User:Rudrasharman/Notes/DNA papers

Sharma 2009
Percentages from their Table 1 (after 6 typo corrections) reconstituted into what is supposed to be a contingency table.


 * Typos corrected before conversion:
 * JK Gujars/Q(xQ5): 2.0% -> 2.04%
 * JK Gujars/R1a1: 40.86% -> 40.82%
 * UP Brahmins/L: 3.21% -> 3.23%
 * Punjab Brahmins/C5: 3.58% -> 3.57%
 * Gujarat Brahmins/C5: 3.33% -> 3.13%
 * Gujarat Brahmins/E: 3.33% -> 3.13%
 * In 8 out of 16 rows (marked with crosses), Sharma et al report percentages which amount to non-integral numbers of persons for the stated sample size. However, the sample size column correctly totals to 621, as reported in the paper; and similarly, the percentages in all 16 rows correctly total to 100% in each case. It follows that the percentages in the 8 erroneous rows either pertain to some other data set altogether or are fabricated for purposes unknown.


 * "Correct" proportionality factors are shown in parentheses after the cross marks, e.g. the sample size for the Himachal Brahmin row has to be a multiple of 19 for the reported percentages to amount to integral values. The MP Saharia row has extremely incommensurate values: the minimal first difference of 0.16 between F* and R2 implies a multiplier of at least 6:1, which jacks the proportionality factor into mid-3 figures at a minimum.