Agreement of Influencers

The indicator shows the extent to which specific influencers (e.g. mothers-in-law, husbands) agree with the target group members (e.g. their wives) following the promoted child nutrition practices. Since the data needs to be collected using a representative sample of influencers, be aware of using this indicator, which might double your M&E-related workload, as your other indicators are likely to concern different respondents. See below for details.

Determine the indicator's value by using the following methodology:

1) Define a limited number (3 to 7) of the desired child nutrition practices that are (or will be) promoted by the intervention. Since you want to make sure that the influencers (e.g. husbands, mothers-in-law) are able to say whether they agree with a given practice or not, avoid selecting practices that they might not know.

2) Decide how many (and/or what types) of the promoted child nutrition practices the influencer needs to agree with in order to be considered as “agreeing with [the target group] following the promoted child nutrition practices”. For example, “at least 4 out of 6 promoted practices” or “all the essential practices and at least two other recommended practices”.

3) For each promoted child nutrition practice, prepare two statements – one describing the optimal practice a target group member should follow, the second the practice you aim to discourage (both need to specify who is supposed to do what and possibly also when). For example, in the case of exclusive breastfeeding, the statements can be:

   - When I have a healthy baby aged 4 months, my wife should give the baby only breastmilk and no other fluids or foods.

   - When I have a healthy baby aged 4 months, my wife should give the baby breastmilk but sometimes also other fluids or foods.

It is very important that the statements (especially the one about the negative practice) are defined in a very neutral way, so that they do not lead the respondent to a particular answer.

4) Conduct individual interviews with a representative sample of the targeted influencers (e.g. husbands or mothers-in-law). After a brief introduction (see below), always read the two statements and ask the respondent which of the two statements s/he agrees with more. See example below. If the respondent is not sure, read the statements again (the enumerators should not rephrase them). Repeat this process for the remaining pairs of statements / practices.

EXAMPLE OF SURVEY QUESTION (Q) AND POSSIBLE ANSWERS (A)

Introduction: Now, I am going to read a series of child nutrition practices that come in pairs. Each time please tell me which of the two practices you agree with more.

Q1.1: When I have a healthy baby aged 4 months, my wife should give the baby only breastmilk and no other fluids or foods.

Q1.2: When I have a healthy baby aged 4 months, my wife should give the baby breastmilk but sometimes also other fluids or foods.

Which statement do you agree with more?

A1: _

1) agrees more with the first statement (only breastmilk)

2) agrees more with the second statement (breastmilk and other fluids / foods)

3) does not know

5) Per each respondent, count how many correct practices s/he agrees with. For example, a respondent might agree with 4 out of 6 correct practices.

6) Count how many respondents agree with at least the minimum number of correct practices (see point 2).

7) To determine the indicator’s value, divide the number of respondents who agree with the minimum number of correct practices by the total number of interviewed respondents (exclude those who responded even once that they do not know). Multiply the result by 100 to convert it to a percentage. 

Disaggregate the data by age group and other criteria relevant to the context of your intervention (e.g. ethnic group or residence in rural vs. urban areas).

1) This indicator is based on collecting data from a representative sample of influencers (e.g. mothers-in-law). However, your other indicators are likely to require collecting data from the members of other target groups (e.g. women of reproductive age). This means that you would have to collect your baseline/endline data using two samples of respondents, which is likely to be quite time-consuming. Therefore, before using this indicator in your logframe, you should consider two main options:

   - Use this indicator only if you also want to collect other data from the targeted influencers (so that the effort is worth it – you do not want to do a separate survey because of only one indicator).

   - Use an alternative indicator, such as “% of [specify your main target group members] who think that their [specify the influencers] agrees with them following the promoted child nutrition practices” (see guidance). This indicator will allow you to use one sample only – the main target group members. However, bear in mind that you will be assessing what the target group members think about the influencers’ attitudes – this is certainly important but it can be different from what the influencers actually agree with.

2) Consider also reporting separately on the individual child nutrition practices. For example, the percentage of husbands who agree with their wives breastfeeding exclusively; the percentage of husbands who agree with children aged 12 to 16 months eating eggs as a part of their meals; etc. Such data will tell you which child nutrition practices the influencers most (dis)agree with, showing you what the focus of your intervention should be.

3) Ensure that the endline survey asks about exactly the same practices (and using the same questions) as the baseline survey did.

4) When training the data collectors, emphasize that they have to verify whether they understood the respondent’s response correctly. For example, when the respondent says: “I agree with the first statement”, they should verify it by saying “Do I understand correctly that you believe that children aged 4 months should receive only breastmilk and no other fluids or foods?” This simple measure will help prevent possible misunderstandings.

5) Change the order of the correct / incorrect answers – do not always, for example, put the correct statements first as this might influence the respondents’ answers.