Displaying 1 result from an estimated 1 matches for "seregf".
Did you mean:
seref
2004 May 27
1
Getting the same values of adjusted mean and standard errors as SAS
...;M"]]
seRegM <- sqrt(mean( ddr$res[ddr$Gender=="M"]**2 ))
sxxM <- sum((dd$Age[d$Gender=="M"]-meanAgeM)**2)
syM <- seRegM * sqrt(1/nM + (meanAge-meanAgeM)**2/sxxM); #0.1103335 cf
0.11032602 - matches to 5 decimal places
nF <- summary(ddr$Gender)[["F"]]
seRegF <- sqrt(mean( ddr$res[ddr$Gender=="F"]**2 ))
sxxF <- sum((dd$Age[d$Gender=="F"]-meanAgeF)**2)
syF <- seRegF * sqrt(1/nF + (meanAge-meanAgeF)**2/sxxF); # wrong:
0.07279221 cf 0.14256466
> dd
Measurement Age Gender Group
1 3.8 94 M 3BCLP
2 2....