A reliable observational study is one with small errors of measurement, one that shows observer agreement (interobserver agreement percentage), stability (stability coefuicient), consistency (reliability coefficient), and dependability (generalizability coefficient) of behavioral observations for individuals on the trait, characteristic, or behavior being assessed in the natural environment.