I'm not sure if I well understood the procedure, but the issue you mentioned may result from the type of the FFQ you used. For example, are you able to precisely calculate energy intake based on FFQ data? I mean, as precisely as you do it with dietary records.

Thanks Agata Chmurzynska

Yes we calculate the energy intake using FFQ. But I think the between person variation is very high. Because the adjustment of the energy intake changes the rank of the people. We have recruited the FFQ data for three times and we have the same problem with almost all of the FFQs.

Dear Agata Chmurzynska

I have to add that the FFQs are also reliable in measuring the energy intake.

Don't you think that the adjustments per se can change the correlations?

Even this FFQ is reliable in measuring energy intakes it gives you different results than energy intakes based calculated from dietary record. Thats why the correlations change after adjustments.

Thanks again. Infact the adjustment per se is changing the correlation and I am not sure about its reason. When we check the energy intake, the correlation between energy calculated from dietary record and energy calculated from FFQ is reasonable, too. Do you think that the energy intake calculated from FFQ is not valid? As I read in previous papers, the energy adjustments reduces the between-person variation and I think the between-person variation is high in our sample. I am not sure if energy adjustment changes the validity correlation with such intensity.

Yes, calculation of energy intake from FFQ data could possible be less reliable than from dietary records.