Exam 8: Assessing Measurement

Correct Answer:

Verified

Reliability refers to the stability of the relationship between observed values and the events that actually occurred. One way to do this is as a byproduct of assessing accuracy. That is, when comparing observed and true values, it is easy to also track the nature of any error. If the consistency of the error is relatively high, for instance, the data might be described as acceptably reliable. Of course, this would have nothing to do with whether the data were acceptably accurate. This way of calculating the reliability of observations involves the same methods for assessing accuracy.
A second way of assessing the reliability of observed values does not require true values. In this approach, the behavior of interest must be preserved in some way, such as in a videotape format. With this advantage, the observer can be presented with the videotape two or more times so that the consistency of observational judgments can be evaluated.

Correct Answer:

Verified

One of the advantages of conceding that data are only estimates of the facts is that it encourages researchers to try to obtain the best estimates possible. This approach helps minimize the risk of basing conclusions on data that are tainted with substantial error. The possibility of error also confronts researchers with the obligation of determining the extent to which error is present in the data, in spite of their best efforts.

Correct Answer:

Verified

There is no simple rule to guide this decision. The decision partly depends on the complexity of the measurement procedures. If measurement involves observing only a single behavior that is clearly defined and easily detected, the data may not need to be assessed very often. On the other hand, if observers must detect multiple behaviors in multiple participants in a complex setting, it may be wise to evaluate the data fairly often. More frequent assessment may also be necessary when observers must also record the status of treatment conditions or behavior of individuals other than participants who are involved in key ways. The frequency with which data assessment is conducted also depends on what these evaluations reveal. That is, the better the assessment outcomes, the less often such assessments may need to be made. On the other hand, if assessment data show certain kinds of problems, it is best to be cautious and conduct assessment frequently.