Erratum: Depressive symptoms among Mexican adolescent girls in relation to iron status, anaemia, body weight and pubertal status: results from a latent class analysis - CORRIGENDUM (Public health nutrition)

The authors apologise for typographical errors in the above article. In multiple instances it is stated that both multinomial logistic regression and linear regression were used for modelling. However, only logistic regression was utilised since the outcome for all models is dichotomic (class membership yes/no). This error is present on pages 408 (abstract), 410 (Methods) and 413 (Table 5). These errors do not affect the results, as logistic regression was originally employed, but was mistakenly labelled as linear regression. Furthermore, there are errors in Table 4 under the category “Iron deficiency (serum ferritin concentration＜15 μg/l)” and “Iron deficiency erythropoiesis (sTfR concentration＞8·3 mg/l)†”. In both instances, the first row should be labelled “no”, and the second row should be labelled “yes”. For the same table, all analyses have been re-run, and the corrected values for the iron deficiency erythropoiesis adjusted model have been updated.