week 10: t tests: single sample and dependent means

Question 1

Z tests –> t tests

Answer 1

when population variability is known (σ2 or σ), we can use Z tests

Question 2

when to use a t test

Answer 2

in the absence of population data we must estimate population parameters from sample data
n-1 instead of n

Question 3

Estimating the population standard deviation from the sample data

Answer 3

the estimate of the population variance from sample data is S2
and, therefore, the estimate of the population standard deviation from sample data is S

Question 4

Degrees of Freedom (df)

Answer 4

n-1

Question 5

Estimating the standard deviation of the comparison distribution

Answer 5

Sm=S/power of N

Question 6

How do we know it is a one-sample t test?

Answer 6

Need to compare sample mean (M) to a population mean (μ)

We need to estimate the population standard deviation (σ) from sample standard deviation (SD)

Question 7

The comparison distribution

Answer 7

t distribution instead of norma distribution (changes shape depending on degrees of freedom)

Question 8

t table appendix

Answer 8

the first column of the table lists the degrees of freedom (df)
the table repeats the 3 columns of cut-off scores for commonly used significance levels (.10, .05, .01) twice across the page for both one-tailed and two-tailed tests

Question 9

The t test for dependent means (repeated measures)

Answer 9

often we don’t know the population mean

Question 10

Confidence Intervals around the Mean

Answer 10

When population σ known

When estimating σ from sample data
M +/− (tcrit)(SM)

Question 11

repeated measures t test

Answer 11

when we have two scores on the same variable for each person and want to test whether there is a systematic difference (before vs. after; two different treatments)

Question 12

matched pairs t test

Answer 12

sometimes we match individuals across a range of variables and then only test them once