Dysregulation of the adrenal cortex has been assessed with measurement of salivary cortisol. So far salivary cortisol is routinely measured with immunoassay (IA). However, liquid chromatography-tandem mass spectrometry (MS) is known to offer better specificity. We compared the concentrations of salivary cortisol measured by MS and IA at basal and stress induced conditions and evaluated reasons for the difference in method-dependent cortisol results. Saliva samples (n=2703) were collected from 169 children (age range: 8-14 years; 81 healthy children; 55 with internalizing and 33 with externalizing disorders) under circadian conditions and during the Trier Social Stress Test for Children (TSST-C). Biochemical analyses were performed with MS for cortisol and cortisone, IA (IBL, RE62011) for cortisol, and enzyme kinetic assay for α-amylase. MS and IA showed mostly comparable results for circadian activity and TSST-C response with similar statistical power. However, IA measured cortisol concentrations about 2.39-fold higher than MS. We found that this difference in measured values between MS and IA was mainly due to different standardization of IA compared to MS. In addition, at cortisol IA concentration below 5 nmol/L, cross-reactivity with cortisone was found to contribute to the lower concordance between MS and IA. Immunoassay and LC-MS/MS were largely comparable in the interpretation of salivary cortisol dynamics in stress research. But the IA method revealed a restricted accuracy in the measuring range below 5 nmol/L.