**Steps
to Targeting**

**Here is the Step by Step Process:**

For ease of presentation in this exercise use **only
those districts in division 2. **To do this select data, select cases and press the
if… button. Select the div variable and move to the right and type "=2" so
that the formula reads "div=2". Press *Continue* and then *OK*.

**1. Rank districts by Prevalence**

Try creating a Case Summaries list using the Bangladesh data set:

- Open
*bdeshc.sav* - Under the
*Data*menu, go to*Sort Cases*. *Sort By*, select the variable prevalence low ac in males and females (*acprvfm*)*Sort Order*: Descending- OK
*Statistics*menu- Summarize: Case Summaries
- Select
**div, district,**and**acprvfm**(prevalence of low ac). (Make sure limit to first 100 cases is selected) *Statistics*:*Cell Statistics*- none- Click on
*OK*

**2. Add Number of
Malnourished, and Percentage of Total Malnourished**

To add the number of malnourished and percent total
malnourished, run the same Case summaries and add the variables labeled **maln# **and **totmalpt**
to the list run before, then press *OK.*

The result should be something like the following.

**Remember that since we selected only
district two the %total malnourished will not equal 100%.**

If the number of malnourished
children was not given as it is here, then it could easily be calculated using the *Transform,
Compute *option in the menu. The new target variable would be named **maln#** The
prevalence of low arm circumference would be multiplied by the population of children
under five. In this case, the population under five is not given, therefore an estimate
must be made by taking the average for that region of the world (probably about 20% of the
total population). The Numeric Expression would be **acprvfm*(pop*.2)**

**3. Rank by the number of malnourished**

__To rank by the number of malnourished, go back to the previous routine of sorting
the data set by a variable (in this case use maln#). Recall that sorting is done by
clicking on __*Data, Sort cases, *insert **maln# **as the variable
(descending order) and click on *OK*. This will order the data with the
greatest number of malnourished at the top and the least number of malnourished at the
bottom of the list instead of the current ranking by district.

**4. Note which districts go up to the top of the ranking list**

__The top five districts for number of malnourished children are: Dhaka,
Kishorenganj, Mymensingh, Madaripur, and Jamalpur, in that order. Take a note of the
order of these districts and which ones are at the top so that they can be compared to the
order in the next steps. __

**5. Rank again by prevalence and make initial selection**

__Under __*Data*, then *Sort cases*…
sort by prevalence of low ac (**acprvfm**), descending

__Here you see that the top five in prevalence are
highlighted in red and the top five in number of malnourished are highlighted in yellow.
This is the comparison between districts with the highest prevalence vs. the
highest number of malnourished and it shows how they do not always perfectly coincide.
Depending on what the intervention strategy is and how much per head the cost of
implementing, the district targeting will be modified to most efficiently provide for the
most children in need and those that are the worst off (considering both numbers and
prevalence). This will depend on the amount of resources available as well and the
specification of the funding agency.__

**6. Add an Indicator of Health
Services (measles immunization rate here called meas)**

Click on *Statistics, Summarize, Case
Summaries *and insert the variables **div, district, acprvfm, maln#, totmalpt,
meas** and click on *OK.*

When estimating the other potential influences, an indicator of health service might be helpful. Here Jamalpur, Madaripur, and Narsingdi do not have good access to health care (at least measles immunization). This may be helpful in furthering the decision on which areas to provide alleviation programs to and which ones might already have a high level of services provided.