Tutorial SEM dengan SmartPLS 4 (Pemula)

Summary of the Subtitles (SmartPLS 4 Beginner Tutorial)

• The video explains how to perform PLS-SEM analysis in SmartPLS 4 from start to finish, aimed at beginners.
• It demonstrates working with questionnaire data (reported as 139 questionnaires) containing multiple items/indicators grouped as A1–A9, B1–B7, and C1–C6.

---

Workflow: Importing Data → Building the Model → Running Analysis

1. Prepare data in Excel

   • Questionnaire response data is moved into Excel.
   • The file is saved as CSV (via Open CSV / workbook → CSV).

2. Create/load a SmartPLS 4 workspace and project

   • SmartPLS 4 is opened (noting it differs from PS3/SmartPLS 3).
   • A new workspace and new folder should be created (to avoid using an invalid workspace folder).
   • A new project is created (example name: “sam/sam1”).
   • The CSV file is imported into the project.

3. Define the PLS-SEM model

   • A model is created and set to PLS-Mode.
   • Latent variables are added (example: A, B, C).
   • Indicators/items are connected to each latent variable:
     • A: A1–A9
     • B:
       • B1–B3 → linked to b11
       • B4–B7 → linked to b22
     • C: C1–C6

4. Draw relationships (paths)

   • The model includes paths such as:
     • A → B11, A → B22
     • C → B11, C → B22
   • Indicator placement/layout differs from SmartPLS 3, so alignment options (left/right) may be needed.

---

Running the PLS Algorithm and Interpreting Outputs

• The tutorial runs the procedure via:
  • Calculate → PLS Algorithm → Start Calculation

Outer model: Validity & Reliability

The outputs are used to check:

• Convergent Validity

  • From outer loadings / loading factor

• Discriminant Validity

  • Via cross-loading (described as a “not determinant” / cross-loading check)

• Reliability

  • Cronbach’s Alpha
  • Composite Reliability
  • Average Variance Extracted (AVE) (mentioned as “eva”)

Inner model: Model fit / explained variance

Additional metrics include:

• R Square (to assess inner model effectiveness)
• Effect sizes / GOF are mentioned as additional measures

---

Hypothesis Testing (Bootstrapping)

• The video covers hypothesis testing (inner model / path significance) using bootstrapping.
• Bootstrapping takes time.
• Interpretation uses path coefficient significance, with colors described as:
  • Red path = hypothesis rejected
  • Green path = hypothesis accepted
• Outer loadings from bootstrapping are also shown to inspect indicator-to-construct relationships.

Example hypothesis outcomes (as stated in subtitles)

• 2 hypotheses accepted
• 2 hypotheses rejected
• A specific case is mentioned (e.g., CGB2 / H0 accepted, and A→B11 rejected), though the exact labeling is slightly unclear due to subtitle noise.

---

Moderation Analysis (SmartPLS 4 vs. SmartPLS 3)

• The tutorial explains adding moderation variables in SmartPLS 4:
  • A moderation effect is selected (workflow differs from SmartPLS 3).
  • A moderation line is drawn from the moderator to the relationship (connecting in the “middle”).
  • Example: using C as a moderation variable between A and B (affecting relationships to B11 and B22).

Moderation results described

• Without moderation, paths are accepted/rejected similarly to the baseline model.
• With moderation, effects change:
  • Moderation strengthens one relationship (described as accepted)
  • Moderation weakens / becomes not significant for another relationship (described as rejected for one path)
• Overall summary from the subtitles:
  • Moderation on B11 appears stronger/accepted
  • Moderation on B22 appears weaker/not significant

---

Reusing/Adding New Data and Re-running Analysis

• The tutorial shows how to add new data (or switch datasets):
  • SmartPLS 4 is described as easier than SmartPLS 3 for managing multiple datasets.
  • Instead of only one dataset, SmartPLS 4 allows choosing among multiple datasets.

Steps described:

• Select a dataset (example: Data 3)
• Adjust/replace data inputs
• Run PLS Algorithm again
• Re-check validity/reliability results (statuses/colors such as rejected vs accepted may change)

---

Main Speaker / Source

• Single tutorial presenter (name not identified) demonstrating the SmartPLS 4 steps throughout.