In community-based studies of stress and immunity, saliva samples offer a non-intrusive way of gathering biological data. Cotton-based devices are widely used in cortisol research, but some may affect assay results. We compared assay reliability and perceived acceptability of three saliva collection methods: passive, cotton ‘salivettes’ and cellulose-cotton tip ‘eyespears’. Compared to passive collection, salivettes reduced the concentration of cortisol (p = .001) and sIgA (p = .002). Eyespears did not reduce cortisol or sIgA concentration, and showed less interference in the rank ordering of cortisol (reyespear with passive = .90) and sIgA scores (reyespear with passive = .96) compared to salivettes (r cortisolsalivette with passive = .79; r sIgAsalivette with passive = .66). The comfort and acceptability of both cotton-based devices were rated positively. Cotton-cellulose eyespears could offer methodological advantages for collecting saliva to measure cortisol and sIgA levels, and, because they can be held during sampling, may be useful for research with children and the frail elderly.