Medical research in general and equipment comparisons in particular suffer from a deficiency and disparity of statistical analysis. We studied four semi-automatic sphygmomanometers in comparison with the Hawksley random zero mercury manometer. The Astropulse 80 showed marked deviation from the standard machine in average blood pressure recorded and greater variation of readings. The Infrasonde M3010, Bosomat II and Infrasonde D4000 were similar to the standard machine. To help comparability in future studies we suggest the following guidelines; blood pressures to be taken under standard conditions, data to be shown as scatterplots, analysis of variance to be performed and regression analysis to be used only to give an approximate overall view. Finally, we urge the manufacturing companies to conduct such studies and include the results in their brochures.