Is the month of Ramadan marked by a reduction in the number of suicides?

For Muslims the month of Ramadan is a time of fasting but during the evenings after sunset it is also an occasion for family and social gatherings. Therefore, according to the BertillonDurkheim conception of suicide (that is based on the strength of social ties), one would expect a fall in suicide rates during Ramadan. Is this conjecture confirmed by observation? That is the question addressed in the present paper. Surprisingly, the most tricky part of the investigation was to find 5 reliable monthly suicide data. In the Islamic world Turkey seems to be the only country whose statistical institute publishes such observations. The data reveal indeed a fall of about 15% in suicide numbers during the month of Ramadan (with respect to same-non-Ramadan months). As the standard deviation is only 4.7% this effect has a high degree of significance. This observation, along with the fact that other occasions of social gathering such as Thanksgiving or Christmas are also marked by a 10 drop in suicides, adds further credence to the B-D thesis. Version of 26 July 2013 Key-words: Suicide, group effect, social interaction, social gathering, Durkheim, Ramadan, Thanksgiving, Christmas 1: Institute for Theoretical and High Energy Physics (LPTHE), University Pierre and Marie Curie, Paris, France. 15 Email: [email protected] 2 Suicide and social ties Suicide is a phenomenon which often seems to frustrate our “instinctive” expectations. For instance, in the northern hemisphere, whether in South Korea or in Turkey, the suicide rate is lowest in November and highest in May or June. Another surprising observation is that suicide is not in the least affected by socio-political events. Thus, even first magnitude events such as the Pearl Harbor or 9/11 20 attacks had no visible effects on suicides. For 9/11 this is true not only at country-wide level but even for suicide rates in New York City itself and whether at monthly, weekly or daily level (Roehner 2007, p. 205-208). To this day the main guideline in the understanding of suicide is still based on the discoveries made by the sociologists Louis-Adolphe Bertillon and Emile Durkheim in the late 19th century. They 25 found that married people had lower suicide rates than bachelors or widowed persons of same age. Moreover, they observed that among married people, suicide rates decreased with increasing number of children (Bertillon 1879 p. 474, Durkheim 1897, chapter II,3). This lead them to the hypothesis that the propensity for suicide was determined by the number and strength of family bonds. So far, however, little (if any) evidence was available to show the effect of social links beyond the family 30 circle. The present paper shows that in Turkey during the month of Ramadan the suicide rates are about 15% lower than during the same time of the year not marked by Ramadan. Why do we think that this observation can be accounted for by the Bertillon-Durkheim conjecture? To see it more clearly one must give a closer look to the characteristics of Ramadan. During the month of Ramadan Muslims fast during the day from dawn to sunset but after sunset they break the daily 35 fast by sharing food with those in need and celebrate with family and friends. This makes Ramadan a time of social gathering. Of course, at this point it is impossible to know whether the fall in suicides is due to fasting, to social gatherings or to some other facet of Ramadan. In order to strengthen the contention that it is indeed social gathering which is the key-factor one must analyze other events marked by social gathering, 40 such as for instance Thanksgiving in the United States. This will be done briefly at the end of the paper. 3 Data We need to ask ourselves the following questions. (i) What suicide data do we need? (ii) Do such data exist? (iii) Are they reliable? We need monthly data from a country (or a region within a country) whose population is over 90% 45 Islamic. Unfortunately, in most Islamic countries the suicide rates as reported by the World Health Organization are very low. In Egypt, Syria, Pakistan, Jordan, Malaysia, Indonesia, Kuwait the rate is under 2 per 100,000. In several (if not all) of these countries suicides are under-reported. This is shown convincingly in the case of Malaysia by T. Maniam (1995) and is related with the fact that many suspicious deaths are just classified as undetermined deaths. Eventually, Turkey turned out to 50 be the only mostly Muslim country for which we were able to find monthly statistics which seem fairly sound. These data seem to exist since 1974 but Internet availability is limited to the period 2000-2012. Over these 13 years the month of Ramadan shifted from December to July (every year it moves backward by some 10 days). Most often Ramadan overlaps two successive months but the degree of overlapping may change 55 widely: from one half in each month (as in 2004 when it covered 15 October to 14 November) to a perfect match with a calendar month (as in 2008 when it covered 1 September to 30 September). Needless to say, when a calendar month is a mixture of Ramadan days and non-Ramadan days, this is not a favorable situation for the purpose of our experiment. Therefore we restricted our analysis to those cases in which the number of non-Ramadan days was less or equal to 5. As 60 an illustration, in 2002 the Ramadan was from 6 November to 5 December which means that in November there were 5 non-Ramadan days. In other words, based on our criterion, 2002 can be included in our sample of favorable cases. Altogether there are n = 6 favorable cases, namely F = {2000, 2002, 2003, 2005, 2008, 2011}. Method There will be two phases in our analysis. In the first we simply wish to get an overall view of suicide 65 numbers during the Ramadan periods. Fig. 1a shows the number of suicides per month. It reveals two features which will be of importance