The physiological stress response is frequently gauged in animals, non-invasively, through measuring glucocorticoids in excreta. A concern with this method is, however, the unknown effect of variations in diets on the measurements. With an energy dense diet, leading to reduced defecation, will low concentrations of glucocorticoids be artificially inflated? Can this effect be overcome by measuring the total output of glucocorticoids in excreta? In a controlled laboratory setting we explored the effect in mice. When standard mouse chow – high in dietary fiber – was replaced with a 17% more energy-dense diet, fecal mass was significantly reduced. As circulating levels of corticosterone and the total output of corticosterone metabolites over time remained unaffected, the result was an overestimation – more than a doubling – of the corticosterone metabolite excretion if expressed as concentrations. Similar results were obtained for testosterone metabolites. Although measuring the total output is not feasible in, for example, wildlife studies, the present findings highlight the perilousness of relying on concentrations of hormones in excreta with no associated information of the dietary intake as even moderate changes can exert a great influence.