Please enable JavaScript.

Coggle requires JavaScript to display documents.

Inferential Statistics: the statistical procedures that are used to reach…

- - - - Probability of the result.
      - How close the estimate is likely to be to the true population value.
- - - - Two-tailed → any difference.
      - One-tailed → directional hypothesis (e.g., treatment > control).
  - - - β₁ = effect size.
      - With dummy variable (0/1):
        
        Effect size (β₁) = mean difference (treatment – control).
- - - - Independent effect of each factor (β₁ and β₂).
  - - - Treatment effect depends on instructional time.
      - Main effects cannot be interpreted alone.
    - - Main effects (β₁ and β₂) are meaningful and interpretable.
- - - - Purpose: Classification
      - Function: Assign observations to nominal categories using linear combination of predictors
      - Use: Marketing, customer/product classification
      - Note: Similar to multiple regression but with nominal dependent variable
      - 3. Logistic Regression (Logit Model)
        
        Purpose: Predict binary outcomes (0 or 1)
        
        Distribution: Logistic
        
        Use: Medical sciences (e.g., probability of heart attack)
        
        Effect Size: Odds ratio
        
        4. Probit Regression (Probit Model)
        
        Purpose: Predict probability of outcome (0–1 or discrete 0/1)
        
        Distribution: Standard normal
        
        Use: Actuarial science, finance, insurance, credit scoring
        
        Note: Similar to logistic regression, but logit easier to compute/interpre
        
        5. Path Analysis
        
        Purpose: Examine directional relationships among multiple variables
        
        Function: Dependent variable in one equation can be independent in another
        
        Use: Social science research, testing complex nomological models
        
        6. Time Series Analysis
        
        Purpose: Analyze variables over time
        
        Use: Forecasting stock markets, urban crime rates
        
        Special Consideration: Correct for autocorrelation