It's been a few years since my 3 stats courses, so pardon if I'm wrong, but aren't his p-values for the logistic regression kind of... important?
Intercept and log(hits) sure, but the rest have rather high p values indicating that it's likely to be present even if it has no effect (lower p-value, more robust, < 0.05 => < 5% chance of incorrectly stating it has an effect (when it does not)).
Intercept and log(hits) sure, but the rest have rather high p values indicating that it's likely to be present even if it has no effect (lower p-value, more robust, < 0.05 => < 5% chance of incorrectly stating it has an effect (when it does not)).
Or am I missing something basic here?