Nowadays, there are still many diseases with limited or no reliable methods of early diagnosis. A popular approach in clinical diagnostic research is Raman spectroscopy, as a relatively simple, cost-effective, and high-throughput method for searching for disease-specific alterations in the composition of blood plasma. However, the high variability of the experimental designs, targeted diseases, or statistical processing in the individual studies makes it challenging to compare and compile the results to critically assess the applicability of Raman spectroscopy in real clinical practice. This study aimed to compare data from a single series of blood plasma samples of patients with Alzheimer’s disease and non-demented elderly controls obtained by four different techniques/experimental setups – Raman spectroscopy with excitation at 532 and 785 nm, Raman optical activity, and surface-enhanced Raman scattering spectroscopy. The obtained results showed that the spectra from each Raman spectroscopy technique contain different information about biomolecules of blood plasma or their conformation and may, therefore, offer diverse points of view on underlying biochemical processes of the disease. The classification models based on the datasets generated by the three non-chiroptical variants of Raman spectroscopy exhibited comparable diagnostic performance, all reaching an accuracy close to or equal to 80%. Raman optical activity achieved only 60% classification accuracy, suggesting its limited applicability in the specific case of Alzheimer’s disease diagnostics. The described differences in the outputs of the four utilized techniques/setups of Raman spectroscopy imply that their choice may crucially affect the acquired results and thus should be approached carefully concerning the specific purpose.