I mean for binary predictors only

You could use OLS regression with a robust covariance matrix. See the following article, for example.

Cheung, Y. B. (2007). A modified least-squares regression approach to the estimation of risk difference. American journal of epidemiology, 166(11), 1337-1344.

Seyyed Amir Yasin Ahmadi see the link: Article A Review of Adjusted Estimators of Attributable Risk

Best. D. Booth

Dear Bruce Weaver David Eugene Booth

Thank you for your suggestions

I saw your suggested papers and also their references.

I found that any researcher-designed model can be used for manual calculation of AR. However, I am seeking for my researcher-designed method.

regarding logistic function:

P=exp(bX)/(1+exp(bX))

I think b is the very beta coefficient. If my adjusted beta=1,

can I consider P(bx=1)-P(bx=0) as adjusted AR?

e.g.:

> exp(1)/(1+exp(1))

[1] 0.7310586

> exp(0)/(1+exp(0))

[1] 0.5

> 0.7310586-0.5

[1] 0.2310586

>

thus, adjusted AR = 0.23

Am I right?!!

Nope you used the wrong formula. See the link : https://www.google.com/search?biw=1600&bih=795&ei=m8etXsifLoO0tAb524TYCQ&q=how+to+do+logistic+regression+in+Spss+examples%3A&oq=how+to+do+logistic+regression+in+Spss+examples%3A&gs_lcp=CgZwc3ktYWIQDDIFCCEQoAE6BAgAEEc6BggAEA0QHjoFCAAQkQI6AggAOgUIABCDAToOCAAQ6gIQtAIQmgEQ5QJQ9oA4WILIOGC2wDpoAXACeACAAecBiAGGGpIBBjMzLjIuMZgBAKABAaABAqoBB2d3cy13aXqwAQY&sclient=psy-ab&ved=0ahUKEwiI2aHY8pXpAhUDGs0KHfktAZsQ4dUDCAw

The adjusted risk is the Odds ratio value from the regression coefficient. You should have shown your data. Best, David Booth to calculate it see the second link below See: https://www.google.com/search?biw=1600&bih=795&ei=W8utXoitFYy5tAaKiKPQCw&q=adjusted+risk+ratio&oq=adjusted+risk+ratio&gs_lcp=CgZwc3ktYWIQAzICCAAyAggAMgIIADICCAAyAggAMgYIABAWEB4yBggAEBYQHjIGCAAQFhAeMgYIABAWEB4yBggAEBYQHjoECAAQRzoFCAAQzQI6DggAEOoCELQCEJoBEOUCOgUIABCRAjoFCAAQgwE6BAgAEEM6BAgAEA1Qk58tWK_xLmCqjC9oAXACeAOAAcoBiAG-IZIBBjQzLjUuMZgBAKABAaoBB2d3cy13aXqwAQY&sclient=psy-ab&ved=0ahUKEwjIxuqh9pXpAhWMHM0KHQrECLoQ4dUDCAw&uact=5

2nd link: https://www.google.com/search?biw=1600&bih=795&ei=Ys-tXuvDI5SStAbBt6iwDQ&q=odds+ratio+from+logistic+regression+coefficient&oq=odds+ratio+from+logistic+regression+coefficient&gs_lcp=CgZwc3ktYWIQDDIECAAQRzIECAAQRzIECAAQRzIECAAQRzIECAAQRzIECAAQRzIECAAQRzIECAAQR1AAWABg8awCaABwAngAgAEAiAEAkgEAmAEAqgEHZ3dzLXdpeg&sclient=psy-ab&ved=0ahUKEwjr_MeN-pXpAhUUCc0KHcEbCtYQ4dUDCAw

Best, D. Booth