Significance and effect size in treatment design when treatment-response in measured

Question

I have a number of plants and all of them undergo a 'treatment'. Then I measure the amount of plants that react to the treatment and the amount that do not. Is there a way to calculate a test statistic (to test for the significant effect of treatment)? If yes, is there a way to infer an effect size measure?

E.g.: R code:

n_total <- 120
n_response <- 100
n_no_response <- 20

Or would it be appropriate to use McNemar test in this case in the following way and the use effect size measure as described here:

tab <- cbind(c(n_total, 0), c(n_response, n_no_response))
mcnemar.test(tab) 

Answer 1

As you've described it, with no control group, & assuming plants react independently, the population effect size would simply be the proportion that respond to treatment. Estimate it by the proportion that respond in the sample & calculate an appropriate confidence interval. NB the crucial assumption here is that whatever you're seeing that leads you to say some plants are reacting to treatment couldn't have come about for any other reason.