Part A
A researcher is conducting a study to determine if a newly developed drug, X, is an effective treatment for treating Basal cell carcinoma (i.e., a common kind of skin cancer). The researcher collects a sample of 2,000 participants from the population of people suffering from Basal cell carcinoma. As part of this study, the researcher must formulate a null and alternative hypothesis and select an alpha (i.e., significance) level for the study.
First, formulate a null and alternative hypothesis for the researcher. Use the proper symbology for null and alternative hypotheses in your response (H0 and H1, respectively).
Next, select a significance level for hypothesis testing (i.e., .05 or .01).
What are the pros and cons of the significance level you selected (i.e., use the terms Type I and Type II error in your response)?
Part B
The researcher has selected a parametric test to run the data analysis in order to determine if drug X is an effective treatment for Basal cell carcinoma.
What are some assumptions you can make about the researcher’s data distribution based on this information? Discuss at least three assumptions as part of your response.
Part C (see note below about G*Power software)
The researcher must determine the statistical power for his design.
With G*Power, use the level of significance, you chose in part 1 of this assignment (i.e., .01 or .05) and the sample size (i.e., 2,000) to calculate the statistical power for this study.
Assume a medium effect size for this calculation (i.e., .5).
Based on your calculation, does the researcher have good statistical power? Why or why not?
How will statistical power effect the researcher’s results?
Indicate whether you are conducting a pirori or post hoc statistical power test, and explain why the statistical power test you are conducting is a pirori or post hoc.
Part D
The researcher found that drug X is a significantly effective treatment for Basal cell carcinoma. In part C of this assignment you were asked to assume a medium effect size.
What is an effect size?
How does having a medium effect size affect the researcher’s interpretation of her findings (i.e., is having a significant result meaningful in this case)?
How would the interpretation change if the effect size was large?
How would the interpretation change if the effect size was small?
G*Power software
For Part C of this assignment, you will need to download G*Power software (i.e., this is free software) onto your computer. Here is the link where you can find G*Power software: http://www.gpower.hhu.de/en.html
You must scroll down on the page until you find the correct version of G*Power software (i.e., mac or windows).
To effectively use G*Power to complete week 3 assignment C: In G*Power, there are three pull-down menus, I have listed them below as well as what selections you will need to make.
Test Family (menu): F-tests (selection)
Statistical test (menu): ANOVA: Repeated Measures, within factors (selection)
Type of power analysis (menu): Post hoc: Compute achieved power – given α (alpha), sample size and effect size (selection) The statistical power (i.e., the answer to the question) is located in the output parameters (bottom half of screen, right side), as the last value in that column: Power (1 – β (beta) err prob).
These instructions will make much more sense once you see the G*Power screen. If you have any questions, please feel free to contact the instructor.
