Pergamon 0031-9384(95)00142-5 Physiology & Behavior, Vol. 58, No. 5, pp. 883-889, 1995 Copyright © 1995 Elsevier Science Inc. Printed in the USA. All rights reserved 0031-9384/95 $9.50 + .00 Menstrual Synchrony: Only in Roommates Who Are Close Friends? LEONARD WELLER, .1 ARON WELLERt, AND OHELA AVINIR:~ *Department of Sociology, Bar-Ilan University, Ramat-Gan 52900, Israel, E-Mail: F42153@BARILAN, ~'Department of Psychology, Bar-llan University, Ramat-Gan 52900, Israel and ~'Department of Sociology, Bar-Ilan University, Ramat-Gan 52900, Israel Received 18 April 1995 WELLER, L., A. WELLER AND O. AVINIR, Menstrual synchrony: Only in roommates who are close friends? PHYSIOL BEHAV 58(5) 883-889, 1995.--This study examined menstrual synchrony among roommates residing in 20 university housing units. Menstrual synchrony was examined for roommates who were close friends and roommates who were not close friends, and by housing unit. The impact of personality, social-interaction and menstrual-related factors was also studied. To avoid limitations found in previous studies, (i) data were collected prospectively and (ii) over the entire academic year; (iii) high participation rate was obtained (86%); and (iv) women who used oral contraceptives were excluded. Menstrual synchrony was found among roommates who were close friends. Synchrony was not found for roommates who were not close friends, nor by housing units. The phenomenon of menstrual synchrony may more likely occur among close friends and women with intensive social contact than under the conditions common to university dormitories. Dormitories Friends Personality Pheromones Menstrual synchrony Roommates Self-monitoring Women MENSTRUAL synchrony, the convergence of the onset dates of the menstrual flow, was first documented by McClintock some twenty years ago among women living in a college dormitory and among close friends (22). The majority of the research since then has continued to examine menstrual synchrony among women living together (9 studie:~ of women living in dormitories and 4 studies of women living in private residences, three of which were comprised of lesbian couples) and among close friends (7 studies). In addition, two studies examined menstrual synchrony among women who work together (21,33) and two studies exam- ined synchrony between mothers and daughters (30,31). For a period of almost 20 yr the accumulated research, almost unequivocally, continued~to confirm the original findings (22). It is not uncommon for the literature to refer to the "McClintock Effect." In the last three years the picture has changed drasti- cally. One reason has been the reporting of negative findings (27,33,36), in addition to the previously reported negative find- ings (8,17) plus the reanalysis of one of the classical studies (24) whose findings have now been shown not to have demonstrated synchrony (37, p.579). But by far the most serious challenge has been the contention (37) that McClintock's procedure (22) for calculating synchrony was flawed, so that neither she nor the subsequent researchers utilizing her approach actually demonstrated synchrony. McClintock's (22) procedure, followed by almost all other researchers who collected menstrual data over time, was to compare the initial onset differences for pairs to their final onset differences a number of months later. She found that the mean onset absolute difference was 9.2 days in October and 6.6 days in March. Wilson (37) claims that there were three basic errors inherent in McClintock's (22) data analysis. Specifically, he avers that McClintock (22) and those following her procedure failed to recognize that about one-half of the pairs of a sample synchro- nize by chance; that they had miscalculated the absolute differ- ences between initial menses onsets; and excluded from the sample subjects or some onsets of subjects to fit the number of onsets to the specifications of the research design. All three errors are seen as increasing the probability of finding a statistically significant level of menstrual synchrony in a sample. After reanalyzing data from three major articles (11,22,24), Wilson (37) concluded that none had provided evi- dence for menstrual synchrony. Thus none of the studies measur- ing synchrony over a period of time have clearly supported the existence of the phenomenon. Another approach to studying synchrony has been developed by us in a series of studies (29-31,33). In this procedure we study women who have been together for a considerable period, 1 To whom requests for reprints should be addressed. 883