Menstrual Cycle Length as Function of Age |
Treloar, A.E., Boynton, R.E., Behn, B.G., & Brown, B.W. (1967), "Variation of Human Menstrual Cycle through Reproductive Life," in the International Journal of Fertility, 12(1), 77-126.
Looking at this graph, we can see that there is more variability in cycle length for women who are younger than 20 or older than 40 (breakpoints are the blank spaces to the right and left of center). Or, as the graph itself is labeled, between "menstrual years 0-6” and "pre-menopausal years" -8-0.”
The Bedouin study discussed in the previous section included women "between 13 and 50 years of age." If we assume that in the graph above, the variability for menstrual years 0-6 and pre-menopausal years -8-0 is NOT due to pheremonally-induced menstrual synchrony, and that a similar long-term study of Bedouin women would result in a similar graph, how does this affect the interpretation of the Bedouin study data?
In order to have valid results in the Bedouin menstrual synchrony study, would it be necessary to EXCLUDE women before the age of 20 and after the age of 40? If not, how do you distinguish between variability (changes in cycle length) due to exposure to the pheremones of other women, and variability due to non-pheremonally-induced hormonal fluctuations, more common in women at the beginning and end of their menovulatory lifetimes? Should some kind of “variability coefficient” be introduced into the analysis of menstrual cycle onset data?
Another question: Does the theory of menstrual synchrony in and of itself allow one to construct a graph of human menstrual cycle variability? That is, is the graph above what the theory of menstrual synchrony would predict, or is there some other variability distribution across the menovulatory lifetime that the theory of menstrual synchrony would predict?
Okay, so maybe the above graph is tougher than 92 tape measures! Let’s turn our attention now to the next point of contention: the random groups method.