Goss-Sampson (2020), Statistical Analysis in JASP - pp. 34-37

Question 1

Data Transformation

Answer 1

Data transformation involves modifying data to:
- Improve normality for parametric analyses.
- Address violations of assumptions (e.g., skewness, kurtosis).
- Simplify computations or prepare data for analysis.

Question 2

Data Transformation:
Key Applications:

Answer 2

• Normalizing Data:
• Reduces skewness for better adherence to statistical test assumptions.
• Creating New Variables:
• Calculate differences (e.g., pre- and post-treatment scores).
• Apply formulas or transformations (e.g., logarithmic transformations).

Question 3

Common Transformations

Answer 3

Logarithmic Transformation (log10):
- Reduces skewness in positively skewed data.
- Applied as log10 (y), where y is the variable to transform.
- Useful for data spanning several orders of magnitude.

Square Root Transformation:
- Addresses moderate skewness.
- Applied to positively skewed continuous data.

Difference Computation:
- Subtract one variable from another (e.g., pre- and post-treatment scores).
- Example: Difference=Variable 1−Variable 2.

Other Built-in Operators:
- Arithmetic operations (e.g., addition, subtraction).
- Exponential or logarithmic functions.

Question 4

Visual Comparison: Transformed vs. Raw Data

Answer 4

• Graphs of raw vs. transformed data demonstrate changes in distribution:
• Untransformed Data: Skewed or uneven distribution.
• Transformed Data: More symmetrical, approximating normal distribution.