The argument is made that because Muslim majority countries have a high Google Trends’ search volume index number for queries termed as bestiality related, this for some reason translates into tolerance for bestiality in the Muslim community and by extension possible allowance for it in Islamic law.
The amount of criticisms that can be stacked up against the methodology used for such an argument would require a separate article. A concise analysis of the data and then review of the faulty nature of such a methodology has been provided here.
Google trends’ Regional interest section enumerates the “Search Volume Index” (SVI) values of regions in descending order. The index calculates the interest of a particular search term in a region, or in other words, indicate where a search term is comparatively more popular. Apparently, Google trends does include, in its calculations, a few different combinations of words, which it is programmed to consider as similar to the entered search query, but these combinations are far from comprehensive. The SVI itself is calculated as the ratio of number of queries for a particular term to the number of total queries, which is then normalized to give results within the range of 0-100. This gives SVI values which are essentially a comparison of the popularity of a search term between different regions. The region with the most interest will always have the value of 100. It should be pointed out that, this does not represent the actual search volume of a particular term, but its “interest” by comparison.
The image above, shows Google Trends’ results for the search term “Google”, in which seven SVI values representing seven distinct regions are enumerated in descending order. These results show that, according to algorithm calculations, Mexico has a 68% interest in this particular search query as compared to Tunisia. The trends’ calculations provide 0-10 different SVI values, corresponding to a particular region, for each search term. For the sake of understanding of how the following analysis is performed: if another search term were to be analysed and from a total of seven values shown, if the first three regions were found to be the same, it would mean that in the acquired data consisting of a total of 14 (7 from the first search term and 7 from the second) SVI values for these two terms, two SVI values correspond to Tunisia, Algeria and Morocco each. This means that these three countries are represented two times in this data. Such a representation of data will be used in this analysis.
Google Trends’ SVI values for 15 terms which are presumably bestiality related were examined, namely: “Dog sex”, “Dog sex porn”, “Camel sex”, “Animal sex”, “Pig sex”, “Donkey sex”, “Cat sex”, “Horse sex”, “Cow sex”, “Goat sex”, “Snake sex”, “Monkey sex”, “Bear sex”, “Elephant sex”, and “Fox sex”. The time period chosen was “2004 – present” (the present being February, 2016).
These 15 search terms resulted in 146 SVI values (each corresponding to a particular region) which include a total of 37 different countries repeating in varying numbers. Of them, there were Muslim majority (14), Christian majority (14), Buddhist/Hindu (6) and divided populations (3). The data at this point is pretty evenly divided.
The most intriguing find from the data was the comparatively high prevalence of the South Asian Bloc (SAB) namely Pakistan, Bangladesh, India and Sri Lanka, and in it the high representation of Pakistan. In the top three places, of the 45 different values, 31 (i.e. 69% of the total) of them corresponded to SAB regions, of which, Pakistan was repeated 14 times, India (6), Sri Lanka (6) and Bangladesh (5). In the top five, they made up 53% of the total number of values, Pakistan (15), India (12), Sri Lanka (7) and Bangladesh was 8 times repeated. In such a rudimentary organization of data, Pakistan does have an uncharacteristically high representation but the overall representation of SAB is curious.
After this, analysis of SVI values ≥ 30 was done, assuming that below 30 the value was too low and virtually any country in a different time period or while analysing a different combination of words used in the search term or for some other anomalous reason could end up being included. In such a data, a total of 87 individual SVI values were attained, of which there were values belonging to Muslim majority (34), Christian majority (19), Buddhist/Hindu (27) and divided populations (7). Excluding the SAB, and for some degree of equivalency the top two for Christian majority, there were Muslim majority (11), Christian majority (10), Buddhist/Hindu (6). The even division mentioned before is again apparent.
Important to note at this point, is the sensitive fluctuation of included data; countries with SVI’s of 29-31 could have been included in the data at one time and after a few hours may have had to be excluded because of their value fluctuations. This was observed in the case of Malaysia, for which one SVI value was excluded in the beginning but after just a few hours had to be included in this data set.
From the above data it is apparent that for these particular search terms the SAB has a high degree of results tilting effect. In-fact, Pakistan and Bangladesh made up 68% of the total of Muslim majority countries representation in the “SVI values ≥ 30” dataset and India and Sri Lanka made up 78% of the total Buddhist/Hindu representation. Of the 27 different countries present in this dataset, just four countries of the SAB made up 51% of all the SVI values ≥ 30, while the remaining 49% of the values were divided among rest of the 23 countries.
As mentioned before, Pakistan has high representation in this data even by the uncharacteristically high standards of the SAB. But here, one of the legitimate criticisms that can be made is the ignorance of search terms in local languages. For example, the search term “Dog sex” would translate as “Kutta sex” for the people of Pakistan, India and Bangladesh, but Pakistan and Bangladesh were absent from the recorded SVI values here. A somewhat similar result was also obtained for “Goat sex”.
Regarding the specific period used for analysis, it was observed that Sri Lanka which was at the time of this writing (February, 2016) top placed for the search term “Elephant sex” – from June, 2010 to July, 2014, had one of its neighbours at the top spot.
“Dog sex”, “Dog sex porn” and later separately, “Dog sex video” were analysed to show the difference caused by just single word changes. Tanzania which comes at top for the first and third terms, is absent from the top ten for the second term. Kenya which comes at 10th for the first and 9th for the third term, is at the top for the second term. Pakistan which comes at 2nd for first two terms comes at 5th for the third term. Such is the inadequacy of the methodology, in using such small sample sizes with so few permutations.
The inadequacy of the methodology involved became more obvious when terms such as “Dogorama” (a zoophile pornography film), “Sex with cat”, “Sex with sheep”, and “Sex with pig” were analysed.
- A North American country which was observed only once (at 7th position) in the previous “SVI values ≥ 30” dataset (comprising of a total of 87 values) was at the top in the first three search terms (for the first term, it was the only country) and 3rd (although SVI value was < 30) in the only three regions for the fourth term.
- A West European country which was observed zero times in the same previous dataset, was absent for the first term and second among the next three terms (only two countries were recorded for the second and third terms). The West European nation had SVI values similar to the North American country.
It should be mentioned here that, not all modified combinations similar to these, gave such drastically different results. The screenshots of the Google trends data used, were saved for future reference.
Regarding the inherent inaccuracies in such data, Google back in 2010, officially made the following statement:
“We do our best to provide accurate data and to provide insights into broad search patterns, but the results for a given query (…) may contain inaccuracies because the sample size is too small for the results to be statistically sound.”
Another problem in obtaining reliable results, comes in the form of the disparity in the bans put on such websites. It is likely that when it comes to adult content related searches, users will make direct use of adult websites’ search engines instead of indirectly going through Google search. For regions with active restrictions for such websites, the indirect method of using Google search is likely to be used which in turn drives up those regions’ representation in Google Trends. On the other hand, however, such restrictions on internet content can also push users towards using various proxy services, in which case these regions might show a decrease in their Google Trends representation. Fairly odd behaviour is observed when regions implementing active bans on adult sites, somehow, still manage to appear in the analytics graphs of visitors by region, presented by such sites. These even include cases where censored regions even manage to get into the top three spots.
In summary, although one Muslim majority country does have an exceptionally high representation for the particular search terms considered, even which is criticisable from several avenues like exclusion of searches in regional languages and same searches using different word combinations; no other Muslim majority country shows any extraordinary behaviour, they possibly might have, in some different time period, but not in the one analysed (which was the most comprehensive to use for such a scenario). Therefore, the relation of bestiality with Muslim countries even in such fallacious methodology is unsupported. The most reliable conclusion that could be established are the peculiar results for the South Asian Bloc, even which could be completely thrown out by attributing this uniqueness entirely to something as simple and unrelated as a difference in the popularity of making Google searches in English instead of other regional languages.
As has been discussed earlier, analysis like these are too incomplete to give any reliable conclusions. Any of the countries that end up being shown under a bad light in this analysis, should not be used as arguments against their “morality”; for which reason, care was taken to maintain, as much as possible, the anonymity of such regions.