Analyze Data

Factor Analysis Example

Questionnaires often include a series of related questions. For example, respondents in a tourism survey might be asked to indicate the extent to which they agree or disagree with various statements relating to travel. These might include statements such as:

  1. I like to learn new things when I travel.
  2. I like to meet people from the local area when I travel.
  3. I like to be on my own when I travel.
  4. I like to travel as part of a group.
  5. I like to know a lot about a place before I travel there.
  6. I prefer to travel to places where my language is spoken.
  7. I like to try the local cuisine when I travel.
  8. When I’m traveling, I try to avoid places where I won’t be able to have the same level of comfort and convenience that I have at home.
  9. When I travel, I like to stay in luxurious hotels.
  10. When I travel, I like to go to places I’ve heard a lot about.
  11. When I travel, I like to go places that few other tourists visit.
  12. I like to participate in adventurous activities when I travel.
  13. I like to be pampered when I travel.
  14. The vacations that I enjoy most are those where I just get to relax.
  15. I try to pack as much into each vacation as I can.
  16. I don’t like to make a lot of plans before a trip — I prefer to just figure things out as I go.
  17. I like to return to destinations that I’ve enjoyed in the past.
  18. I like to experience nature when I travel.
  19. I like to visit well-known cities when I travel.
  20. I prefer having more short vacations to having fewer longer vacations.
  21. …

If there are a large number of such questions, trying to understand and interpret the responses can be overwhelming. Factor analysis addresses this problem by grouping questions into categories based on response patterns — specifically, through calculating the correlations between responses to different questions. For example, answers to the questions about going places few other tourists go and meeting people from the local area are likely to be positively correlated — if people like one, they probably like the other and if they dislike one, they probably dislike the other too. There is likely to be a negative correlation between answers to the questions about traveling alone and as part of a group — if people like one, they’re likely to dislike the other. Factor analysis groups questions based on their correlations, making it possible to identify higher-order categories which can often lead to more meaningful interpretation and be more useful in subsequent analysis. In the travel example, these higher order categories may be things such as authenticity, which would encompass meeting local people, trying local food and going to destinations that don’t attract a lot of tourists, and comfort, which would include things such as returning to familiar destinations, going places where one’s own language is spoken, and having the same level of comfort and convenience the traveler is accustomed to at home.