MERGING GAM RATES FROM DIFFERENT SURVEY ZONES

[Geoffrey MurithiParticipant]

In my area of operation we have traditionally segmented the area into four separate survey zones,we have not been combining the data from the various zones to obtain a cumulative figure.I am lost on how i can combine the GAM rates from the various survey zones to obtain a cumulative GAM rate for the entire area.Could some help me understand how i can go about this?.

[deeps.kuParticipant]

Well according to me.. if you have 4 zones and those make area to 100% then you make take percentage of percentage to combine GAM rates:

For instance:
Zone 1 is 15% of total area and GAM rate is 12% so this means actually GAM rate is 1.8% (15%*12%) and if you take same for all remaining zones, then you might get what you need.

This is just a way to do it.. may be you find some technical way of doing it.

[SMARTKeymaster]

The process of calculating a combined GAM rate is a bit more involved.

When surveys are undertaken in separate strata (zones), the sizes of the populations (and sample size in each stratum) are different, and the distribution of the sample for each strata is not proportional to the sample calculated for the entire population.

Weightings are correction factors to adjust these differences in probability of selection/inclusion (differences in the size of the populations). The weight for each child is calculated taking into account their unequal probability of being selected between strata (zones) and is inversely proportional to their chance of being selected/included in the survey. The standardised weights for each strata are then calculated and used to determine a cumulative GAM. This will correct for the bias of a non-representative sample (differences in the size of the populations in each stratum/zone).

For example, to calculate cumulative GAM for rural (population: 20,000, sample size 800, 15% prevalence) and urban strata (population: 75,000, sample size 600, 5% prevalence):
Calculate the weights for each stratum.
Rural: 20,000/800 = 25.
Urban: 75,000/600=125
Calculate the standardised weighting factors.
Rural: 25 / [(25+125)/2*] = 0.333
Urban: 125/ [(25+125)/2*] = 1.667
*Divide by the number of strata. i.e., if you were to have 4 strata/zones, divide by 4.
In ENA Options tab, select Showing columns for measure, clothes and weighting variables.
Enter data for each of your strata in the Data Entry Anthropometry tab.
Identify strata (zones) by numbers (e.g., 1-4), and
Enter the corresponding standardised weight factor in the WTFACTOR column for each strata.
Select Weighted Analysis of Survey from the Extras menu for your weighted prevalence (it will not appear in the Results tab).
See the ENA Software Manual: https://smartmethodology.org/survey-planning-tools/smart-emergency-nutrition-assessment/

• This reply was modified 11 years, 2 months ago by  SMART.

[Author]

Posts