ANALYSIS OF VARIANCE (ANOVA) Flashcards
what is the normal distribution?
data is symmetrically distributed with no skew.
what are parameters?
for the normal distribution they are constant values that determine the shape and position of the distribution
Mean, μ
the average - determines where the peak of the distribution is
Standard deviation, σ
the variability - determines the spread of the distribution
what do we assume for ANOVA?
samples have a normal distribution
what percentage is +/- SD’s?
68%
what percentage +/- 2 SD’s?
27%
what percentage is +/- 3 SD’s?
5%
assuming two normal distributions with the same variance what is the null hypothesis?
it’s true when the two population distributions are mostly (ideally, entirely) overlapping
H0: 𝜇1=𝜇2
assuming two normal distributions with the same variance what is the alternative hypothesis?
it’s true when the two population distributions have minimal overlap
H1 : 𝜇1≠𝜇2
what must you ask when rejecting H0?
- Given our uncertainty about the means (μ1 and μ2) of two populations, how likely is it that the observed difference between their estimated (sample) means (𝑥̅_1−𝑥̅_2) is due to chance?
- accepting H1, would be suggesting that the observed difference between the sample is not due to chance
what is an experimental error?
refers to all uncontrolled sources of variability in an experiment
what are the 2 types of experimental error?
- measurement error = the measuring device used is not properly calibrated or bias
- individual differences error = the samples they comprise, will have naturally occurring differences that are practically very difficult to control, typically more extreme when small sample sizes are used.
what does between group variation estimate?
estimates both experimental error and treatment effects
what does within-group variance estimate?
estimates experimental error only
