Over the past decade, the use of portable optically stimulated luminescence (OSL) readers has expanded rapidly, enabling the recovery of latent stratigraphic information from bulk sediments of late Quaternary age. These applications assume that luminescence signals from sample aliquots are reproducible, yet no study has explicitly examined preparation strategies that optimize inter-aliquot consistency. Consequently, no standardized methods for sample preparation exist in portable OSL research. This study initiates this discussion by testing how aliquot preparation and sample size influence reproducibility. Four samples were collected from postglacial eolian dunes in Alberta, Canada. For each sample, 20 aliquots were prepared using four approaches, yielding 320 aliquots: (a) monogranular layers of equal volume portions (~1.5 g), (b) monogranular layers of equal-weight (1.5 g), (c) monogranular layers mounted on double-sided adhesive segments of equal area (~1.0 g), and (d) multilayer portions of equal weight (12.0 g). All aliquots were analyzed with a portable OSL reader. Statistical results showed that for blue-OSL signals, reproducibility varied both among samples and across preparation methods. However, multilayer aliquots prepared with equal weights (12 g) consistently produced the highest reproducibility, suggesting that larger grain populations enhance inter-aliquot consistency. For IRSL signals, no preparation method offered a clear advantage.