the grade only
LAB 7 Factorial Design Experiment
Lab 7
¨ Think of encountering someone on the street or meeting someone for the first time… ¤ What are the first few features you look at to decide
who this person is?
Lab 7
¨ In our experiment… ¤ Every participant in our sample, on each trial:
n Saw an image of a n Male or female n Whole face visible or upper face visible or lower face visible
n Categorized gender of person in image n Rated confidence of categorization
Lab 7
¨ Our experiment… ¤ Explore how we categorize faces as male and female
n What visual information about a face is most important?
¤ Our independent variables? n Actual gender of face in picture
n How many levels?
n Part of face visible in picture n How many levels?
¤ What’s the name of this factorial design? n 2 x 3 within-subject design
¤ Dependent variables n Accuracy of categorization, Confidence of categorization
Lab 7
¨ How do we categorize faces? ¤ Which part of face do we use to categorize? ¤ Does this differ for male faces vs. female faces?
1. Eyes 2. Mouth 3. Whole
1. Female faces # correct/8 # correct/8 # correct/8
2. Male faces # correct/8 # correct/8 # correct/8
Eyes Mouth Whole
Female faces confidence confidence confidence
Male faces confidence confidence confidence
Lab 7
¨ Hypotheses ¤ Main effect:
n Whole face easier to categorize than partial face n H1: µ whole face (male or female) > µ lower face or upper face (male or female) n H0: µ whole face (male or female) ≤ µ lower face or upper face (male or female)
n Upper face easier to categorize than lower face (Campbell, Benson, Wallace, Doesbergh, & Coleman, 1999)
n H1: µ upper face (male or female) > µ lower face (male or female) n H0: µ upper face (male or female) ≤ µ lower face (male or female)
Lab 7
¨ Hypotheses ¤ Interaction:
n Lower face easier to categorize in male pics than female pics (Armann & Bulthoff, 2012)
n H1: µ lower face male > µ lower face female n H0: µ lower face male ≤ µ all other conditions
IDing Interactions: Graphs
¨ An interaction occurs when the lines on a graph of the results are not statistically parallel
YES: Crossover Interaction
IDing Interactions: Graphs
¨ An interaction occurs when the lines on a graph of the results are not statistically parallel ¤ Is it an interaction?
YES: Spreading Interaction
IDing Interactions: Graphs
¨ An interaction occurs when the lines on a graph of the results are not statistically parallel ¤ Is it an interaction?
YES: Spreading Interaction
IDing Interactions: Graphs
¨ An interaction occurs when the lines on a graph of the results are not statistically parallel ¤ Is it an interaction? ¤ More different the
slopes, the more likely there’s an interaction
0
5
10
15
20
25
Silence Mozart
NO: Lines appear nearly parallel
IDing Interactions: Graphs
¨ An interaction occurs when the lines on a graph of the results are not statistically parallel ¤ Is it an interaction?
YES: Crossover Interaction
IDing Interactions: Graphs
¨ Interaction: when the lines on a graph of the results of an experiment are not statistically parallel ¤ If any parts of the lines are non-parallel, there is likely
to be an interaction ¤ HOWEVER, this is only a rule-of-thumb
n Must run statistics to determine whether or not an interaction is present or not
Lab 7
¨ Method ¤ Include all three subsections
n Participants: summarize number of participants, basic demographic information (number of male, female participants, average age)
n Materials: describe online experiment materials in detail n Procedure: describe the experiment from the researcher’s
perspective (what did participants experience) and be sure to include the factorial nomenclature
Lab 7
¨ Results ¤ What do you predict (hypothesis)? ¤ Let’s look first for interaction first (it’s most important)… ¤ Let’s look for main effect of gender… ¤ Let’s look for main effect of part of face visible… ¤ Report all three, whether significant or not
n “There was (was not) a significant interaction between ______ and ______, F(__, __) = ___, p = (<) ___.”
n “There was (was not) a main effect of..., F(__, __) = ___, p = (<) ___.” Do this for both possible main effects
¤ Describe the results in plain English ¤ Include line graph as Figure 1
Lab 7
¨ Title Page ¨ Abstract ¨ Introduction ¨ Method ¨ Results ¨ Discussion ¨ References ¨ Figure