Chapter 9: Experimental Designs: Within Subjects Designs

Question 1

a _____ design uses a SINGLE group of participants for all the treatment conditions. The single group and thus each participant is exposed to all levels of the IV

Answer 1

a within subjects design

Question 2

A within subjects design is the ultimate ____ groups because the group in one treatment condition is absolutely identical to the group in every other condition

Answer 2

a within subject design is the ultimate in EQUIVALENT GROUPS

Question 3

a within-subject design is also known as a ____ because repeated measurements of the same individuals under different conditions are being compared

Answer 3

aka a repeated measures design.

Question 4

T/F there is no need to control extraneous variables in a within-subjects group design

Answer 4

false, extraneous variables still need to be controlled for

Question 5

Pros and cons of a within-subjects group design

Answer 5

pros

1) requires relatively fewer participants in relations to a between subjects design
2) ELIMINATES individual differences as a CONFOUND because the participant is the same in all conditions
3) Eliminates high variance: each individual will serve as own control/baseline and therefore possible to remove variance caused by individual differences.
4) it is possible to measure differences between participants without involving individual differences. ( when the individual differences are consistent across treatments, they can be measured and removed from variance in the data)

cons:

1) series of treatments are administered at different times and thus time related factors may act as confounds or act as a threat to internal validity
2) participant attrition: not everyone will complete the study, causes a SHRINK IN SAMPLE SIZE AND AN EXAGGERATION OF VOLUNTEER BIAS

Question 6

T/F individual differences act as a threat to internal validity in a within-subjects design

Answer 6

false; there are no individual difference-confounds because the same participant is being used amongst all treatment conditions

Question 7

what’re the 2 major sources of potential confounds?

Answer 7

1) environmental variables ( time of day, room differences between treatments
2) time-related variables ( factors that may influence the individual of the course of period of study)
- includes history, maturation, changes in instrumentation, regression to the mean, testing effects.

Question 8

What are testing effects

Answer 8

effects directly related to the EXPERIENCE obtained by participating and being measured in previous treatment conditions

Question 9

What are order effects

Answer 9

Effects caused by the fact that individuals go through a series of treatments IN ORDER, and performance in any treatment may be influenced by treatments that occurred in EARLIER conditions.

Question 10

name 3 different aspects of order effects and give examples

Answer 10

1) carry over order effects
when an order effect is caused by a SPECIFIC TREATMENT, one treatment condition produces lasting change that carries over to subsequent condition

ex/ a learning memory technique learned in an earlier treatment condition that helps with the recall in a new treatment condition

2) contrast order effects: a SUBJECTIVE condition is influenced by its contrast with a previous treatment condition

ex/ participants who enter a moderately bright room may think that it is poorly lit if they came form a brightly lit previous condition

3) progressive error order effects: general experience accumulated during the study
- practice effects: progressive IMPROVEMENT in performance as one gains experience
- fatigue: negative PRACTICE EFFECTS, progressive decline in performance as one gains experience.

Question 11

How can order effects act as a confounding variable

Answer 11

1) can produce changes from one treatment to another that are not truly caused by the IVs
2) occurs if the order effect varies SYSTEMATICALLY with treatment conditions (ex/ first treatment will always effect the second treatment)

Question 12

3 Methods of dealing with time related threats and order effects of a within-subject design

Answer 12

1) controlling time- must balance short intervals (increased order and carryover effects) with long term intervals (increased time related effects like maturation and attrition)
2) switching to a between subjects design: if expected order effects are obvious (ex/ carry over effects are obvious), a between subjects design eliminates the threats of confounding from order effects
3) counter balancing: matching treatments in respect to time.

Question 13

What is counterbalancing in respect to matching treatments to time

Answer 13

counterbalancing reduces order effects because different participants undergo different treatment conditions in different orders. DISRUPTS systematic relationship between times and order of treatment conditions and thereby eliminates confounding from TIME and ORDER RELATED effects.

Question 14

Pros and Cons of using counter balancing to mitigate order effects

Answer 14

pros; prevents order effects from accumulating in one particular treatment- order effects are spread evenly across all treatment (ex/ fatigue effects will be present in the 4th condition for some, 1st condition for others).

cons:
COUNTER BALANCING AND VARIANCE:
order effects are not completely eliminated, they ARE HIDDEN.
- this can disrupt treatment means. They may inflate treatment means
-this is bad if absolute value of the mean is important for the study.
- counterbalancing mau add the order effects of some individuals within each treatment but not to all of the individuals. Scores are thus increased within each treatment, which adds to the variance within treatments

COUNTERBALANCING AND ASYMMTERICAL ORDER EFFECTS
-one treatment might produce more of an order effect than another treatment. (ex/ one treatment may be more demanding than another treatment, so the people who experience that treatment first are more susceptible to fatigue in the next treatment than those who were not in the demanding treatment prior)

COUNTERBALANCING AND THE NUMBER OF TREATMENTS

• its hard to completely counter balance if there are a lot of treatments.
• you need to create a latin square in order to correctly partial counterbalance.

Question 15

What is partial counter balancing

Answer 15

ensures that each treatment condition at one time within the order of treatments, so that each treatment condition occur FIRST in sequence for one group of participants, second for another ,etc.

you must create a latin square in which each condition proceeds and follows each condition exactly once.

Question 16

What is a 2 treatment within subjects design? Pros and cons?

Answer 16

similar to a 2 group mean difference design in a between subjects setting, but instead there is only one group that is being exposed to only 2 different treatment conditions.

pros: easy to counter balance, increases likelihood of obtaining significant differences.
Cons: you only have 2 data points

Question 17

what kind of statistical test can be ran for a two treatment design involving interval or ration data?

Answer 17

you can compute means, run a PAIRED DIFFERENCE T TEST, can run correlation tests, or do a repeated ANOVA SINGLE FACTOR TEST

Question 18

what kind of statistical test can be ran for a two treatment design involving ordinal data?

Answer 18

WILCOX test to evaluate significant difference

Question 19

Advantages and disadvantages of a multiple-treatment design

Answer 19

pros: more likely to reveal a functional and more convincing cause-effect relationship between IV and DV
cons: more time is needed to complete treatments, increased attrition, difficult to completely counter balance because there are more than 2 treatments (n!)

Question 20

Typical statistical analysis for multiple treatment designs utilizes repeated ___ measures

Answer 20

ANOVA

Question 21

Compare WS and BS designs

Answer 21

1) individual differences are eliminated as a confound in within subjects designs, but is a problem for BS designs
2) time related factors and order effects: potential issues for a WS design, but not BS.
- fewer participants are needed in a within subjects design

Question 22

What is a matched subjects design

Answer 22

a mixture between within and between subjects designs. uses separate group for each treatment condition, but each individual group is matched one-to one with an individual in every other group to mimic a within-subjects design.

• attempt to eliminate individual differences as a confound, but are not as effective as using a WS design
• faster than a WS design because there are multiple groups.
• becomes extremely difficult as the number of matched variables increases and number of different groups increases.