Within subjects design Flashcards
In terms of variance your regression assumes that
There is no variance on your x
Another assumption that no longer applies here is
Independence of data
Composite variables
Are another form of aggregation
When is it good to use a compound scale
To deal with poor properties of a likert scale
Simple form of aggregation assumes that______
each question contributes equally – they do not!
Issues with power analysis
Rules of thumb, problematic because it requires you have an idea of how big an effect you are going to see, how much noise
Why are rules of thumb different depending on the field?
Because they depend on the effect size and expected variance
This is not a within subjects design
A cross sectional design (is between)
Is a way of treating pre-post data
Using a differential score (careful with Lord’s paradox)
Lord’s paradox
You have to do this to compensate for random assignment missing here
Counterbalancing
Counterbalancing
You make sure all your subjects experience all conditions in a random order
This is the ideal form of counterbalancing
Full counterbalancing
ABC…
ABC, ACB, BAC, BCA, CAB, CBA
Reverse counterbalancing
Middle levels never experienced as first or last
Any two levels always experienced sequentially
AB—-BA
ABC—CBA
Latin square or William Latin square
Another type of couterbalancing
Every level is presented in every order
ABC
BAC
CBA
Is the only way to get an A after B half the time and B after A half the time
A latin William’s square not the standard
Simpliest repeated measures design
No predictors/IVs, but each person is measured repeated times
Now the average performance is the intercept
In a within subjects design the subject
Subject experiences multiple levels of the continuous predictor
Random effects
Each subject has its own slope and intercept. random because each subject was selected randomly
