Slideshow version

This last section is all about marginal vs conditional distributions in two-way tables. Let’s look at our table from last section to see where we can find these different distributions.

  Morning Evening Total
Coffee 45 15 60
Tea 10 42 52
Total 55 57 112

First up is marginal distributions, which you can find on the margins (i.e. edges) of the table. These only look at the distribution of one variable. The marginal distribution of preferred drinks would be

  Morning Evening Total
Coffee 45 15 60
Tea 10 42 52
Total 55 57 112

The marginal distribution of preferred time to drink would be

  Morning Evening Total
Coffee 45 15 60
Tea 10 42 52
Total 55 57 112

If you need percentages for either of these, divide each number by the total, so 112 in this case.

Conditional distributions on the other hand, involve both variables. The conditional distribution of preferred time to drink for coffee drinkers would be

  Morning Evening Total
Coffee 45 15 60
Tea 10 42 52
Total 55 57 112

The conditional distribution of proffered drink for evening drinkers would be

  Morning Evening Total
Coffee 45 15 60
Tea 10 42 52
Total 55 57 112

And again, if you need percentages, divide by the the total. The catch here is to use the relevant total. For the first one, the total is 60. The second example’s total was 57.