Using Rating Scales with Market Research Software
Rating scales of various kinds are common in market research. Scales can be purely numeric, such as “On a scale of 1 to 5, with one being the worst and five being the best, how would you rate your experience with Acme widgets?” Or they can use labels for each point in the scale, such as “How would you rate your experience with Acme widgets? Do you find them excellent, good, fair or poor?” When you use labeled scale points, your software will convert them to numbers for your analysis.
If you are doing a mail or web survey, your software should offer a choice of how you visually present a scale. One option for a numeric scale is a fill-in-the-blank format. An option for either a numeric or wholly labeled scale is a grid or matrix format, in which the rows represent different items to be rated and the columns offer the rating choices. This is probably the most commonly used format. Another option is a multiple-choice or series of multiple-choice questions. Yet another option for web surveys is to use sliders. These can be engaging, when used in moderation.
Good market research software should also help you analyze the answers to your scale questions. You could display the results in a table that matches the grid format mentioned above, with the rows being items and the columns showing which percentage of the respondents chose each choice. This is often the simplest and most compact way to analyze the overall answers chosen for a series of items. One limitation in this method is that it makes it cumbersome to compare how different subgroups within your sample rated the items. For example, you would have to create separate tables for men and women, which makes it hard to compare their answers.
There are two main ways around this problem. One is to analyze the ratings for each individual item as a multiple choice table, with the rows representing the choices and the columns representing different demographic groups. This makes it easy to see differences between those groups for one item at a time. Another approach is to create a table in which the rows are the items and the columns are demographic groups, with each table cell showing the mean or average rating given by each group to each item. This method makes it easy to compare and contrast how different groups rate different items and is probably the most common ways survey researchers analyzed rating data, but it is not without controversy. Our article oncorrelations discusses this issue.
Your market research software should let you analyze rating scales in any of those formats, regardless of format used to collect the data.