Subject Designs (Within Subject - Between-Subject Design) Flashcards
How can I manipulate my IV ?
1.Quantitively
(parametric)
2. Qualitatively
(non-parametric)
What kind of subject-designs are there?
- Between-Subject Designs
2. Within Subject Designs
What are Between-Subject Designs?
That you randomly assign participants to to different groups
-exposing each group to different levels of the IV
What are the subcategories of Between-Subject Designs?
- Single-Factor Designs
- Single Factor Randomized group designs
- Matched Group Designs
What is a Single Factor Randomized Group ?
Assigning subjects randomly to different levels of IV - forming different groups
What are subcategories of Single Randomized Group Designs ?
- Randomized Two group design
2. Randomized multi Group designs
What is a randomized two group design?
You assigning subject of your sample to different conditions of your IV.
What are the steps for randomized two group designs?
- take a sample of the whole population
- Randomly assign them to two different groups (conditions/levels of your IV)
- Comparing the mean of each group
What are Matched Group Designs(Steps )
- assessing one or more characteristics of your subjects
- you group similar subject together in one group
- Randomly assigning each participant from each group to a treatment condition
What is a randomized Multi-Group design
You add 1 or more levels of your IV –> adding more treatment conditions
= systematic variation of the amount of IV
What are the subgroups of Matched- Group Designs?
- Matched-Pair -Designs
2. Matched Multi-Group Designs
What is a Matched Pair Design?
- You asses characteristics from a subjects and match them in pairs.
- Randomly assign each person to two different groups (treatment conditions)
What is a Matched-Multi Group design?
- Assessing characteristics of your sample.
- Grouping similar subjects in same group
- Randomly assign each participant to multiple levels of your IV.
What is a main characteristic of a matched Design?
you don’t select your subject randomly
- rather on specific characteristics
What is the advantage of a randomized two group design ?
- simple
- requires less subjects
- no pretesting or categorization necessary
What are disadvantages of a randomized two group design?
- limited amount of information about the effect of the IV
- no information about the nature of the relationship
- sensitivity to the effect of the IV
- greater subject differences influence the performance on the DV
What is problem of the randomized multi-group design?
there could be a strong correlation of the subject characteristics and the DV
What is a within subject design?
that all subject are exposed to all the levels of a IV (treatment conditions)
What are advantages of the within subject design?
- subject related differences (e.g. age, gender, IQ) are identical across treatment conditions = perfect match
- related to matching designs
- performance difference not due to error variance
What are the downsides of within subject designs?
- high demand for subjects because they are exposed to all treatment conditions.
Main problem= Carry over effects
What are the two subcategories of within-subject designs?
- Single factor two level design
2. Single Factor multi level design
What is a single factor two level design`
all subjects are exposed to all levels of IV
. one half in one order
second half the other way around
= counterbalancing
What are sources for Carry over effects?
- Learning
- Contrast
- Habituation
- Sensitization
- Fatigue
- Adaptation
What are solutions to deal with carry over effects?
- Counterbalaning
- Reduce error variance
- Treatment order
What is counterbalancing
assigning the various treatments of the experiment in a different order for different subjects
What is the advantage of counterbalancing?
distribution of any Carryover effects across treatments
- no produced differences in the treatment outcome/mea
- avoid that you misinterpreted the observed changes as effects of your IV
- separating carryover effects from the effects of your IV
What are the two options for counterbalancing?
- Complete Counterbalancing
2. Partially counterbalancing
What is complete counterbalancing ?
providing every possible ordering of treatments
- at least assign one subject to each oder
What is partially counterbalancing?
- making a list of all possible orders
2. randomly choose a selected number from these order
When do you use within-subject design?
- subject differences are strongly responsible for changes in your DV (correlation with DV)
- limited number of subjects available
- carryover effects are absence or at least limited
What is error variance?
that the variability of your DV is caused by extraneous variables (subject related. e.g. gender, IQ)
What is the problem with error variance?
That you cannot control extraneous variables in the natural environment
- difficult to determine the real effects of your IV
How can you deal with error variance?
- Reducing error variance
- Increasing effectiveness of IV
- Randomize error variance across groups
What is Statistical Analysis?
That some error variance will remain
- checking if observed effect was due to your manipulation or the error variance
= using inferential statistics
