Table 3: Results of final linear and logistic regression models predicting health outcomes from specific types of yoga practicea in combination with influential demographic predictors ( ).

Health outcomeFinal predictorsb S.E.t

MindfulnessBreath workc1.60.295.51*
Philosophy study.31.084.09*
Subjective well-beingMeditationc2.80.713.95*
Philosophy study.76.203.87*
BMI ( )Vigorous posese−.05.02−3.27*
Philosophy study−.16.07−2.22**
Fruit and vegetables/day ( )Standing posese.024.013.31*
Gentle posesf.36.113.16*
Sleep disturbanceVigorous posese−.07.14−5.50*


Vegetarian statusgGentle posesf.73.249.05*/2.07*
Philosophy study.24.0615.84*/1.27*
Alcohol consumptionhGentle posesf−.48.1314.37*/.621*

Specific yoga practices: physical poses (standing, vigorous, inversions, and gentle), breath work, meditation, and yoga philosophy study. Each final model includes all predictors included in the final model. Demographic covariates were included if they had at least a small ( ) and significant ( ) correlation with the health variable. No demographic variable met these criteria that was not included in the final model. ≤ or > once per week. dGender coded “0” = male “1” = female. eDays per month. for 30 minutes per week. gVegetarian status coded no = “0”, yes = “1”. hAlcohol consumption (number of drinks one consumes on a typical day when one drinks) coded ≤2 drinks per day = “0”, >2 drinks per day. iRace coded “0” = other, “1” = white. Abbreviations— : unstandardized beta weight. SE: standard error. t: t score for linear regressions. Wald: Wald statistic for logistic regressions. OR: odds ratio. BMI: body mass index. Note: For all measures, higher scores indicate more of the concept measured. * level (2-tailed). ** (2-tailed).