Chapter 7: Hypothesis Testing and Inferential Statistics

Question 1

Null versus Alternative Hypothesis

Answer 1

Null hypothesis is the one that is actually tested (inferential stats/f and t tests). Alternative hypothesis is typically the research hypothesis. If null is rejected, alternative is accepted.

Question 2

2-tailed versus 1-tailed hypothesis tests

Answer 2

2-tailed is non-directional and is basic/theoretical research. 1-tailed is directional and hypothesis tests should generally be used in applied research.

Question 3

Type 1 versus Type 2 errors in hypothesis testing

Answer 3

Type 1 (alpha) error: if the sample data is a fluke and there really is NO EFFECT in the population
Type 2 (beta) error: if the sample data is a fluke and there really IS AN EFFECT in the population

Question 4

What traits does a significant effect have?

Answer 4

A large test statistic and a small p-value indicates a significant effect, and the conclusion is that the null hypothesis is likely false and should be rejected

Question 5

What is inferential statistics?

Answer 5

T-test and f-test; can be used to evaluate whether numerical difference is statistically significant. Test statistic is compared to a “critical value” to determine statistical significance. Significant p-value will be small (p<.05).

Question 6

What is descriptive statistics?

Answer 6

Involves reporting a measure of central tendency and/or measure of variability for each treatment condition

Question 7

When to use a 1 tailed or 2 tailed hypothesis test?

Answer 7

-if the interest is whether the treatment conditions (e.g. group 1 and 2) differ significantly from one another that is a non-directional or 2-tailed test
-if the interest is whether one particular treatment condition (e.g. group 1) is significantly better than the other (group 2) that is a direction or 1-tailed test

Question 8

Are different statistical tests performed for non-directional and directional tests?

Answer 8

The math to compute the t and f statistics are the same, only differences is (1) whether the test is even done and (2) how the test results are determined to be significant or not. In a directional test I need to test if no numerical difference.