Summary Descriptive Table
Return to One-Way Page 1
Using Descriptives (bdeshc.sav):
In the Data Editor menu, click on Statistics, Summarize, Descriptives.
Select the variables for low arm circumference (male, female, and total)-acprvm, acprvf, acprvfm, % measles immunization-meas, Percentage of households with iodized salt-hhiod, and population-pop, and use the arrow key to place them in the Variables box.
Under Options, you can select other statistics to be generated (e.g. Mean, Minimum, Maximum, Std. Deviation) and click on Continue.
The date set is at district level (N=64 districts) in Bangladesh, thus each case is a district and the minimum and maximum give the range of district values, for example. Note similarities of low arm circumference in males and females and overall. The prevalence of households with iodized salt has a noticeable range displaying a large disparity between districts and their access and utilization of iodized salt. Thus one would suspect that those districts with low prevalence of iodized salt would be at a higher risk of goitre and iodine deficiency. The next section on targeting will cover more on this topic. For population, all you are seeing is the range and mean population per district. There is a box to select in the Descriptives command that will produce the sum of the populations, which would be the total population for Bangladesh.
Using Frequencies (bdeshc.sav):
1. Click on Statistics, Summarize, Frequencies.
2. Choose the variables exbfcat (exclusive breastfeeding <2% at >5 mo. or >2%), hiodcat (salt iodization > or < 45%), latcat (safe latrine > or <45%), meascat1(measles immunization 0-70%, 70-85%, or >85%), and wathcat (using unsafe water sources < or > 50%). Move each of these variables from the variable list on the right into the Variable box on the left, using the arrow key.
3. Click to remove the check mark in the Display frequency tables box if you would not like to see the five individual frequency tables.
4. Click on the bottom box labeled Statistics.
5. Under Dispersion,
select Std. Deviation, Minimum, and
6. Under Central Tendency select Mean.
7. Click OK.
For categorical data, the descriptive statistics table is far less useful than the frequency tables because the summary information is often difficult to interpret (depending on the coding scheme). For example, the mean value might not be as useful if the categories are not simple bivariate (dummy code) 0 and 1.
If you were to run all of the frequency tables as well, there would be five additional tables that followed, which would all appear in the same format as this table for exclusive breastfeeding.
Although the Descriptive statistics are slightly less useful for categorical variables and the frequency tables much more useful, the descriptive statistics can show some relevant information such as the number of cases, missing cases, and generally what direction the responses are tending towards. In the Statistics table above, in the second column labeled exclusive BF the categories are 1 and 2 and the mean is 1.3. This can be interpreted so that it reads, 30% of the districts have >2% that are breastfeeding after 5 months of age; but this information is more easily accessed by looking at the frequency table. The descriptive statistics tables are clearer when looking at continuous variables and dummy coded categorical variables assuming the values 0 and 1. For most categorical variables, the frequency tables probably provide the necessary information in a cleaner format. The frequency table for exclusive BF categories includes a frequency (absolute number), a percentage, valid percentage and cumulative percentage column as well as a Total for each. Normally,the frequency and valid percentage columns are the most useful columns. For this table, it seems that a significant number (30%) of the districts have continued duration of exclusive breastfeeding after 5 months of age and in these districts, it might be useful to educate mothers on the importance of complementary feeding by 6 months of age.
Return to top of Page