What model to use for a 'spraying experiment'?

Question

I separate my plants into four groups of 16, and I spray them to keep the bugs away:

group 1: I spray daily
group 2: I spray every 2 days
group 3: I spray every 3 days
group 4: I spray every 4 days

at the end of the 12 days, I check number of plants that's bug free

group 1 is: 12
group 2 is: 10
group 3 is: 7
group 4 is: 6

How should I fit a model to this?

Answer 1

First you should get your data into R! so create a variable for your response (number of bug free plants - or you might want to consider having this as a bug free proportion of the total), and a variable for your factor (ie sprays per day is probably the best way).

Then plot those two variables. There's probably an obvious relationship. The trick then is to somehow model that...

If your response is a proportion the best model is probably some form of logistic regression. You should also think carefully about the best form of your explanatory variable; with most ways of going about modelling, you would like this to have a linear relationship to some transformation of the response variable. Read up on logistic regression, and good luck!