Survey research can be classified into two categories, according to the time period being studied. Cross-sectional surveys take measurements at a single point in time. Longitudinal surveys take measurements at multiple points in time. These two types of surveys are useful for different purposes, but can be misused if not properly understood. Consider the following example.
The head of Global Consumer and Market Intelligence for a global athletic shoe manufacturer once called me with a very specific request. She had developed the hypothesis that people develop emotional attachments with the first brand of athletic shoes they buy, making them loyal to that brand throughout their life. To test this, she wanted a survey to measure the first brand of athletic shoes a person bought and the number of pairs of each athletic shoe brand they purchased in later years.
Now, imagine you are a survey respondent who is asked to list the brand name of every pair of athletic shoes you purchased, by year, for every year of your life. Even if you were willing to invest the time and effort to complete this survey, would you be able to provide accurate information? Would you be able to correctly recall what brand of athletic shoes you purchased when you were 24 years old? And would you remember if you bought those when you were 24 instead of 23 or 25?
My client was trying to measure a longitudinal phenomenon with a cross-sectional survey. Rather than mixing these conflicting time periods, surveys should match the time period of the survey with the time period of the attitude or behavior being measured.
Cross-Sectional Surveys
Many market research surveys are cross-sectional surveys, capturing a “snapshot” at a single point in time. Cross-sectional surveys may measure consumer awareness, attitudes, or behaviors at that moment, or at least within recent memory. Cross-sectional surveys can be useful for measuring current or recent phenomena, but are not often the best way to measure very old events.
For example, my shoe client might have conducted a survey to ask consumers what brands of athletic shoes they currently own, and how many pairs of each brand. Or, perhaps she could have asked people to name the brand of athletic shoes they had purchased most recently. A cross-sectional survey would be appropriate to measure these questions.
Longitudinal Surveys
In contrast, longitudinal studies are designed to measure phenomena over time. Longitudinal surveys might ask respondents the same question at multiple points in time, to determine how something has changed over time. Longitudinal surveys might measure different, but related, phenomena over time, to examine delayed cause-and-effect relationships.
For example, my shoe client might have conducted a longitudinal study each year to ask the same consumers what brand of athletic shoes they wear most often. Or, if she wanted to know the effect of Super Bowl advertising on sales, she might have conducted a survey in February to ask respondents whether they recalled seeing her brand’s Super Bowl ad, and then followed-up with a survey in May to measure whether or not they had bought her brand in the intervening months.
Choosing a Cross-Sectional or Longitudinal Survey
Cross-sectional and longitudinal surveys each have their own merits and drawbacks. To determine whether to conduct a cross-sectional or longitudinal survey, consider the research question. If the research question has to do with measuring a phenomenon at a single point in time (e.g., what brand currently has the highest brand awareness in our category?), then a cross-sectional survey is probably the better method. However, if you need to understand how a phenomenon changes over time for the same individual (e.g., does buying my brand at a younger age make someone more likely to become brand loyal?), then a longitudinal survey might be the better choice.
A third option that companies often adopt is to conduct a tracking survey. Tracking studies are repeated cross-sectional surveys. Like longitudinal surveys, tracking studies repeatedly measure the same variable over time. Like cross-sectional surveys, tracking studies measure a different group of respondents in each wave. This is often easier than tracking the same respondents from wave to wave over a period of months or years, but still allows the researcher to analyze trended results.
Conducting Your Survey
For help with your next market research survey (cross-sectional, longitudinal, or tracking), contact the survey experts at MMR Strategy Group.
Dr. Justin Anderson
Vice President
MMR Strategy Group