Variables, Interaction of

An Example: Interaction of Variables in ANOVA

Imagine that you are interested in determining the best way to do your laundry. You decide that the cleanliness of your clothes is your outcome variable (dependent). Cleanliness is scored on a scale from 1 to 10, where 1 is least clean and 10 is most clean. You have separated your clothing into two different kinds: not very dirty clothing (NVD) that you wore when you were at school/work and very dirty clothing (VD) that you wore to exercise and complete outdoor chores. You also have recently purchased two different types of laundry detergent: Detergent X has few chemicals and Detergent Y has lots of chemicals.

You might ask three different questions based on this scenario:

• 1.
  Is there a difference in how clean your clothes are depending on whether they are NVD or VD at the start (ignoring type of laundry detergent used)?[Page 1847]
• 2.
  Is there a difference in how clean your clothes are depending on whether they are washed using Detergent X or Detergent Y (ignoring how dirty the clothes were at the start)?
• 3.
  Does the effect of being washed using Detergent X or Detergent Y differ based on whether the clothes were NVD or VD at the start?

If you ask question 1 or question 2, then you are interested in main effects (i.e., effect of one independent variable on a dependent variable). If you ask question 3, then you are interested in the interaction of variables. In other words, you are asking if the effect of one independent variable (how dirty clothes are to start: NVD or VD) on a dependent variable (cleanliness of clothes) depends on the level of a second independent variable (type of laundry detergent used). It would be a mistake to only consider each variable individually because it may be a combination of variables that achieves the desired result (in this case, the cleanest clothes).

...