RT Journal
A1 Golomb BA
T1 TOo sweet to be real?—reply
JF Archives of Internal Medicine
JO Archives of Internal Medicine
YR 2012
FD September 10
VO 172
IS 16
SP 1270
OP 1270
DO 10.1001/archinternmed.2012.3388
UL http://dx.doi.org/10.1001/archinternmed.2012.3388
AB 
Indeed, an explanation based on differential honesty in relation to adverse behaviors related to BMI would seem similarly germane to other variables, like calories, saturated fat, and exercise engaged in. Yet for these variables, reporting bias did not obviate statistically significant BMI associations in the expected direction—much less produce significant associations in the reverse one. Nor does this explain why chocolate consumption frequency, but not amount, demonstrated the favorable association.