pathnottaken
5-Year Member
- Joined
- Jan 28, 2013
- Messages
- 189
It's been 30 years since I took statistics but doesn't the sample have to be random in order for the n=30 or greater heuristic to be valid? If the two women who failed weren't random female marines, but a subset of the larger cohort, then the sample size argument is irrelevant.
Thanks DHinNH for addressing this. The samples have to random and independant. The population that the samples are pulled from is also improtant (normal and continuous). For my analysis I ASSUME (yes all stats make an --- out you an me) That the repsonse variable (female and male success) is from single population that is bimodal (pass, fail). The data set sampled has a 12% failure rate. Comparing the results 2/2 and 12/108 faliure/total sample size of group rejects the hyposis that the two group are different....
I bring this up because one concern that can be asked is the endurance of females compared to males and how endurance crittical to many missions. I tryied to find data that compared females to males in terms of endurance. The data I found was from the WS100 mile 2012 race (316 racers finished the 100 mile running race - 52 female and 264 males) When looking at the finish times of the two groups there is NO statistical differnce between the male and female finishing times (women - average time 24 hours 15 minutes, males 24 hours 43 minutes)
I can remember when I was in high school there was still a debate if women should be allowed to run long distances as they "were" not made for it. So much for that debate!!!