none of the variables were significant

[Gee Smarty]

Hey everyone,

I have a set of data (n=3000, 48% males and 52% females) from a cross sectional study.
I got 20 explanatory variables (let's say X1, X2, X3,.... X20) and binary outcome. I conducted logistic regression for whole (male and female), male-only, and female-only population.

From fully adjusted logistic regression (adjusted for all explanatory variables), I found only X1 and X2 were significantly associated (p<0.05) with the outcome for the whole (males and females) and female-only population.
But, for male-only population, none of the explanatory variables were significantly associated with the outcome.

Do you have any thoughts why there's no variable were significant among males?

I would appreciate your help.

Ooh, yeah, and wish you a very mery christmas too!

Gee

[bugman]

Do you have any thoughts why there's no variable were significant among males?



Not without a better understanding of your data set.

[krytellan]

Do you have any thoughts why there's no variable were significant among males?

What bugman said. Also, possibly just because none of the variables account for males. If it ain't there, it ain't there (sometimes).

[Gee Smarty]

Thanks Bugman and Krytellan for your replies.

Sorry for the incomplete information.
The outcome was back pain (yes/no), X1 was awkward posture, and X2 was stress.

X3 to X20 were variables in other physical factors (repetitive, lifting, vibration tools, etc), psychosocial factors (dissatisfaction, social support, etc), and environmental factors (cold, hot, etc).

Does it make clearer?