You are on page 1of 19

Analyzing

Data
PSY 106
Fall 2021
Descriptive Statistics
• Central Tendency Measures
• Mean, Median, Mode
• Standard Deviation- how far from the ‘average’
Decision Rule for Levene’s Test

• Helps to ensure that the distributions of the


outcomes for each independent group are
comparable and/or equal. If groups are not
equal you can have spurious findings. Should
be conducted on outcomes between groups
that have similar variances.

• If p<= .05 the variances are significantly


different. Interpret the BOTTOM row of results
for t.
• If p> .05, the variances are not significantly
different. Interpret the TOP row of results for t.
Test of Homogeneity of Variance with
ANOVA
• Helps to ensure that the distributions of the outcomes for each
independent group are comparable and/or equal. If groups are not
equal you can have spurious findings. Should be conducted on
outcomes between groups that have similar variances.
• IF p > .05 you have met assumption of homogeneity of variance and
can proceed to conduct your One-Way ANOVA
• If p < .05, you have violated the assumption of homogeneity of
variance and will not be able to conduct a One-Way ANOVA.
• Kruskal Wallis test
Between subjects

• Divided into group/conditions


• Results are compared between the
Between different groups
Subjects vs. Within Subjects
Within
• No division
Subjects • Exposed to all conditions in study
• Results are analyzed/compared
WITHIN each participant (pre/post) to
examine the difference
Provide us the relationship
between two variables

r = .xx

Correlations
-1.00 - +1.00

Strength of the relationship


between variables
T-tests

• t-test for independent means (independent samples t-test)


• Comparing two groups on an outcome
• E.g., Males and Females on rates of anxiety
• Statistical significance: p < .05
• Practical Significance: Strength of relationship between your variables.
• Effect size: d=.xx
• Null hypothesis
• Accepting and Rejecting
• Independent Samples t-test vs. Paired Samples t-test
Anatomy of a t-test for Independent Means
Factorial Design
PSY 106

Garcia
Factor- independent variable (IV) that causes the
outcome or changes the dependent variable.

Factorial design is used when you have more

What is a than one independent variable.

Factorial We look for the interaction between those IVs.

Design? Experimental design that allows us to establish


cause and effect

Allows us to examine the impact of multiple IV's


on the DV
Anatomy of a Factorial Design

• Depending on the number of independent variables included in the study the factorial design can look very
different.
• Ex. 2 x 2 x 2
• The numbers indicate the number of independent variables and the number of levels (conditions) within it.
• So, for example you want to study the effect of being in a relationship and class standing on student’s academic
grades.
• You would have a 2 numbers indicating the number of IV’s (class standing and relationship)
• You would have 2 (or more) levels of relationship (i.e., in a relationship, not in a relationship)
• You would have 4 levels of class standing (i.e., freshman, sophomore, junior, senior)
• So, it would be a 2 x 4 factorial design
# of IV’s each indicating one IV and its levels
2x4 Factorial Design
IV: Class Standing (4 levels)
Freshman Sophomore Junior Senior
Freshman in Sophomore in a Junior in a Senior in a
IV: Relationship Status (2

In a Relationship Relationship Relationship Relationship Relationship

Not in a
Relationship Freshman not in a Sophomore not in Junior not in a Senior not in a
Relationship a Relationship Relationship Relationship
levels)

DV: Academic Grades


Sometimes we may have an IV that
the researcher has no control over
such as gender (which cannot be
manipulated) and this becomes a
Hybrid hybrid design

Design
Limits the ability to make a causal
conclusion
• An interaction in this design refers to the ability to
determine if the influence of one IV on the DV is changed
depending on the level of another IV.
• For example:
• You have a study that wants to examine the impact of
following a diet and exercise plan has on weight.
• So, a 2 (diet, no diet) x 3 (yoga, running, weightlifting) design
• That is two IV’s (diet and exercise plan)

Interactions • BUT…when you add in exercise (levels indicated by the


type of exercise) that may also impact/change the
relationship between diet and weight, and we call that
an interaction. (interaction effect hypothesis)
• Hypotheses:
• Main effect: Being on a diet will lead to lower weight.
• Interaction effect: Those who are on a diet and have a
running exercise plan will have the lowest weight. Etc..
• You may have more hypotheses based on the
number of combination with levels of IV’s.
2 x 3 Factorial Design

Exercise (3 levels) Second Independent Variable


First Independent Variable

Yoga Running Weightlifting

Diet Diet/Yoga Diet/Running Diet/Weightlifting


Diet (2 levels).

No Diet No Diet/Yoga No Diet/Running No Diet/Weightlifting

One interaction, but each cell is a


different interaction
• When using a factorial design, the
Two-way statistical test used to examine the
relationships is a Two-way ANOVA.
Analysis of • This test allows us to include the multiple
Variance IV’s and multiple levels of the IV’s.
• It tests for any main and interaction effects
(ANOVA) • Takes all combinations of levels of IV’s
Vignettes

A description of a
hypothetical This is one way that
situation or even to researchers can
which participants manipulate an IV
react.
Enjoyment of Party Crossover Interaction

Introverts
Extroverts

Small party Large party

You might also like