Variation during autorefraction: Influence of two different target types

An investigation was performed on seven subjects of varying ages in order to determine the influence of two different target types on autorefractor results. A single autorefractor capable of presenting two different targets was used. Multivariate statistical methods which only recently have become available in optometry are used to illustrate graphically the results obtained and formal hypothesis tests further demonstrate that most of the eyes in the sample display similar variation or behaviour with both targets. The mean refractive state likewise is not significantly different for both targets for five of the seven eyes at a 99% confidence level. Departures from multivariate normality (such as resulting from outliers) are shown to be of great importance in order to reach appropriate decisions with respect to target influence on autorefractor measurements and an example is provided to illustrate this point. In the absence of departures from normality it is believed that this study exhibits only weak evidence for differences in variation or mean refractive state with the different target types.