1) The visual scale is used to make it easier for the respondent to choose the response that best represents her/his feelings. If you do not wish to use the visual scale, you will have to ask the respondent: Would you say that you strongly agree, somewhat agree, somewhat disagree or strongly disagree with the following statement? [read out the given statement]
2) Keep in mind that the fact that someone has a positive attitude towards a certain behaviour does not automatically mean that s/he will adopt it (think of the many health and other practices we all approve of but do not necessarily follow).
3) An alternative version of the indicator is “the average value of the [specify the behaviour] Attitude Index” – for example, “the average value of the exclusive breastfeeding Attitude Index”. The value of the Attitude Index is calculated by giving each response a score where the most favourable response (e.g. strongly agree) receives 4 scores, the second most favourable response (somewhat agree) receives 3 scores, the next one (somewhat disagree) receives 2 scores, and the least favourable (e.g. strongly disagree) receives 1 score. As the next step, calculate the total scores received by all respondents and divide them by the number of respondents (exclude those who did not know). The resulting number is the average value of the Attitude Index. This Index is able to record more subtle changes in the attitudes of the target group but it does not tell you the proportion of the target group members with favourable attitudes (unless you calculate this separately).
4) If you want your indicator to capture people’s attitudes towards several different behaviours, you can rephrase it to: “% of [specify the target group] with favourable attitudes towards at least X out of X promoted behaviours”. For example, “% of mothers of children under two with favourable attitudes towards at least 5 out 7 promoted infant and young child feeding practices”. To calculate the value of the indicator, use the same methodology as described above, with the only exceptions being:
- when analysing the data, calculate the number of respondents with favourable attitudes towards the minimum number of behaviours (e.g. at least 5 out of 7)
- to calculate the indicator’s value, divide the number of respondents with favourable attitudes towards the minimum number of behaviours by the total number of respondents. Multiply the result by 100 to convert it to a percentage.
5) Measuring people's attitudes towards a certain behaviour can also be based on responses to several statements related to the same behaviour (as opposed to using just one statement as proposed above). For example, you can use 3 different statements and if the respondent is positive about at least 2 out of 3 statements, s/he can be taken as 'having favourable attitude'.
6) In the statements, ask only about behaviours the respondent is likely to be familiar with (otherwise you will receive many ‘I don’t know’ answers that are difficult to analyse).
7) If you are concerned about the data being affected by social desirability bias (respondents saying what they think you want to hear rather than what they really think), consider presenting the negative behaviour in the statement. For example: “Children younger than six months can receive breastmilk but also water, juices and other fluids and meals” or “When children have diarrhoea, they should be given only very little water to drink”. In this case, the respondents who hold such unfavourable attitudes are more likely to admit to them. If you use this approach, make sure that you reverse the meaning of the pre-defined answers (strongly agree will be a negative answer; strongly disagree will be a positive answer).
8) The guidance on the use of this indicator was adapted from Roll Back Malaria Partnership (2014) Malaria BCC Indicator Reference Guide.