Quick Overview: Types of Data
Continuous
Ratio: e.g., height, weight Interval: e.g., temperature, shoe size
Discrete/categorical
Ordinal: e.g., first, second, third Nominal: e.g., #93, #13
How you use a scale determines the type of data!
Example. #93, #13 on NHL jerseys is nominal, but it could be ordinal in your study if you chose to make the top player #1, and so on.
What Type of Data Will You Collect?
Questionnaires
Interval rating scales: e.g.,
Please indicate the extent to which you believe the defendant’s claim Do not believe at all 1 2 3 4 5 6 7 Believe completely
Open-ended format: e.g.,
In your own words, please indicate if, and why, you believe the defendant’s claim
Close-ended format: e.g.,
Please check the box that most closely matches why you believe the defendant’s claim No scientific evidence Eye witness testimony Alibi
How Will You Analyze Your Data?
Interval rating scales: means, continuous data T-tests, ANOVA, correlation, regression Open-ended format: many types of responses (steer clear!) – data needs to be coded for reliability Closed-ended format: frequency data Chi-square
T-tests (experimental)
Independent measures,
One IV (discrete, two levels/groups), one DV (continuous) Example: Difference between men and women in levels of implicit self-esteem
Dependent (repeated) measures
One DV (continuous, measured twice) Example: Difference in men’s implicit self-esteem levels before and after a complex problem-solving task
ANOVA (experimental)
One-way ANOVA
One independent variable (discrete), one dependent variable (continuous) Single main effect only (i.e., a difference in means between at least two levels of an IV)
Two-way ANOVA
Two independent variables (discrete), one dependent variable (continuous) Two main effects, one interaction Definition: An interaction occurs when the influence of one IV on the DV variable depends on the level of the second IV.
Correlation/Regression (no causality)
Correlation
Two continuous variables
Relationship between two variables
Regression
One or more predictors, one DV (continuous)
Used for predicting one variable from one or more other variables
Chi-square (my favourite!)
Chi-square Goodness-of-Fit test
One discrete variable – observed vs. expected frequencies
Chi-square Test of Independence
Two discrete variables – observed vs. expected frequencies
How to Enter Data
You MUST identify your questionnaires with a code Choose SHORT, but meaningful, variable names (no hyphens/spaces allowed) Add label if necessary Add values (i.e., 1 = male, 2 = female) Choose number of decimal spaces
Grouping Variable = IV
‘Define groups’ (enter 1, 2 where you have defined 1 and 2 in your dataset)
Test Variable = DV No options (e.g., Cohen’s d, charts) available
Repeated Measures T
‘Analyze’
‘Compare means’
‘Paired-Samples T-Test’
Paired Variables = pre-/post-test measures No options (e.g., Cohen’s d, charts) available
One-Way ANOVA
‘Analyze’ ‘Compare means’ ‘One-way ANOVA’ Factor = IV Dependent List = DV Post hoc (Tukey) No options ‘Analyze’ ‘General Linear Model’ ‘Univariate’ Fixed Factor = IV (do not use random factor) Dependent Variable = DV Post hoc = Tukey Plots Options: Estimates of effect size
Two-Way ANOVA (factorial)
‘Analyze’
‘General Linear Model’
‘Univariate’
Fixed Factor = IVs (do not use random factor) Dependent Variable = DV Post hoc = Tukey Plots Options: Estimates of effect size
Correlation
‘Analyze’
‘Correlate’
‘Bivariate’
Variables = measures Default = Pearson
Use Spearman for ranked/ordinal data