selection bias: absence of comparability between groups being studied. the causal relation between exposure and disease is distorted due to (a) procedures used to select or exclude subjects, (b) differences between participants and nonparticipants, and (c) differences in follow-up or drop-out rates that differ according to a causal exposure. information bias: has the information been gathered in the same way? incorrect determination of exposure and/or outcome; if bias is differential by group this may increase or decrease the relative risk/odds ratio; if non-differential "noise in the system" the odds ratio may be shifted towards 1 (no association). some types of information bias: diagnostic suspsicion bias, surveillance bias, recall bias, reporting or wish bias, bias in abstracting records, bias in interviewing. confounding (bias) is an extraneous factor blurring the effect? the causal relation between exposure and disease is distorted due to association of the exposure with other factors that influence the outcome. the effect of the exposure is mixed together with the effect of another exposure, leading to bias. what is a confounder? if we want to know whether factor A is a cause of disease B, then a third factor X is a confounder if Factor X is a  known risk factor of disease B and Factor X is associated with Factor A, but is not a result of factor A. Some approaches to addressing confounding: during study design, restrict study to those without confounder or match on confounder; during data analysis, stratify by confounder or adjust for confounder in multivariate models. 

something "different" distorts the planned comparison.