hi..

i just did some question about regression. the data are:

x: 2 3 4 5 6 7 8 9 10
y: 1.7 1.1 0.5 0.7 0.2 -0.3 -1.2 -0.9 -1.8

y values are in log.

i calculated the SS reg and SS total, but i got smaller value of SS total. i guess this happened because the negative value in y. so, can SS error have negative value in the ANOVA. if so, what does it mean? how can we say about the negative in SS error?

thank you..

I think there is no error in ANOVA. Rsq is .94863. But coeff. b is negative indicating inverse relationship of y & x.