Simple vs Complex Matrices

Length of the survey is one of the most common problems affecting survey research. The longer the survey – both in questions and in the time it takes the user to complete it – the more likely the respondent will drop out before the survey is finished. Researchers look for many ways to cut down on the length of time it takes to answer the survey, which is why matrix questions are easily one of the most common choices for survey design.

The goal of matrices is to place as many questions on a page as possible, in order to reduce clicking and make it easier for the user to answer. Questions are placed on a scale. On the left side are all of the questions/statements, while at the top is the scale itself, such as “Rate your overall satisfaction from 1 to 5.” Respondents go down the line, read the question, and select the appropriate point on the scale within the matrix.

There are generally two types of survey matrices. The first is the simple matrix. The simple matrix is the standard matrix applied in most surveys, with a scale at the top and questions on the left.

Most respondents are familiar with this type of survey design and find it cognitively intuitive. The other type of matrix is a complex matrix. Complex matrices have what can best be described as “interruptions” that make it slightly less intuitive.

For example, in a complex matrix, the user adds subtitles within the left hand column that don’t require the user to answer, followed by questions below each one. For example: Q: Rate how satisfied you are with the following aspects of each division: Sales (no scale) Customer Service (followed by scale) Responsiveness (followed by scale) Marketing (no scale) Customer Service (followed by scale) Responsiveness (followed by scale) etc… All of this is part of one matrix.

In addition, the researcher may have a different scale included within the matrix, though this is far less common.

Rarely is there a reason to use a complex matrix, when it would be nearly as simple to simply break up the complex matrix into several smaller simple matrices that are easier to answer and make more sense for the user. While complex matrices are interesting – and it’s possible they have some research use in very select studies (perhaps where a comparison of answers is more important). But most researchers – especially corporations – need to keep their surveys simple, and complex surveys add a needless complication.