Practice quiz answers – final exam - fall 2007 si

Practice Quiz Answers – Final Exam - Fall 2007 SI

1. Error Mean Square is based on mean of group variances
T/F
2. In one way ANOVA, when the F value is substantially greater than 1 it means thatthere is a treatment effect or some sort of difference between the compared groups
T/F
3. In an ideal world for any researcher you are testing the effects of one treatment on a group or organismsand there is no demonic intrusion or gradient effects of any kind, which statistical test would you use to ensure that you have the best power possible that will truthfully report any significantdifference of these treatments? Hint: if there is no gradient effect (sunlight, moisture, etc) then you should probably avoid blocking your experimental design.
Answer: 2-sample t-test
4. I want to testthe effects of Vitamin E on skin so I recruit 50 participants and put lotion with Vitamin E on one arm and a placebo on the other arm. Which statistical test should I use to analyze my results?Answer: Paired t-test
5. In ANOVA, why is the statistical hypothesis a one-tailed and the scientific hypothesis a two-tailed?
Answer: The scientific hypotheses compare the means of groups and thestatistic hypotheses looks specifically for treatment effect which is anything greater than 1.

6. Fill out this hypothetical RB ANOVA table:
Source DF SS MS F P
Treatment 2 500 250 16.129 0.00
Block9 400 88 5.677
Error 18 100 15.5
Total 29 1000
How many total experimental units were there? 30
How many treatments? 3
How many blocks? 10
Was there a treatment effect in thishypothetical experiment and why? Yes, F » 1.0

7. My experiment has a sample size of 72 mung beans, 4 experimental groups and 6 blocks since I suspected a moisture gradient in the room. What are the degreesof freedom in this experiment? Is my experiment replicated, explain? What statistical test should I use to analyze my data and why? What implications do I have to account for if there was no...