As a highly sensitive food, the safety of meat is an important issue in the context of food safety. The current analytical methods for detection are based on different, invasive proce-dures. The analysis of these foods therefore causes avoidable food losses in the case of foods that are still suitable for consumption. Spatially offset Raman spectroscopy (SORS) is a po-tential alternative analytical method, as it offers the possibility to measure through packag-ing. Using packaged chicken breast as an example, we simulated two different storage condi-tions and compared these samples to control samples using SORS. To validate the group as-signment and to monitor the resulting changes, total bacterial count and NMR spectra of the meat extracts were measured and evaluated. We overcame the various difficulties in the multivariate data evaluation of the through package measured Raman spectra and were able to classify samples deviating from the control group with very high accuracies and simulta-neous absence of false positive results.