University:
York University
Course:
AP/ADMS 2320 | Business Statistics
Academic year:

2022
Views:

90

Pages:

2
Author:

Niklas

two-way anova remember= SST=SSB+ SSE the experimental design is important, to avoid bias. thered main goal of different experimental design is to study effects of different treatments on a population one-way anova uses a completely randomised design: observations are assigned randomly to a specific level of treatment. -2 explanatory variables, so observations are affected by 2 factors, -if there's only one observation for any combination of treatments, there's no replications there are 2 SSB-like sums, and SSE will have contributions from both treatments -the first index is summation for first treatment Co vice versa Loeg is mean for first treatment xi is j for second treatment Ngh Log levels 9 levels for groupRS LD R= levels blocks row SSR = h ∑(x̄i.-x̄..)² no of elements column SSC = g ∑(x̄.j-x̄..)² h SSE = ∑∑(xij-x̄i.-x̄.j+x̄..)² i=1 j=1 SST = SSR + SSC + SSE dfSST = gh - 1 dfSSC = h - 1 no of blocks dfSSR = g - 1 dfSSE = (g-1)(h-1) MSR = SSR / (g-1) MSC = SSC / (h-1) MSE = SSE / ((g-1)(h-1)) Test Statistics: MSR / MSE MSC / MSE we need 2 critical values