**Mean
Outcome Scores for Sanitation and Education**

The routine for running tabulations with the Mean scores is probably
familiar from previous exercises. We will look at both the mean weight/ age z-score
and the mean score for a variable called **waprev**, which is a trick used to
calculate prevalence by coding malnourished as 1 and not malnourished as 0. Run the
mean outcome for these two variables by education (DLOWEDN) and access to toilet
(NOTOILET).

1. Open

keast4j.sav2. Click on

Statistics, Compare Means, Means.3. Enter the variables

wazandwaprevin theDependentvariable list using the arrow key.4. Enter the variable

DLOWEDNin layer 1 of theIndependent(s)variable list using the arrow key, then click onNextto proceed to Layer 2.5. Enter the variable

NOTOILETin layer 2 of theIndependent(s)variable list.6. Click on

OK.

These results can easily be transferred to a more readable table, like that given in the main text, but you must first be able to tease out the information you are looking for. There is a slight change that will be made to the prevalence column to make it presentable...just multiply by 100 to give the percent that are malnourished. The first group on the left are those that have higher education (primary or more) and the difference in levels of malnutrition between those with and without access to toilets is large (mean scores are -0.958 and -1.749 respectively). There is also a correspondingly large difference in the prevalence of malnutriton (WAZ <-2 SD) for those with and without access to toilets in the better education group (15.2% and 50.0%, respectively). When you move down to the next section for low education (<primary), you will not find a large difference between those with and without access to toilet. There is also not a large difference in the prevalence data. This supports what we were finding in the regression analysis.