Evaluating Normal Approximation

Normal Approximation Evaluation

• Assess the normality of variables in datasets graphically.

Epo Trial Overview

• Study objective: Compare erythropoietin effects and hemoglobin transfusion thresholds on neurological recovery.

• Design: Randomized clinical trial with 200 patients (erythropoietin: n = 102; placebo: n = 98).

Assessing Normality

• Age variable assessment:

  • Histogram is right skewed.

  • Departure from straight line in normal probability plot indicates non-normality.

Example of Normal Data

• Characteristics of normally distributed data:

  • Histogram: bell-shaped and symmetric.

  • Normal probability plot: data points closely follow the 45-degree line.

Dealing with Non-Normal Variables

• Apache variable is not normally distributed:

  • Right skewed histogram.

  • Middle points deviate from straight line in normal probability plot.

• Transformation strategy:

  • Apply log transformation for right skewed data to achieve normality.

Normal Probability Plot Anatomy

• Plots data on the y-axis and theoretical normal quantiles on the x-axis.

• Linear relationship indicates nearly normal distribution.

• Construction of the plot is complex, often accomplished with software.

Practice with NBA Heights

• Assess histograms and normal probability plots for NBA heights.

• Analyze normal distribution conformity based on plot characteristics.

Understanding Skewness in Normal Probability Plots

• Right skew: data points bend up to the left of the line.

• Left skew: data points bend down to the right.

• Short tails: S-shaped curve indicates narrower distribution.

• Long tails: Points start below and above the line indicate wider distribution.