Many peacebuilding evaluation designs utilize a mixed-method approach to data: they seek to collect and use both qualitative and quantitative data. Yet quantitative data seems to be viewed with some degree of suspicion amongst peacebuilders. A common perception is that quantitative data is too ‘black and white’ and does not account for unique perceptions and worldviews of the surveyed participant. There are, however, multiple ways in which to capture complexity in quantitative data.
Surveys and questionnaires are a frequent tool for both qualitative and quantitative data collection, and can also be used for quantifying the qualitative. It’s a great tool for measuring, on a large scale, how people perceive certain issues. Certainly, key informant interviews and focus group discussions are critical in generating such data, but these have limited utility: extrapolation to the larger population can be difficult. This is where surveys come in handy.
Likert Scales: Assessing Attitudes and Perceptions
The general principle behind Likert Scales is to ask participants to respond, in terms of the extent to which they agree, to a series of statements about a particular topic.
A key assumption behind the use of such a tool is that the strength and/or intensity of experience or thought is linear. Respondents must select where they lie on a linear continuum of responses that generally include at least five potential answers along the following or similar lines.
- Strongly agree
- Agree
- Undecided
- Disagree
- Strongly disagree
The University of Connecticut has developed an excellent list of potential scales that are grouped in accordance to what the scale is attempting to measure: agreement, frequency, importance, quality and likelihood. This list is not meant to be exhaustive in the possible scales or the groupings.
Using such a scale allows you to measure the attitudes and perceptions of broader populations. An excellent supplemental tool for such data includes the traditional qualitative tools such as focus group discussions and key informant interviews. The qualitative tools will allow you to ‘dig beneath the surface’ of the broader survey data and look at specific motivating or key driving factors behind such attitudes and perceptions—but both sets of data are important in gaining as complete a picture as possible of the environment and the changes brought about as a result of the intervention.
03/27/2013 at 3:34 PM
Thank you for this post!
I sometimes find that there are discrepancies in people’s personal rating systems. This is the issue I have with the scale system, that although it does allow for views to be expressed it does not layout criteria of what each level should entail. For example, I may put down that I strongly disagree with something because I disagree with 50% of the idea. While another person may only put down that they strongly disagree with something only if they 100% disagree with everything that is proposed.
This is why I find that the level system in which outlines criteria that each level would be in order to be classified in this level to be a more effective measure of people’s attitudes and perceptions. This encapsulates what is expected of a certain level and leaves less room for individual interpretation of what each level means.
However, this method also has its draw backs which include the fact that it takes more time to outline criteria. As well as, the fact that it takes the survey participants longer to fill out as they have to read through more criteria and could make them less likely to take the survey.
11/10/2012 at 6:26 PM
I really appreciated reading this piece. Coming from a social sciences background and working on communication projects currently, I often face the opposite objection: People love quantitative data and distrust qualitative data. The idea of combining the two to create a larger picture is never discussed. I am a bit curious on the actual application though. Would you suggest using a survey on a large population and interviewing representatives for more in-depth responses to those questions, or would you suggest surveying all those that you interview to get an idea of the general trend in addition to detailed personal accounts?