Author Question: A researcher has already written a study proposal and performed a power analysis, which indicates ... (Read 103 times)

WhattoUnderstand · **on:** Jul 8, 2018

A researcher has already written a study proposal and performed a power analysis, which indicates that the sample size will have to be at least 36 if a type II error is to be avoided.

The researcher then decides to set the level of significance at p <.01 rather than p <.05 but wants to maintain power at 0.8. What will the researcher have to do in order to maintain sufficient power?
a. Increase the effect size.
b. Set beta at 0.60.
c. Use a different statistical test.
d. Increase the sample size.

Question 2

A researcher studying depression in relation to five predictor variables decides to include five more variables in the study because the literature reveals that they may also be involved in depression.

Consequently, the researcher will do which of the following?
a. Use a larger sample size.
b. Set beta at a lower level.
c. Perform additional reliability and validity tests.
d. Change the design of the research from correlational to descriptive.

rekilledagain · **Reply #1 on:** Jul 8, 2018

Answer to Question 1

ANS: D
You determine the sample size needed to obtain sufficient power by performing a power analysis. Power analysis includes the standard power (usually 80), level of significance (usually set at 0.05 in nursing studies), effect size, and sample size. In order to demonstrate statistical significance at a more stringent level (a decreased alpha), the sample size must be increased. The effect size is determined by the magnitude of effect that the intervention produces and, consequently, cannot be adjusted.

Answer to Question 2

ANS: A
As the number of variables under study grows, the needed sample size increases. Adding variables such as age, gender, ethnicity, and education to the analysis plan can increase the sample size by a factor of 5 to 10 if the selected variables are uncorrelated with the dependent variable. Setting beta lower, performing additional reliability and validity tests, and changing the design from correlational to descriptive will not solve the problem of the altered power of the study, due to addition of five variables. Only increasing the sample size will be a practical solution.