Science

Skepchick Quickies 8.14

Amanda

Amanda is a science grad student in Boston whose favorite pastimes are having friendly debates and running amok.

Related Articles

29 Comments

  1. Yeah, I learned about the original MHC/birth control thing way back in undergrad, which would have been a couple years after the original study came out.

    It is interesting that they didn’t find a preference in the non-bcp women, contrary to the prior study (at least according to the wired article – I’ll have to track down the study at some point.)

    I do think that the press release statment “”Not only could MHC-similarity in couples lead to fertility problems but it could ultimately lead to the breakdown of relationships when women stop using the contraceptive pill, as odour perception plays a significant role in maintaining attraction to partners.”” is stretching things just a tiny bit though. :) I’m sure scent plays a role in mate preference at some level, but so many other things play into the question that I’d want to see a study looking at actual mate choice preferences between BCP and nonBCP using women before I’d believe that an odor preference is going to have that drastic of an effect on final mate preference.

    Now the real question – does this mean that women who are husband hunting must insist that their dates refrain from wearing cologne or deodorant to avoid any false signals? ;)

  2. Blake, thanks for pointing that out. 49 people? How on earth can any findings be statistically significant with such a small sample size? I can hear my college stats professor’s voice scoffing in my head.

  3. How on earth can any findings be statistically significant with such a small sample size?

    You could do it, but the difference you’re looking at would have to be dramatic. The more subtle the difference you’re looking for, the larger the sample size needs to be.

    The ridiculous example would be “Analysis of a random sample of 100 women reveals that a statistically significant number of women have breasts”.

  4. Here’s a non-ridiculous example of using small sample sizes.

    Check out the entry on early experimental design on cures for scurvy.

    In this case, the variable isn’t how many people recovered, but how quickly they recovered given the various treatments. It’s a sweet, sweet trick in experimental design and execution: get from a qualitative value “did they recover” to a quantitative one “how quickly did they recover”. Helps out the math tremendously.

    If he’d tried to use “did they recover” as his measurement, he’d need a lot more people in each category for any of the conclusions to be valid.

  5. A small sample size could stand as a proof of concept, anyway. It could support the case for a more indepth study with a larger sample size. Especially when different studies of small sample sizes tend toward similar results.

  6. Bah, you folks have taken this thread completely off the rails. Talking about “sample sizes” and “statistical significance”…what utter tosh.

    I came here for comments about Robo-monkeys cyberbeakflipping Bird-bots, and I ain’t leavin ’til I sees ’em :-P

  7. On one of my other forums, I saw a name for months if not years before realizing that it was probably a pseudonym. Since my mother’s maiden name is Cox and most of my cousins have that for a last name, it didn’t immediately register with me that somebody probably wasn’t really named Dixon Cox.

    So I’m a little clueless sometimes …

  8. As nice as the idea of bio-bot is – I dont think RatBot, Cat-o-Tron and MechaDog would sell very well.

    RatBot would run off and get stuck in a drainpipe somewhere, Cat-o-Tron would refuse to obey instructions and shred the curtains out of spite, and MechDogzilla would spin in circles chasing its own power cable and ocassionally leak oil on the carpet. Cyber-Chimp would be even worse as it makes loud screeching noises and flings bearing grease at you

    The proposed iShotGun would be sleek and impressive – but with no user serviceable parts and being a totally sealed unit would mean you couldnt put the cartriges in.

  9. @Lyc — You forgot, MechDogzilla would also run away three times in 24 hours, pee in the kids’ bedroom, poop on a pile of Indiana Jones themed Legos, chew up a stuffed animal, gnaw through a brand-new leash in 5 minutes, claw a window screen to pieces, chew large pieces off a wooden door frame, rip holes in a carpet, and completely destroy a pillow and a couch cushion.

    Oh wait, I mean MY dog would do that. Just think how much worse she’d be if she was a MechDogzilla.

    *sigh* A dog seemed like a good idea two weeks ago …

Leave a Reply

You May Also Enjoy

Close
Close