Grid questions are particularly susceptible to yielding poor quality responses, because they are easy for respondents to speed through with little care or thought. It is true that respondents who “straight-line” through a grid can be identified, and their answers excluded from the survey results. But this is not possible with respondents who just click random answers.
However, there are various techniques that can be used to give better grid results. Here are three easy ones.
Use shorter scales
The easier your online survey questions are to answer the better will be the results. And this includes grids, with a 3 or 5 point scale, for example, being easier to answer than a 10 point scale. Of course, you may need a 10 point scale on occasion, but as a rule the shorter the better.
Manage the pace
You can ask your survey provider to prevent respondents from speeding through your grid questions by only showing the “Submit” or “Continue” button (to click to the next page) after a certain number of seconds. And it can be helpful to precede the grid with an introduction which explains that the grid will be on the screen for x seconds.
Insert check elements
A grid question usually comprises of a number of elements, or rows. So, one useful technique is to add an element which is simply there to catch the speeders. For example, if one of your elements is “I like apples” then your check element could be “I don’t like apples”. You might want to be a bit more subtle than this, but you get the idea. When someone gives exactly the same answer to your two “opposite” elements a warning alert can appear, asking the respondent to ensure they read the elements carefully; or they can be screened out, with the agreement of the sample provider.
Of course, there is more you can do, such as breaking up grids into individual elements, or repeating the scale every few rows. Whatever approaches you take, quality control is worth considering.