November 22, 2012.
Lecture 9 - Student’s T-test for Single Samples
General Procedure for Testing a Null Hypothesis
Applying General Procedure for Testing a Null Hypothesis
Calculate an appropriate statistic based on your sample’s data, study design, and your
o Sign Test: sum of pluses
o Z Test: obtained mean obtained Z score
o T Test: obtained mean obtained t
Choose the appropriate null hypothesis sampling distribution that corresponds to your statistic.
o Sign Test: Binomial Distribution
o Z Test: sampling distribution of the Mean (one group and know population mean and
o T Test: sampling distribution of t (one group and DON’T know population SD, or have
Decide whether the statistic is likely to have come from a population modeled by the null
hypothesis sampling distribution
o Sign Test: calculate p value (probability of getting that # of pluses if only chance
o Z Test: Calculate p value (probability of getting that mean if only chance is operating)
o T Test: calculate p value (probability of getting that mean or mean difference if only
chance is operating)
Types and Uses of t Test
o Compare one group mean to a known population mean when you don’t know SD
o Compare one group mean to some other null value of interest (e.g., scale midpoint)
o Compare two group mean in independent groups experiment
o Compare two group means in repeated measures experiment
Correlated groups, not independent of each other
Using a Single Sample t Test
o How much do people think other UBC students are attracted to risky behaviours like
drinking, partying, having unprotected sex?
How much do you think the typical UBC student thinks partying is attractive?
o Scale of 1 – 6, 6 = very attract