Parametric Statistics requires the assumption that quantitative data exhibits normal distribution

  • parametric statistic is a stronger way of testing hypothesis than non-parametric statistic

quantitative data can be either:

  • normal
  • right skewed
  • left skewed

if data does not exhibit normal distribution, we can either:

  • use a mathematical transformation on the data that leads to normal distribution
  • use another method that does not require normal distribution
    • rank-based method
    • non-parametric testing

3 commonly used transformations for quantitative data:

  • logarithm
  • square root
  • reciprocal

we call these transformations variance-stabilizing because, their purpose is to make variances the same

transform values of a variable to take on normal distribution

https://machinelearningmastery.com/how-to-transform-data-to-fit-the-normal-distribution/

plot data as histogram and see what transformation is needed