

Log transform then z-score
Hi, new to stats. I am doing linguistic structure work on 4chan threads where post rate is an IV. because different boards move at different speeds i am z-scoring post rate. But when plotting the z-scored post rate and the DV, I got what looked like a hyperbola. After log transforming them, I get a weak linear relationship. Because you can’t log a negative, I log the original raw post rate then z-score. the first image is the raw scores and the second is with post rate logged then z-scored and the DV logged.
I am wondering if this is completeley wrongheaded or okay. thanks.