Chapter 12- more than one IV

Question 1

Interaction effect

Answer 1

In a study examining the effect of cell phones on driving quality, age is a variable that could affect the results. Young drivers might be less affected by cell phones since they grew up with them, but could also be more affected since they have less driving experience. Adding another independent variable allows researchers to look for an interaction effect- whether the effect of the original independent variable (cell phone use) depends on the level of another independent variable (driver age). The interaction between these variables allows researchers to ask whether the effect of cell phones depend on age.

Question 2

Intuitive interactions (2)

Answer 2

1. Crossover interactions

2. Spreading interaction

Question 3

Crossover interaction

Answer 3

Results can be described saying “it depends”- you might only like hot food or cold food depending on which food it is. You would always want ice cream to be cold and pancakes to be hot. Graph forms an X

Question 4

Spreading interaction

Answer 4

Results can be described saying “only when”- it matters when you say “sit” to a dog when you’re holding a treat, but if you don’t have a treat, the dog won’t sit at all. The lines of the graph touch- one is horizontal, the other has a slope. Holding a treat and not having a treat are independent variables that impact the amount of time the dog sits (the dependent variable).

Question 5

Factorial design

Answer 5

A design where there are two or more independent variables (factors)- tests for interactions. The 2 independent variables are commonly crossed to create 4 unique conditions (cells). Each independent variable has 2 conditions.

Question 6

Participant variable

Answer 6

A variable like age, where the levels are selected (measured) instead of manipulated. They aren’t true independent variables, but they’re called independent variables in factorial designs for the sake of simplicity.

Question 7

How can factorial designs test limits?

Answer 7

If there is no interaction between variables (age of the drivers and whether they used their phone or not) the graph will be two parallel lines. Cell phone use did not interact with/depend on age.

Question 8

How are factorial designs related to external validity?

Answer 8

When researchers test an independent variable in more than one group at once, they’re testing whether the effect generalizes. If the independent variable affects the groups in the same way, it suggests that the effect of cell phone use generalizes to drivers of all ages.

Question 9

How do interactions show moderators?

Answer 9

In factorial designs, a moderator is an independent variable that changes the relationship between another independent variable and a dependent variable. The effect of one independent variable depends on (is moderated by) the level of another independent variable. Body weight moderates the affect of alcohol on aggression- people with higher body weights are more likely to be aggressive when drunk.

Question 10

Example of using interaction to test a theory

Answer 10

One memory theory states that adults remember more than children do because of their accumulated knowledge they can use to make connections. The study recruited a group of children who were chess experts and a group of adults who were chess novices. They also performed two memory tasks- remembering digits and the location of chess pieces. This was a 2 by 2 factorial design. Adults could better remember the digits, but the children could better remember the chess pieces. These results show the interaction predicted by theory- children’s memory capacity can be better than adults when they have expertise in the topic.

Question 11

In a factorial design with two independent variables, how many results are there to evaluate?

Answer 11

Two main effects and one interaction effect

Question 12

Main effect

Answer 12

The overall effect of one independent variable on the dependent variable, averaging over the levels of the other independent variable. You take averages across the row/column to get 2 marginal means.

Question 13

How is it estimated how large each main effect is?

Answer 13

The difference between the means is calculated, and the confidence interval is calculated to determine if it’s statistically significant.

Question 14

How are interactions estimated from a table?

Answer 14

Start by computing two differences (make sure to go in the same direction both times. This differences can indicate if the results are statistically significant- need to determine if the differences are different from each other.

Question 15

When looking at a graph from a factorial design study, how do you know that there’s an interaction?

Answer 15

The lines aren’t parallel. For bar graphs, imagine drawing a line to connect the bars of one variable, and see if these lines are parallel or not

Question 16

Are interactions or main effects most important?

Answer 16

When a study shows both a main effect and an interaction is almost always more important. Main effects generally don’t give the whole picture. For example, the overall difference (main effect) for body weight hides the fact that body weight influences aggressive behavior only when men are drunk. Most answers to questions in daily life reflect an interaction.

Question 17

In a cell chart, where do you find the interaction and main effect?

Answer 17

The main effect is the difference between the marginal means in a specific row/column. The interactions are found by looking at the cells with the means for each condition. These are measurements for the dependent variable.

Question 18

Independent groups/between subjects factorial designs

Answer 18

Both IVs are studied as independent groups. A 2 times 2 independent groups factorial design has 4 groups/cells- if you have a total of 100 participants, 25 would be in each condition. Different groups of participants experience each condition

Question 19

Within groups factorial designs

Answer 19

Both IVs are manipulated within groups. Only one group of participants in all 4 groups/cells- the same participants are exposed to all 4 conditions at separate times. Order effects are a potential issue- this can be resolved with counterbalancing

Question 20

Mixed groups factorial designs

Answer 20

One IV is manipulated as independent groups and the other is manipulated within groups. This design is intermediate between the within groups design and the independent groups design in terms of number of participants. With 100 people, 50 people experience 2 conditions, and the other 50 people experience 2 separate conditions. For example, one group of participants is asked to shop alone with a purpose, then the next day they shop with a purpose with another person. A separate group shops alone without a purpose, then the next day shop with someone without a purpose.

Question 21

How do the number of cells increase as levels are added?

Answer 21

2 times 3 factorial design (6 cells/conditions), 3 times 4 factorial design (12 cells/conditions). The X format is the number of levels of each independent variable.

Question 22

How do the number of cells change as the number of independent variables increase?

Answer 22

The format remains the same. If each variable has two levels, it’s 2X2X2 (8 cells)

Question 23

How many results are you concerned with in a 3 way design?

Answer 23

Three main effects (one for each independent variable), three separate 2 way interactions and a 3 way interaction. Main effects only test one independent variable at a time, so we would only mention one at a time.

Question 24

What does it mean for a 3 way interaction to be significant?

Answer 24

This means that the two way interaction between two of the independent variables depends on the level of the third independent variable

Question 25

How is a 3 way interaction detected graphically?

Answer 25

The graph must show 2 different 2 way interactions. There must be 2 sets of lines that intersect/aren’t parallel in different ways. If the two sets of lines intersect but look exactly the same (in terms of shape and variable represented) there is not a 3 way interaction

Question 26

How can you tell if an empirical journal article used a factorial design?

Answer 26

It will be mentioned in the methods section, and X format will be used. The results section will mention whether the main effects and interaction sections are significant.

Question 27

How can you tell if a news article is talking about factorial design?

Answer 27

Look for “it depends” or “only when” when talking about an interaction between 2 factors. You can also look for any participant variables like age, personality, gender, or ethnicity. In these stories, the participant variable often moderates another independent variable. When there is a moderator, there must be an interaction and a factorial design to test it.