The aim of the study was to refine and improve the inter‐observer reliability of an observation schedule, developed by Le May&Redfern, designed to record the amount and type of nurse‐patient interpersonal communication The new technique involved observation of non‐verbal interactions using a portable computer programed as an event recorder and a tape recorder to record conversation The observation schedule was employed m an acute/rehabilitation care of the elderly ward and inter‐observer reliability was tested on two occasions The second observer used a handwritten version of the schedule Reliability coefficients were acceptable (≥ 0.60 or ≥ 70% agreement) for all the schedule components although the recording of non‐verbal response was inadequate. Use of the portable computer produced no additional benefits over the handwritten schedule The observation confirmed previous research which found that relatively little expressive touch takes place between nurses and elderly patients A simplified version of the handwritten schedule together with tape‐recording of verbal communication is likely to incre