The inconsistence may be caused by the sample size in the subgroup of 129 participants.
Assuming a binomial distribution for the number of deaths from all causes, the sample size (for both groups combined) for an all death endpoint is Nall = 2 (1.96 Sqrt [2 vallH0 ] +.84 Sqrt [vallH0 + vallHA])/ d- e), where vallH0 = (p + k)(1- p- k) and vallHA= (p+ k- d+ e) (1- p- k+ d- e) are the variances for one subject under the null and alternative hypotheses respectively.
The inconsistency between our study and Baumgartner's may be caused by the sample sizes, the exposure durations and altitudes.
Given the large number of SNPs and the low probability that any specific one causes disease, the sample sizes in association studies need to be large enough to achieve adequate statistical power.
Each TE located in an inversion is eliminated from the frequency calculation of that TE, which causes variation in the sample size of different TEs.
The presence of outliers is known to cause heteroskedasticity, especially when the sample size is small [ 9] and therefore we believe that this is the most likely source of the heteroskedasticity.
However, doing so would have caused a significant reduction in the sample size.
14 The study was originally of approximately 17?000 births, but the subsequent exclusion of Northern Ireland, death, emigration, and other causes of attrition reduced the sample size, although the cohort has remained broadly representative of the target population.
13 The study originally comprised approximately 17 000 births, but the subsequent exclusion of Northern Ireland, death, emigration and other causes of attrition reduced the sample size, although the cohort has remained broadly representative of the target population.
For instance, the extension of χ 2 test [15] to higher dimensions suffers from the curse of dimensionality [40] caused by the space sparsity unless the sample sizes are large enough.
