From: gts (gts_2000@yahoo.com)
Date: Sat May 24 2003 - 11:11:03 MDT
Harvey Newstrom wrote:
> gts wrote,
>> 3) Of the ten parameters measured, only the differences in ONE of
>> them was statistically significant in all tests: HDL.
> It's not quite as obviously one-sided as you make it sound.
It is one-sided for this one parameter. That is all I'm saying. As I wrote
in my last message, the parameters are debatable.
> Remember that over 40% of the participants were confirmed to
> quit their diets and many more were believed to be poor at
> following their diets.
That is indirect evidence for rather than against the statistical
differences found between the groups. The net effect of the cheaters in both
groups would have been a wash on the means of the parameters, but their
measurements would have increased the variances of the measurements, making
the differences seem less significant.
> As such, I believe the end-of-year measurements are not as reliable as the
earlier
> measurements which showed:
>
> Atkins: Conventional Diet:
> 6 month triglycerides: +15.0% -7.6%
> 3 month LDL(bad) cholesterol:+5.4% -7.4%
> 3 month total cholesterol: +1.7% -5.4%
The first comparison above is not statistically significant with p = .13
The second comparison above is not statistically significant with = .38
The third comparison above is statistically significant at 3 months but
becomes insignificant at 6 and 12 months.
-gts
>
> The raw data is not as obviously one-sided as you imply. If
> the numbers were all good and stayed good for Atkins, I would
> agree. But there are some troubling data early on when
> adherence to the diets was better. Given these earlier
> numbers, I am not sure I would choose the Atkins diet.
This archive was generated by hypermail 2.1.5 : Sat May 24 2003 - 11:21:53 MDT