It's been a long, busy time. I hope I am now back to my once a week posting- both in terms of advancing and in terms of article reading!
I finished the two statistics courses, and the lesson learnt is: What you get out of a study depends on who does the statistics. It is really easy to massage data by taking out outliers. It is easy to get effects by choosing which test you use. And since so much is up to the discretion of the researcher, it is hard to be able to be critical enough to not trust the study, and not cynical enough to be able to reach conclusions that make sense. Or, reworded- throwing out research is the easy solution, but not a smart one.
I have great news- I began an algorithms course, and then I applied (and won!) a two month scholarship to Dataquest. Dataquest is a website that teaches data skills in R and Python. It's a great system, where you can study for ten minutes at a time. I decided to apply to a diversity scholarship (for women etc) on a whim, and was really lucky to get it. I now have two months to finish as much as I can. And yet, I want to stick with the algorithms course. The trick is to see it as a positive challenge, but it is fun, and I am enjoying it thoroughly.
On to the physical part. This week's article (Beneficial associations of low and large doses of leisure time physical activity with all-cause, cardiovascular disease and cancer mortality: a national cohort study of 88,140 US adults by Zhao et al) is another how much exercise should we do article. They took a large cohort, examined 88,000+ adults between 40-85, examined how many died in later surveys of cardiovascular disease and cancer-related deaths, and then calculated a hazards ratio and all that stuff. They found that even 10-59 minutes of moderate activity (read:walking) a week is enough to reduce risk (18%, no less). And the more you do, the stronger the effect, to a point. 10 minutes a week! That is going to the bathroom time. That's nothing. There are a number of possibilites:
1. This is absolutely true.
2. This is a poorly done study where the preconceived notions dictated the results.
3. Something in between. Well, obviously.
The 18% comes from the 0.82 hazard ratio- hazard rate (dying at a certain time) of those doing exercise/those not. 0.82 means smaller risk. And the exercise questionnaire was given only initially-over 36,000 did NO activity. How could people do no activity? 10 minutes a week! This is the problem with questionnaires.
All the studies I've looked at show trends- just be a little active. But the results they show seem highly simplistic. And don't take into account population differences.
