Module 03 Activity

Scenario

Describe the data honestly before testing anything about it.

What you build

A descriptive summary, a distribution chart and a paragraph in which every claim is traceable.

Steps

1. Compute mean, median, std, IQR and quartiles. Where mean and median diverge, find the values responsible.
2. State which ddof you used, and use ddof=1 for any sample standard deviation.
3. Group by the reporting category carrying n, the sd and the standard error, and check the group sizes sum to the row count.
4. Plot the distributions before testing any means; note bimodality, truncation or floor effects.
5. Label every error bar, and never judge a difference by whether two intervals overlap.
6. Write each finding with the measure, the size, the comparison and the period named.

Evidence to hand in

• The full descriptive summary with its ddof stated.
• Your choice of central measure, with its justification.
• The grouped summary including n and se.
• The distribution chart and what it showed.
• The labelled error bars, if any.
• The findings paragraph.

Review checklist

• The ddof convention is stated wherever a standard deviation appears.
• Every grouped row carries n and a measure of uncertainty.
• Group sizes sum to the row count, or the dropped keys are explained.
• The distributions were plotted before any test of means.
• No finding claims a cause, and every one names its period.

Mark complete