Exam 9: Experimental Designs: Within-Subjects Design

One general concern for within-subjects research is that participation in one treatment condition may influence an individual's scores in a later treatment condition. This creates a threat to internal validity known as order effects.

Which manipulation of the time period between treatment conditions reduces the chances of history influencing the results of a within-subjects experiment?

At a minimum, how many participants would be needed for a matched-subjects design comparing three different treatment conditions with 20 scores in each treatment?

One disadvantage of within-subject designs is that they tend to require more participants than are needed for between-subject designs.

To completely counterbalance a study, each possible order of treatments must be presented to a participant.

Provide an example of order effects in a within-subjects research study and explain why order effects are a concern for researchers.

Counterbalancing involves having different groups of participants move through the series of treatments in different orders.

For a within-subjects experiment with data measured on an interval or ratio scale, researchers typically compute a mean score for each treatment condition.

In a completely counterbalanced within-subjects experiment, ____.

A problem with counterbalancing is that it can increase variance in the data by adding order effects to some individuals but not others.

In a within-subjects research study, factors that change over time, such as history and maturation, can be threats to ____.

When one treatment condition has a lasting effect on the participants and influences their scores in later treatments, the study is confounded by ____.

If the measurement device changes during the course of a within-subjects study so that the participants' scores are influenced, internal validity is threatened by statistical regression.

Compared to a between-subjects design, the variance in a within-subjects design is greatly reduced because the ____ are measured and statistically removed.

In a Latin square, ____.

In a within-subjects experiment comparing two treatments, the researcher is concerned that the results may be confounded by outside events that occur between the first and second treatments. One way to reduce the likelihood that internal validity will be influenced by an outside event is to ____.

Within-subjects designs are most useful when individual differences are ____.

A within-subjects design would be appropriate for a research study that compares problem-solving ability for participants with three different levels of self-esteem (high self-esteem, medium self-esteem, and low self-esteem).

How many participants would be needed for a within-subjects experiment comparing four different treatment conditions with a total of 20 scores in each treatment?

Describe a matched-subjects design and explain how this design attempts to avoid the major problems that exist with a between-subject design and a within-subjects design.