10

Question 1

Assumptions of two sample t test? 3•

Answer 1

•Both samples are random samples

•Both populations have normal distributions

•The variance if both populations is equal

Question 2

Welch’s t test?

Answer 2

Compares the means of two normally distributed populations that have unequal variances.

Question 3

When do you want to use Welch’s t test and what is the drawback? 2•

Answer 3

•When population variances are thought to be more than ~10x different (SD more than ~3x different)

•Not as powerful as 2-sample t test when variances of populations are equal. They are harder to calculate.

Question 4

!!!x^2 contingency test?

Question 5

Levene’s test?

Answer 5

Compares the variance between two or more groups.

Question 6

Realized reproductive success?

Answer 6

Assigns paternity by genotyping.

Question 7

Apparent reproductive success?

Answer 7

Assigns offspring to “social males”

Question 8

!!!Detection derivations from normality by histogram?

Question 9

Quantiles?

Answer 9

Values of the x variable that divide the range of a probability distribution into continuous intervals with equal probabilities.

Question 10

Shapiro-Wilk test?

Answer 10

Used to test statistically whether a set of data comes from a normal distribution.

Question 11

Why is Shapiro-Wilk test not recommended?

Answer 11

•at low sample size, little power to reject normality
•At high sample size, t tests are ok even with moderate deviations from normality because of the central limit theorem.

Question 12

!!!What are the options if the assumptions of Shapiro-Wilk test is not valid?

Answer 12

•if sample sizes are large and deviation from normality is not too big, parametric tests may be OK anyways.
•Transformations
•Non-parametric tests
•Bootstrapping
•!!!!!