13.07.2015 Views

Contents

Contents

Contents

SHOW MORE
SHOW LESS

You also want an ePaper? Increase the reach of your titles

YUMPU automatically turns print PDFs into web optimized ePapers that Google loves.

CHAPTER 6: Independent Groups Designs 197experimenter, but this is not essential. What is essential is that entire blocks betested at each level of the extraneous variable, which, in this case, is the fourexperimenters. The balancing act can become a bit tricky when there are severalextraneous variables, but careful advance planning can avoid confoundingby such factors.Key ConceptKey ConceptSubject Loss We have emphasized that the logic of the random groups designrequires that the groups in an experiment differ only because of the levels of theindependent variable. We have seen that forming comparable groups of subjectsat the beginning of an experiment is another essential characteristic of the randomgroups design. It is equally important that the groups be comparable exceptfor the independent variable at the end of the experiment. When subjectsbegin an experiment but fail to complete it successfully, the internal validity ofthe experiment can be threatened. It is important to distinguish between twoways in which subjects can fail to complete an experiment: mechanical subjectloss and selective subject loss.Mechanical subject loss occurs when a subject fails to complete the experimentbecause of an equipment failure (in this case, the experimenter is consideredpart of the equipment). Mechanical subject loss can occur if a computercrashes, or if the experimenter reads the wrong set of instructions, or if someoneinadvertently interrupts an experimental session. Mechanical loss is a lesscritical problem than selective subject loss because the loss is unlikely to berelated to any characteristic of the subject. As such, mechanical loss should notlead to systematic differences between the characteristics of the subjects whosuccessfully complete the experiment in the different conditions of the experiment.Mechanical loss can also reasonably be understood as the result of chanceevents that should occur equally across groups. Hence, internal validity is nottypically threatened when subjects must be excluded from the experiment dueto mechanical loss. When mechanical subject loss occurs, it should be documented.The name or subject number of the dropped subject and the reason forthe loss should be recorded. The lost subject can then be replaced by the nextsubject tested.Selective subject loss is a far more serious matter. Selective subject lossoccurs (1) when subjects are lost differentially across the conditions of theexperiment; (2) when some characteristic of the subject is responsible forthe loss; and (3) when this subject characteristic is related to the dependentvariable used to assess the outcome of the study. Selective subject loss destroysthe comparable groups that are essential to the logic of the random groupsdesign and can thus render the experiment uninterpretable.We can illustrate the problems associated with selective subject loss by consideringa fictitious but realistic example. Assume the directors of a fitnesscenter decide to test the effectiveness of a 1-month fitness program. Eightypeople volunteer for the experiment, and they randomly assign 40 to each oftwo groups. Random assignment to conditions creates comparable groups atthe start of the experiment by balancing individuals’ characteristics such asweight, fitness level, motivation, and so on across the two groups. Members

Hooray! Your file is uploaded and ready to be published.

Saved successfully!

Ooh no, something went wrong!