Hypothesis tests of Means Flashcards
when is a paired samples t-test used
if 2 groups are paired eg 2 measurements on same individual at different times
when is an independent samples t-test used
when the population means are from 2 independent groups
usage of one-sample t test
to test whether the mean of an interval scale variable in a sample is consistent with a given value for the mean in the population
assumption in a one sample t test
when sample size is small, sample must be drawn from a population that is normally distributed. checked using histogram
what is the hypothesis tested in one-sample t-test
H0=population mean = a given value
H1 population mean does not = a given value
when is an independent-samples t-test used
to test whether the mean of an interval scale variable in one group is equal to the mean in another group when the 2 groups are independent
assumptions made in an independent-samples t-test
- when sample size small, variable of interest should be normally distributed
- std deviation in the 2 populations is similar
what hypothesis is tested in independent-sample t-test
h0 is that the mean in pop1 = mean in pop2
h1 is that mean in pop1 doesnt equal mean in pop2
how is p value obtained in independnent-sample t-test
test statistic is compared against t-distribution (n1+n2 -2) degrees of freedom if the sample sizes are small or the N(0,1) distribution if sample sizes are large
when is a paired samples t-test used?
to test whether the mean of an interval scale variable in one group is equal to the mean of an interval scale variable in another group when there exists a pairing between groups
what assumptions are made when using a paired samples t-test
when the sample sixe is small - required that the within-pair differences must be drawn from a population which is normaly distributed
what hypothesis is tested in paired sample t-test
h0 is that the mean in pop1= mean in pop2
h1 is mean of pop1 doesnt equal pop2
how is a p-value obtained in paired sample t-test
test statistic is compared against the t distribution with n-1 degrees of freedom if the sample size is small or the n(0,1) distribution if large
