Chapter 10 Flashcards
2 ways to manipulate an IV
between subjects: Different groups of people, each group exposed to a different condition.
within subjects: One group, each person is exposed to each condition (but at different times)
advantages to within subjects
Within subjects is more powerful than between subjects design: (More likely to reject Ho if IV really does have an effect)
Less “noise” (extraneous variability) in the data b/c same subjects are tested twice.
disadvantages to within subject design
However, “carry-over” effects are possible with within subjects designs. (e.g., practice effects or carryover effects such as with different drugs…
denominator of t test
The average difference between the means of two samples drawn from the same population. the estimated standard error
why do we pool variances?
combine variances of both groups and it will weigh the variance according to sample size
assumptions about independent samples t test
he variability of the samples must be roughly equivalent.
Homogeneity of the variances
(Also, as w/ one sample t-test: interval or ratio data, and normal population or large n if pop not normal)
Hartley’s F max test
Test for homogeneity of variance
Large F-max value indicates a large difference between the sample variances
Small value (near 1.00) indicates similar sample variances and that the homogeneity assumption is reasonable
what do you do if the sample information suggest violation of homogeneity of variance assumption
calculate standard error
what is r squared
Proportion of variance accounted for by group/ treatment/ IV.
Range from 0-1. 0=There is no variance that is accounted for by the treatment. 1=all of the variance is accounted for by the treatment.
when do you use paired t
1.) You have two means from the same group of people:
Before/after scores of the same people
IV manipulated within participants
statistical hypothesis for paired t
Ho: μ D= 0 (there is no difference between means)
Ha: μ D = 0 (there is a difference between means)
confidence interval
the range of values that you expect your estimate to fall between a certain percentage of the time if you run your experiment again or re-sample the population in the same way.
