Blog Front Page

Frederic

Charting: Temperature Rise and Pregnancy - Stats Study

This post is a follow-up on my previous post. I'll present below a few more results concerning the size of the temperature rise with respect to pregnancy.

Among the numerous feedback we received, a few interesting issues were raised. I'd like to discuss it below.

Lopsided Histograms
Some readers, well trained in stats, noticed that the histograms departed from the normal distribution which can skew any test of hypothesis that would assume a normal distribution. One reader suggested to run a Kolmogorov-Smirnov test instead of the simple t-test that we ran. I just did that but unfortunately it was again not really conclusive (p value of 0.16).

Temperature Scaling
Some readers pointed out that comparing temperatures may not be the best because everyone has a different metabolism. They suggested to compare the difference between pre-ovulation and post-ovulation temperatures instead. We ran a study to measure the temperature rise.

You'll find below histograms showing the rise for both pregnant and non pregnant charts (500 charts in each group).

Not Pregnant Temperature Rise

Mean: 0.6646
Standard Deviation: 0.1710

Pregnant Temperature Rise


Mean: 0.6652
Standard Deviation: 0.1733


We also ran a hypothesis test (Kolmogorov-Smirnov and t-test) to find out if there is a statistical significance to the difference. We did not find anything conclusive.

Conclusions
The main point that these two studies reveals is that there is no real information about the chance of pregnancy in either the raw value of early post-ovulation temperatures or even in the size of the temperature rise. This is quite important for those who may be worried about these aspects of their charts. Some readers suggested to look at temperature further along in the cycle. We expect that we would probably would find a significant difference because of the embryo implantation (check also the triphasic pattern study we ran previously). Other readers suggested that we may want to look at the difference between pregnant cycles and not pregnant cycles for a given individual. We will certainly look at this in later studies. Stay tuned.

Feel free to send us as usual your comments and feedback as well as suggestions for further studies/articles.

Permanent Link: http://www.fertilityfriend.com/blog/posts/221.html
Posted: Aug 18, 2009 10:02am Eastern Time by Frederic, Founder.

FeedbackShare



IMPORTANT: Text, data, graphics, graphs and photos presented on this page are intended for personal use only. Any other usage as well as any derived work (papers, studies, re-prints, reproductions on web sites...) is prohibited without explicit written consent from Tamtris Web Services Inc., owner and operator of FertilityFriend.com.

FertilityFriend.com is a trademark of Tamtris Web Services Inc. All other trademarks are the property of their respective owners.