Question Answer
what are some reasons for different scores of people on baseline readings? individual differences, researcher error and chance factors
what are chance factors? factors which can have an effect on results but cannot be mitigated for (I.e. rained on one day, not the other)
what is the collective name for individual differences, researcher error and chance factors? these are known as experimental errors.
what is another name for experimental error? within group variation
what are the reasons for between group variations? treatment effect and experimental error
what is F ratio? the statistical test to see the effect of our treatment
what is the formula for F ratio? between subjects variation over within subjects variation.
what would we expect the result of F to be if there was no treatment effects we would expect F to be 1, or pretty close to it.
what is the null hypothesis? that there will be no treatment effects.
what is alpha level? alpha level is another word for p
what is a type one error? that we have rejected the null hypothesis wrongly.
what are the assumptions which must be met before we can do between subjects anova? assumption of normalityassumption of independenceassumption of equal variance
violation of which assumption invalidates a between subject design? independence
what is the assumption of normality? this assumption that the dependant variable is normally distributed in each of the conditions/groups.
what is the assumption of independence? that individual observations are independent within groups and between groups.
what is the assumption of equal variance? (homogeneity of variance) that the experimental error (I.e. within group variability) is approximately equal in each group.
what test can we use to see if homogeneity of variance is met? F max test
how do we calculate F max test? Fmax = largest variance over smallest variance. If if the results are greater than 3 than a correction must be made to alpha level. The usual way is to divide alpha by 2.

Leave a Reply

Your email address will not be published.