Step 2: Choosing your Data Sources

Rationale

If your data sources are not self-evident, you will need to decide from among several alternatives.

There are always multiple ways of capturing information you need on the indicators you have selected.

Different data sources will vary according to:

Accuracy of the information (validity)
Stability of the measures (reliability)
Ease (and therefore cost) of collecting the data
Risks to the data source and to the data collectors

Important:
*Weigh these aspects to select the data sources that will work best for each indicator.

*Use more than one data source for an indicator to increase the accuracy of your evaluation.

*In many SRHR programs you will not be able to collect data from every person, document or event so you will need to select a sample without introducing biases.

Task 1: For each indicator consider all the logical options

Obtaining responses from:

• People who participated in the program or received the messages that were disseminated
• Other people who know the program participants, such as a man’s wife or vice versa, to avoid the bias of the program participant reporting what he or she knows you want to hear and/or to gain alternative (but equally valid) perspectives on the same thing
• Groups of people who can provide information that is not very personal and that won’t be too severely distorted by the group setting

Observing and recording behavior or events -

• Watch people’s behavior directly at home, school, or public venues; on sports playing fields, in bars etc. in ways in which your presence does little to alter the typical behaviors people exhibit
• Observe the content of TV or radio programs, advertisements in which you can register certain kind of information or images
• Attend public meetings, legislative sessions
• Participate with people in daily life or activities, and observe what they do and say or don’t say

Capturing information from documents that you think are reasonably complete

• clinical, official, or meeting records
• newspaper articles
• legislative proposals
• government statistics
• attendance records

Task 2: Select the option(s) that are best for your situation

You may select different data sources for different indicators. Each data source has its pros and cons.

Tips: Selecting Data Sources

 

 

 

 

 

 

 

 

 

Worksheet: Selecting Data Sources

Worksheet: Fill in Column 3 of the Logical Framework

Option 1: Word Download Word version of the log frame in order to work on your own computer. We highly recommend you download it to your PC before working in it. The Word version is best for modifying the form to suit your needs, for example, adding rows and columns.

Option 2: PDF Download PDF version of the log frame in order to print. PDF files are best for viewing and printing.

Option 3: Fill out the form Online..To enable you to view the entire form at a glance, we have restricted the number of characters to be entered into each text box. Therefore, we only recommend using the online form if you do not need to enter a lot of information. In order to use the online form, you will first need to create an account, and then log in

Task 3: Decide whether to use a sample and what kind of sample to use

If you don’t have the time, money or personnel to hear from everyone, review all the documents, or observe all the events, you will need to select a smaller subgroup of the entire population, called a sample.

Choosing a sample can make your evaluation

more feasible and
less costly

In order for your findings to be relevant to the entire group or population the sample needs to resemble the larger group in ways that are directly relevant

• to the kind of information you will be collecting and
• to the objectives of the program

You will use these characteristics to:

• to select the sample in the first place and/or
• check if your sample ended up similar to the population as you intended it to

Tips: Kinds of Samples

Tips: How Big a Sample Do You Need?

Worksheet: How to Decide What Kind of Sample to Select

Whatever means you use to select a sample, always

• make absolutely clear who or what was, or was not included, and why, and
• think out the implications of such decisions on your potential findings