It is an error related to variable lengths differing. I have read previous forums in regards to this but I am not too sure if they apply to me. I have 6 independent variables and TCMD as my dependent variable and a samples size of 10. I am trying to undertake Akaike Criterion Information for my data using:
library("MASS")
Peak_data <- read.table(text="GSP ACPEN SOLPEN NOH ASP EPI TCMD
361 0 0 2497838.8 11276.06 79.7 14828.45
369 0 0 2471221 14200.75 86.9 14483.62
375 0 1 2497838.8 15756.69 100 12445.13
384 0 1 2524456.6 19674.65 119.3 13060.37
392 0 1 2551074.4 15613.37 123.5 13000.04
402 0 1 2577692.2 14292.03 114.5 13308.07
410 0 1 2604310 21585.35 109.3 14038.92
418 0 1 2630927.8 34850.07 120.8 15258.51
421 1 1 2657545.6 30050.26 136.6 13933.98
431 1 1 2684163.4 30969.31 135.2 14236.64
",header = T)
full <- lm(y~., data = Peak_data)
Error in model.frame.default(formula = y ~ ., data = Peak_data, drop.unused.levels = TRUE) :
variable lengths differ (found for 'GSP')
I get this error preventing me from performing backwards AIC using:
stepAIC(full, direction='backward')
Any help is much appreciated!
Thanks in advance!
Your code syntax is correct but the dependent variable naming is wrong. You can use the following code
library("MASS")
full <- lm(TCMD~., data = Peak_data)
stepAIC(full, direction='backward')