The linear combination of model spectra (LCModel) calculation of a parameter for long-term quality control, kT, was introduced, representing the ratio of the temporal and nominal intensities of CH3 groups of lactate and acetate in a quality control phantom. This procedure is a part of the quality assurance of the scanner using fully automatic measurement and calculation of kT parameters, and utilizing Shewhart regulation control charts for continuous evaluation of the magnetic resonance (MR) scanner setting. The application of the kT parameter for the correction of in vivo data increases the precision of molar concentration determination by about 4%. This was tested by the quantitative in vivo MR determination of the molar concentrations of 13 prominent metabolites (N-acetylaspartate (NAA), N-acetylaspartylglutamate, creatine and phosphocreatine (Cr), choline-containing compounds (Cho), myo-inositol, scyllo-inositol, gamma-aminobutyric acid, glutamine, glutamate, glucose, lactate, alanine, taurine) in the white matter and hippocampus of the brain in groups of volunteers, using a short echo time stimulated echo acquisition mode sequence (echo time = 10 ms) and the LCModel technique. The repeatability of the measurement of prominent metabolites such as NAA, Cr and Cho was found to be around 10% (relative standard deviation, n = 6); precision in a group of volunteers (n = 20 and 28, respectively) was in the range of approximately 13-20%. For other metabolites, which are measured with a lower signal-to-noise ratio, the precision can be much lower.