Chapter 11: Factorial Designs

Question 1

Why do researchers design experiments that include more than one independent variable?

Answer 1

because it creates a more REALISTIC situation compared to a single factor design because behavior is usually influenced by a variety of different variables acting and interacting simultaneously.

Question 2

factors

Answer 2

when two ore more independent variables are examined in one study

Question 3

levels

Answer 3

number of conditions for each factor

Question 4

factorial design

Answer 4

a research study involving two or more factors (IVs).

Question 5

Describe an example of a factorial design that uses quasi independent variables.

Answer 5

a factorial design that invovles variables such as age or gender that are not manipulated are called quasi-independent variable factorial designs.

ex/ studying how new video game violence (factor A) and gender (B) are related to aggressive behavior (dependent variable). The gender is a quasi-independent variable because it cannot be manipulated.

Question 6

What kind of factorial design is 2 x 3? How many conditions are there?

Answer 6

this is a 2-factorial design in which there are 2 independent variables with three levels each. There are 6 conditions in this type of factorial design

Question 7

example of a 2 x 2 factorial design

Answer 7

ex/ schizophrenic and control group, who experience placebo or drug treatment. You get four scenarios.

schizophrenic and placebo
schizophrenic andd drug

normal and placebo
normal and drug.

Question 8

What is a main effect

Answer 8

seeing how each individual factor (IV) influences behavior/measure (DV). MEAN DIFFERENCES among the levels of one factor. ex/ differences between the row means define the main effect for the row factor.

the main effects reflect results obtained if each factor was examined on its own.

Question 9

What is an interaction

Answer 9

seeing how the group of factors (both IV’s) act together.

Question 10

if factors work independently, then there is ___ interaction (yes or no)

Answer 10

no interaction.

Question 11

in terms of main effects, how can you tell if there is an interaction between the two independent variables?

Answer 11

if the main effects of one factors doesn’t apply EQUALLY across the mean conditions of the 2nd factor.

Question 12

when graphing results, non parallel lines between factors indicates an ____, meaning that the factors are ____

Answer 12

when graphing results, non parallel lines between factors indicates an INTERACTION, meaning that the factors are DEPENDENT ON ONE ANOTHER.

Question 13

A gap in the graph indicates

Answer 13

that there is a main efffect in the y variable

Question 14

a slope in the graph indicates:

Answer 14

that there is a main effect in the x variable.

Question 15

if the two variable lines in the graph intersect right at the mid point, does this mean there is a main effect in the x axis?

Answer 15

NO! the two slopes cancel each other out. There is actually no change in means.

Question 16

what are the three separate sets of mean differences that a 2 factor study allows researchers to evaluate?

Answer 16

1) the mean differences in the main effect of factor A
2) the mean differences from main effect of factor b
3) the mean differences from the interactions between factors.

Question 17

What is a mixed factorial design?

Answer 17

factorial design where one independent variable is a between subjects design and the other is a within-subjects design.

Question 18

example of a mixed factorial design

Answer 18

ex/ looking at the type of words recalled based on mood state. researchers get either happy or sad people (between subjects design) and make them both listen to positive and negative words (within subjects design because the same group of people listen to pos and neg words, not just one type of word). researchers then measure word recall in both happy and sad groups to see how many positive and how many negative words are recalled. thus, they wanted to see how type of word (independent variable A) and type of mood (independent variable B) affects word recall.

Question 19

Pros and cons of between subjects design? when should you use a between subjects factorial design?

Answer 19

pros: no order effects because conditions can be tested all at once
cons: needs a large amount of participants, individual differences can be a confounding factor.

when to use:

1) when you have a large population
2) when individual differences are small
3) order effects are likely

Question 20

Pros and cons of a within subject design? when should you use a within subjects design?

Answer 20

pros: only one group of participants, fewer participants
- assured equivalence, no individual differences

cons: time consuming, high drop out rates
- testing effects, difficulties to counter balance

when to use:

1) when you have a small population
2) when individual differences are high

Question 21

What is a combined strategy factorial design

Answer 21

when 2 different strategies in the same factorial design are implemented.

1) one factors is a TRUE INDEPENDENT VARIABLE (experimental design)
2) another factor is quasi-experimenta; (preexisting characteristic like gender), or TIME!!!! (not manipulated)

Question 22

example of a combined strategy factorial design

Answer 22

ex/ researchers are measuring performance levels of individuals with high and low self esteem in a room with an audience vs no audience. Audience vs no audience is a controlled variable (experimental), where high vs low self esteem is a pre-existing characteristic within the population they sampled from (quasi-experimental, NON EQUIVALENT, NON EXPERIMENTAL)

Question 23

When is a pretest-post test control group design considered a combined strategy desgin?

Answer 23

if a researcher has one sample of participants and can RANDOMLY ASSIGN them to the two groups.
R-OXO
R-O O

recall: originially, a pretest post test non equivalent control group design was originailly classified as quasi-experimental designs because the two groups are non equivalent and preexisting, they were not randomly selected.

Question 24

if both factors in a 2 x 2 factorial design are a between subjects, what kind of statistical test can you use for analysis?

Answer 24

independent measures 2-factor ANOVA

Question 25

if one factor in a 2 x 2 design is a between subject design, what kind of statistical test can you use for analysis?

Answer 25

a MIXED DESIGN 2 factors anova

Question 26

When both factors in a 2 x 2 factorial design are within subjects, what kind of statistical test can you use for analysis?

Answer 26

a repeated measures ANOVA TEST

Question 27

Applications of factorial designs

Answer 27

1) explaining and replicating a previous study
- how would the treatment effects differ if they were administered in a different environment?

2) reducing variance in a between-subjects design.
- involves using the specific variable that is acting as a confound as a second factor, thereby creating a 2 factor study. 2 factor designs allow variance to be reduced without sacrificing external validity.

Question 28

how could you measure order effects using a factorial design?

Answer 28

usie counterbalancing to measure order effects. the different treatment condition groups are the differences in treatment order.

ex/ 2 x 2 design
group 1 has treatment 1 and then 2
group 2 has treatment 2 and then 1.

the 2 different treatments are a WITHIN subjects design because both treatments are being experienced with the same group of individuals (group one gets both treatment 1 and 2)

the different groups are a between subjects design. (group 1 is different than group 2).

Question 29

if you are using a 2 x 2 factorial design to measure order effects, what can you expect the data to do if there is no order effect?

Answer 29

the data will show a pattern with NO INTERACTION,

Question 30

if there are symmetrical order effects, what can you expect the data to do?

Answer 30

there will be an interaction. treatment 1 influences treatment 2, and treatment 2 influences treatment 1.

Question 31

If there is a non symmetrical order effect, what can you expect the data to do? what does this tell about the data?

Answer 31

the graph will be lopsided and will intersect at a point that isn’t in the middle. The different treatments may produce different levels of practice. ex/ treatment 1 produces a 5 point different on treatment 2, but treatment 2 only produces a 1 point difference to treatment 1

Question 32

A two factor mixed design

Answer 32

aka a pretest post test control group design, which is a type of quasi experimental design that involves two groups, one control and one treatment THAT ARE NON EQUIVALENT. IF these two groups are assembled via random assignment, it is also considered a combined strategy factorial design ( one experimental factor and one quasi)