiple IVs influencing one mediator, which then influences the DV.
🧪 The analysis involves two main regression models: Model 1 (IVs predicting the mediator) and Model 2 (IVs + Mediator predicting the DV).

Analisis Regresi dan Interpretasi Beta Coefficients (Menggunakan SPSS)
⚙️ Regression Model 1: Used to determine the significance and strength of the influence of IVs ($\text{X}_1, \text{X}_2$) on the mediator ($\text{M}$). Significance is checked via the $\text{p-value} < 0.05$.
⚙️ Regression Model 2: Used to determine the significance and strength of the direct effect of IVs ($\text{X}_1, \text{X}_2$) on the DV ($\text{Z}$), and the effect of the mediator ($\text{M}$) on the DV ($\text{Z}$).
💡 The significance of the indirect effect is calculated by multiplying the standardized beta coefficients: $\beta \text{ (Indirect)} = \beta (\text{IV} \to \text{M}) \times \beta (\text{M} \to \text{Z})$.

Jenis Mediasi dan Uji Sobel
🔍 Pure Mediating (Perfect Mediation): Occurs when the direct effect ($\text{X}_1 \to \text{Z}$) is not significant, but the indirect effect ($\text{X}_1 \to \text{M} \to \text{Z}$) is significant.
⚠️ Partial/Sham Mediating (Mediasi Semu): Occurs when both the direct effect ($\text{X}_2 \to \text{Z}$) and the indirect effect ($\text{X}_2 \to \text{M} \to \text{Z}$) are significant.
🔬 The Sobel Test is used to formally test the significance of the indirect effect using online calculators, requiring four inputs: the regression coefficients (B) and the standard errors for both paths ($\text{IV} \to \text{M}$ and $\text{M} \to \text{DV}$).

Key Points & Insights
➡️ Researchers often introduce mediating variables to reconcile theoretical expectations with weak empirical findings between an IV and DV.
➡️ The initial step in testing mediation requires running two separate linear regressions in SPSS using standardized coefficients (Beta) for comparison.
➡️ For the $\text{X}_1$ pathway, the direct effect was non-significant, but the indirect effect was significant, indicating pure mediation.
➡️ The Sobel Test result of $p=0.086$ (for the first path) suggests that formally, the mediation was not statistically significant at the 5% level ($\alpha=0.05$), suggesting the result from the pathway multiplication method ($p < 0.05$) might be less precise.
➡️ Actionable Task: Students are assigned to create a research model with three independent variables and one mediating variable, test it using any data set with at least 30 units of analysis, and perform the analysis.

📸 Video summarized with SummaryTube.com on Nov 26, 2025, 12:43 UTC