Factorial design Flashcards
So far we’ve only explored the relationship between one independent variable and one dependent variable, this is rather artificial. In reality there is for than one variable affecting the outcome. What is this called?
A factorial design: explores how multiple factors influence a specific outcome (multi cause)
ex: happiness influenced by the circumstances born into(parents and family, time and place..), outer world (home, work, country…), inner world (us - belief, feelings, emotions…). likely these all interact to either decrease or increase level of happiness
tu vois genre si tu regarde seulement un facteur ok admettons home you grew up in, résultats vont pt montrer que happiness scores decrease as environment is poor. Mais what if you had a positive outlook on life? Positive outlook + home = different outcome for happiness. This is what factorial design looks into
What are two things we can examine with factorial design?
- The effects of more than one IV on a given DV
- The interactions between the IVs
when there is more than one IV, how are IVs referred to as?
Factors (hence the term factorial design)
What’s one main characteristic of factorial design?
High ecological validity.
External validity is the umbrella term for generalization.
Ecological validity is one piece of external validity, specifically about how real-life-like the study is. Basically, how well does it reflect real life.
example of factorial design and how it relates to ecological validity
e. g. poor exam outcome:
possible factors:
-lack of sleep
- bad mood
- not studying enough
- tricky questions
- noise level
you wanna determine which are responsible for the outcome
What is a single-factor design study?
You examine impact of each IV one at a time. Doing so would require a new set of participant for each study and it wouldn’t tell us anything about interaction of IVs, as opposed to a factorial design.
What can a factor design do?
Looking at various factors all in the same, single study!!!! one big arrow rassemble tous les facteurs.
regarde screenshot iPad
If you have a 2 factor factorial design, how many questions are you answering?
for example: effects of sleep time and study time on exam performance.
you can answer whether sleep time (2 levels/ treatment conditions: sufficient; insufficient) affect exam performance.
You can answer whether study time (2 levels: sufficient; insufficient) affect exam performance
and you can assess interaction between sleep time and study time!
How does a factorial design matrix look like?
Each factor is denoted by a letter (A, B, C…) and each level is represented by a number (1, 2, 3…)
so for example :
A1; A2 (factor A study); A1(sufficient study time)….
B1 ; B2 (Factor B sleep)
One factor on the vertical axis, one factor on the horizontal axis.
regarde screenshot iPad.
What do the cells of the matrix represent ?
Combination of each factor (their interactions with each other!!!)
- COULD BE CODED
ex:
A2B1: insufficient study time with sufficient sleep (and then see how does that affect dependent variable - exam performance)
- COULD BE IN WORDS NO CODES, so just the factor combinations : sufficient sleep and sufficient study in one cell…. Bref you write them out in letters, words.
matter of personal preference, both correct
Is there a specific convention for naming factorial designs? Explain
Yes. The number of digits indicates the number of factors; and the actual value of each digit indicates the number of levels for each factor
2 x 3 (2 factors; factor A - 2 levels ; factor B - 3 levels)
x = BY, you cant pronounce it as X
2 x 2 meaning?
How do you pronounce it?
2 factors, within each 2 levels
ex. study (sufficient; insufficient) and sleep (sufficient; insufficient) .
Pronounced as: 2 by 2 factorial
2 by 2 by 3 (2x2x3) not only tells you about how many factors you have; and how many levels each one has, but also…
how many cells your matrix will have
How many cells does a 2 by 3 factorial design has?
6 cells (do the multiplication!) so 6 combinaisons
How many cells does a 2 by 2 by 3 factorial design has?
12 cells
Lets say you first had a 2 by 3 factorial design. One factor being trait anxiety and the other test anxiety. And now you wanna add sex of participant. How would that look in a matrix?
2 by 2 by 3 = 12 cells
6 cells for each sex
so
A1B1C1 (ex: male participants with high test anxiety level and high trait anxiety level)
A1B1C2
A1B1C3
A1B2C1
A1B2C2
A1B2C3
sameeee thing but instead of A1ça serait A2
bref tu catégorises tes participants dans la combinaison qui les fit!!!! (une autre utliité des matrix, autre que voir interactions entre facteurs!!!!)
Type of information that can be obtained from factorial designs
- Main effects: treatment differences between levels of a factor. In other words, the individual effect of each factor.
main effect for factor A (effect of factor A on DV)
main effect for Factor B (effect of factor B on DV)
- Interaction effects (huge reason why you’re conducting a factorial design in the first place):
- Combined effect of factors when the effect of one factor DEPENDS ON another factor
A x B interaction effect
(what are you factors? what’s the outcome? outcome depends on…. telle ou tell combinaison…. )
Whats a one two-way interaction?
A one three-way interaction?
A three two-way interactions?
A x B
A x B x C (one three-way interaction - une possible de trois interactions)
A x B ; B x C ; A x C (3 possibilities de 2 interactions !)
There is a hypothesis for…
Each main effect and each interaction effect.
How to interpret results for factorial designs? (FOR MAIN EFFECTS!)
Always from your matrix….
if you have a 2 by 2 factorial design:
Step 1: Look for the main effect for A, and that is for each level!!!! (effect of factor A on dependent )
compare column means for the two levels. If means are identical, no main effect (cause the effect of one treatment condition is identical to the other!)
ex lets say factor A is mood when studying. 2 levels: happy or sad. ideally you would like two see a difference between your two levels! or else it would mean the mood during learning doesn’t matter for exam performance.
Step 2: Look for the main effect for B, and that is for each level!!!! (effect of factor B on dependent )
compare column means for the two levels. If means are not identical, you can conclude that there is a main effect for mood during recall.
How to interpret results for factorial designs? (FOR INTERACTIONS!)
Check if difference in row means is consistent across columns.
if difference is the same = no interaction effect
d = difference
What does it mean if there is no interaction effect?
ex:
interaction A x B
If there isn’t an interaction effect in a factorial design, it means that the effect of factor on the dependent variable is consistent across all levels of the other factor. In simpler terms, the two factors influence the outcome independently of each other.
interaction effect occurs when the effect of one factor on the dependent variable depends on the level of another factor.
In a graph, if the lines are parrallel, what does it mean?
That there is no interaction.
M majuscule means ?
MEAN score
