> 4. There is such a thing as girl power when we talk about email
The data presents a correlative link between gender and email activity, but not a causal link. (I would expect that the user's profession, which likely has a biased gender ratio, to be the more pertinent factor).
Additionally, the phrase "they have a higher response rate, of 15% compared to 13% for men" does not make it immediately obvious if it's a statistically significant difference. (at n=5000, it may be, but it would be better with the math present)
As a whole, the analysis could use confidence intervals, because email activity in particular has a very skewed distribution between normal users and power users.
3 comments
[ 6.5 ms ] story [ 13.7 ms ] threadThe data presents a correlative link between gender and email activity, but not a causal link. (I would expect that the user's profession, which likely has a biased gender ratio, to be the more pertinent factor).
Additionally, the phrase "they have a higher response rate, of 15% compared to 13% for men" does not make it immediately obvious if it's a statistically significant difference. (at n=5000, it may be, but it would be better with the math present)
As a whole, the analysis could use confidence intervals, because email activity in particular has a very skewed distribution between normal users and power users.