A researcher wants to increase the generalizability of a planned ...

[shofmannx20]

A researcher wants to increase the generalizability of a planned experimental study's results. What can the researcher do, relative to sampling, that will achieve that goal?
 
  a. Select a sample larger than that recommended by power analysis.
  b. Purposively select a sample that is extremely heterogeneous.
  c. Use random assignment.
  d. Use random sampling.

Question 2

The project team in intervention research failed to operationally define the study variables correctly. Which subsequent actions of the team were affected? (Select all that apply.)
 
  a. The research question posed failed to address the problem accurately.
  b. Results were not meaningful to clinicians.
  c. The methods used to study the problem were inappropriate for intervention research.
  d. The results could not be discussed effectively, in terms of the hypotheses.
  e. A type III error was not avoided.

[jesse.fleming]

Answer to Question 1

ANS: D
Random sampling increases the extent to which the sample is representative of the target population. Good representativeness enhances generalizability of the results. Selecting a sample larger than the size recommended by power analysis results only is a larger sample, not a more representative one. Purposive selection of a more heterogeneous sample does not increase the sample's representativeness. Use of random assignment has no effect on representativeness of the sample, since it is selected before assignment occurs.

Answer to Question 2

ANS: A, B, D, E
The initial focus of the team is to clarify the problem or issue of interest. The project team must be alert to the risk of making a type III error. A type III error involves asking the wrong questiona question that does not address the problem of concern. This error is most likely to occur when the researchers do not thoroughly analyze the problem and, as a result, have a fuzzy or inaccurate understanding of the issue of concern. The solution, then, does not fit the problem. A study conducted on the basis of a type III error provides the right answer to the wrong question, leading to the incorrect conclusion that the newly designed intervention will resolve the problem.

[shofmannx20]

Excellent

[scottmt]

Gracias!