Bangladesh:
Two Independent Variables
Prevalence of Low Arm Circumference
by Latrine and Water Source Quality
Individually - Safe Latrines:
There is an obvious association between malnutrition and poor sanitation, where the areas with poor access to safe latrines have a much higher prevalence of malnutrition (12.0%) compared to areas with better access to safe water (8.5%). The ANOVA tables supports that this difference is significant at the 0.006 level.1. Open bdeshc.sav
2. Click on Statistics, Compare Means, and Means.
3. In the Dependent list, enter the variables acprvfm (prevalence of low arm circumference) and from the left hand variable list
4. In the Independent list, enter the variable for access to safe latrines >45%, latcat and those using other water source >50%, wathcat. (Note: those with >50% using "other water source" will be treated as having bad water; those with with .45% safe latrines will be considered to have good sanitation.)
5. Click on the Options box and check the ANOVA table box under the area labeled Statistics for the first layer.
6. Click on Continue and OK.
Individually - Safe Water:
The tabular means table below for access to safe water also supports an association exists
between malnutrition and poor access to safe water, where the areas with poor access to
safe water have a much higher prevalence of malnutrition (11.5 %) than in areas with
better access to safe water (8.7%)
Together - Safe Latrine and Safe Water:
1. Open bdeshc.sav 2. Select Statistics, General Linear Model (GLM) – General Factorial… 3. Select acprvfm. Move it to the Dependent Variable field.
4. Select wathcat (good/ bad water source) and latcat (good / bad latrines), and move them into the Fixed Factor(s) field.
5
. Click on Options.6. Check Display, Descriptive statistics and Observed power.
7. Click Continue, then OK.
INTERPRETATION: The final table shows the results of the ANOVA for the two independent variables. The areas of interest are marked in red circles with the category name underlined. Both the water source and latrine source have a significant level of association with the prevalence of low arm circumference among children(Sig= 0.013 and 0.007 respectively). The variable labeled WATHCAT*LATCA is an INTERACTION term (like the one introduced in the Kenya example), which tests for an association between water and latrines to determine if the association with low arm circumference is overlapping. In this case the interaction DOES appear to be significant! The Significance level of 0.10 indicates that the effect of water and latrine together is not just by chance but of a level that is worth giving some attention. This is just the tip of the iceberg for interaction…more will be covered in the following section on regression analysis.