3 ms·
New study connects religiosity in US South Asians to cardiovascular disease
- jdale27 6y agoVery skeptical of this study. Here's the full text: https://www.nature.com/articles/s41598-020-79429-1 https://www.nature.com/articles/s41598-020-79429-1. The first thing they did was a proteomic screen of the study participants' blood samples to identify proteins that are associated with cardiovascular disease. They analyzed each of 1305 proteins with a univariate logistic regression model, and a P-value threshold of 0.05, and they found 36 proteins. There's nothing in the materials and methods about multiple-testing corrections or controlling the false discovery rate, which are standard and essential techniques when going on 'omic fishing expeditions like this. Given that many or most of these 36 proteins could be false positives, all the subsequent analysis of the associations between the proteins and the religiosity/spirituality scores and their impact on CVD is highly questionable.
- gus_massa 6y agoWith 1305 tests, and a 0.05 threshold, you expect to get roughly a 1305 * 0.05 = 65.25 false positives. Then they do some multivariate analysis and reduce the number of proteins to 3. Then they do some analysis and some variables are somewhat correlated, but I don't see a clear result.
- jelliclesfarm 6y ago> in which, for example, individuals experiencing adverse life events feel they are being punished or abandoned by their God, or have a crisis of faith. [..] I don’t know if I missed this, but what religion and which God are they referring to?
- jdale27 6y agoMaterials and methods: https://www.nature.com/articles/s41598-020-79429-1#Sec11 https://www.nature.com/articles/s41598-020-79429-1#Sec11 Look for "R/S variables" section.