Hey so I'm trying to perform a power calculation for a longitudinal study. I've been working with the longpower package. I was a bit confused about the meaning behind the pct.change parameter in the lmmpower command when I was trying to calculate sample size for an nlme model. So, for example, in the following command what does the .3 represent.
lmmpower(model.3, pct.change = .3, t = seq(1,7,1), power = 0.80)$n
The package writeup lists it as "the percent change in the pilot estimate of the parameter of interest (beta, the placebo/null effect)" but am having trouble understanding it. If someone could explain it with a simple example I'd really appreciate it. Also not sure if this belonged here or on cross validated so sorry if it doesn't.
Whenever you do a power calculation, you need to specify an effect size, i.e. the magnitude of response that you're expecting from your experiment. This is usually the smallest effect you would reasonably expect to be able to detect, and an effect size below which you would be comfortable concluding that the effect probably wasn't important in terms of your subject area (biology, economics, business, whatever ...)
lmmpower allows you to specify either pct.change or delta; probably the reason that pct.change is emphasized is that it's often reasonable/easier to interpret proportional changes in an effect. Among other things, this makes the values of the parameter independent of the scale (units) on which the response variable is measured. Alternatively you can specify delta, which is the change in absolute units (i.e., the same units as the predictor variable is measured in).
For what it's worth
"the percent change in the pilot estimate of the parameter of interest (beta, the placebo/null effect)"
seems a little puzzling to me too; I would add "as a proportion of" before "beta" in the parenthetical clause.