within participants building, null model, p value

how do we measure the difference for a within-participants study? (test statistic)

calculate how much each individual has changed and then average the individual change

if the null is true what source of variation of population is there

differences due to the normal everyday differences in individuals (random)

what does the null hypothesis say in a within participation design? (using the pizza example)

the two happiness values that a participants gave us were not related to the pizza at all, they were just random variations in the person. They would have reported those two values even if we switched the order of pizzas

if we track the average individual differences that we calculate across the two conditions as we perform this randomization over and over we …; the results we would expect to see if there’s no actual difference between the two conditions

build the null distribution

describe what the null distribution would look like

• symmetric

• centered on 0

• variability depends on variability of the original data

T/F: in order to be able to reject the null hypothesis, we need the differences created by our experimental manipulation to be relatively large as compared to the normal differences within the population

True

if the random variability in the data is small compared to the difference between groups, the result is more likely to be significant

good indicator that it would be a smaller p-value

if the random variability is large compared to the difference between groups, the result is less likely to be significant

really noisy → need much larger diff between groups

in order to have a significant result, the differences due to your experiment must be …

large compared to the individual/random differences between members of your population