The associations with CIMT and atherosclerosis across WHR quartiles were not observed in both sexes, and in the case of WTR, the odds of carotid atherosclerosis across WTR quartiles were lower than WCR quartiles (data not shown).
The highest OR was observed in the third quartile, except in the case of the relationship between per capita income and obesity.
In any case, the odds ratio was observed to have its greatest magnitude, not in the quartiles having a lower per capita income or a lower population percentage with university education, but rather in the third quartile, except in the case of the association between per capita income and obesity.
In the case of a non-Gaussian distribution, median and quartiles are given.
For example, in the case of the normal distributions, examples of subsets T are the lower quartile and the upper quartile.
Quartiles were used in the case of the CAT-Health completion times and test length.
However, in the case of highly chlorinated PCBs, such as PCB178, PCB194, and PCB199, mean HOMA-IR increased from the 1st to 2nd quartile and then clearly decreased through the 4th quartile, forming inverted U-shaped associations.
In the case of family income, we used two approaches to detecting non-linearity: a semilog model in which ECEC use was regressed on the log of income and a quartile model in which ECEC use was regressed on three dummy variables indicating family income quartiles (with the wealthiest quartile excluded as the comparison group).
For all analyses, I examined histograms, quartile-quartile plots, and predicted versus observed plots to ensure that model assumptions of normality (or Laplacian errors, in the cases of epistasis analysis [see below]) were met.
Additionally, regardless of cancer type, our population had a higher proportion of cases in the highest quartile of cSES.
Also, the trend observed in the discovery set of more cases in the highest (Q4) quartile of MUC1core3, MUC1Tn and MUC1STn was not observed (see Table 3B).
