When you have more than one independent variable, sometimes you want to look at how they work independent of each other. In this lesson, we’ll examine main effects in factorial design and how they differ from interactions.
Joanne is a psychologist who studies the television habits of children. She wants to study whether gender plays a role in preferences for live action (or real) television shows over animated television shows.
Specifically, she believes that little girls will prefer live action shows and that little boys will prefer animated shows. But there’s something else, too. Joanne wonders if these preferences will change with age. For example, will children who are age 3 prefer animated shows, while 5-year-olds prefer live action shows?Joanne’s study has two independent variables, gender and age, and one dependent variable, which type of television program is preferred. When a study has two independent variables, it is a factorial design study. Factorial design gets its name from the fact that independent variables are often called ‘factors.
‘Within factorial designs, there are two ways to look at the results. Let’s look closer at one of the ways, called main effects, and compare it with the other way to look at results.
Ok, so Joanne gets her results back, and she has two different questions that she wants to answer:1. Do girls prefer live action and boys prefer animated shows?2. Do older children prefer live action to animated shows?The main effects of a factor are simply whether that factor has an effect on the dependent variable on its own. For example, Joanne wants to know if gender affects preference and whether age affects preference. She can look at the main effects for each of these factors separately, and it will tell her what each of them does.
One way that researchers look at the main effects of their factors is by drawing a table. Each column of the table represents one level (or option) for the first factor, and each row of the table represents one level of the second factor.For example, if Joanne drew her table, she might have three columns: 3-year-olds, 4-year-olds, and 5-year-olds. She’d also have two rows: boys and girls.Now all that Joanne has to do is fill in the mean, or average, for each of the boxes. For example, let’s say that a 5 means that a person prefers live action shows and a 1 means they prefer animated shows.
She’d average the numbers for all the 3-year-old boys and put that number in the box for 3-year-old boys. She’d do the same for 3-year-old girls, and 4-year-old boys, and so on until all the cells of her table are filled up.Ok, so Joanne’s filled up the cells in her table. But what does that have to do with main effects? Well, now she needs the mean for each row and each column. If the average score for the row labeled ‘boys’ is 1.
3 and the average score for the row labeled ‘girls’ is 4.7, she can see that girls prefer live action shows and boys prefer animated shows. There is a main effect for gender on television preference.Likewise, if she averages the age columns, and the 3-year-olds average out to 2.7, the 4-year-olds average out to 3, and the 5-year-olds average out to 3.3, she can tell that there is a main effect of age on television preference.
Main Effects vs. Interactions
Main effects essentially look at the factors individually. Through main effects, Joanne can see if gender has an effect on television preference, and she can also see separately if age has an effect.But what if gender and age work together to influence television preference? For example, what if girls prefer live-action for the bulk of their childhood, while boys change preferences from animation to live-action as they age?An interaction looks at how two or more factors work together to affect the dependent variable.
Essentially, you are looking at how the two factors interact, hence its name.Interactions are more complex than main effects, but they can tell us a lot about the world at large. After all, none of us live in a vacuum, and so there are often multiple things that work together to change us.
But that doesn’t mean that main effects are not as important as interactions. Both main effects and interactions give us different pictures of research. It’s just that they answer different questions.Think about it like eating dinner.
Main effects are like eating your mashed potatoes and savoring the way they taste, and then eating your peas and enjoying them on their own. Interactions are like mashing your peas into your mashed potatoes and getting a conglomerate of both.
Factorial design involves a study that has two or more independent variables, or factors. The main effects of each factor is how it influences the dependent variable on its own, while interactions are how the factors work together to influence the dependent variable.
After watching this lesson, you should be able to interpret how main effects are used in factorial design as well as how they contrast to interactions.