Abstract
M.Phil.
The primary purpose of this study was to investigate in different age groups the various relationships between autorefraction and subjective refraction, between autorefraction and retinoscopy and between retinoscopy and subjective refraction. Multivariate statistical methods are used to explore the nature of the relationships between these three different methods of evaluating refractive state. This is the first occasion in which these statistical approaches have been used to address these relationships within a sample
of black South Africans.
The short-term variation of measurements of the autorefractor used in this study was evaluated by means of an artificial or model eye. Here samples of fifty measurements
each were obtained every two hours over a period of one day.
Scatter plots, meridional profiles and hypothesis tests of equality of variance-covariance and of equality of mean autorefraction are used for analysis of the measurements
obtained. Such analysis suggests that the variability in autorefraction across the day was of small magnitude and less than that typically found with human eyes. The implication is
that many other factors besides the instrument itself are responsible for the variation found when using an autorefractor on a human eye. These factors include changes
in accommodation, ocular fixation, the blinking process and attention. The environment within which the autorefractor is placed is important and changes in, for example, temperatur may have subtle but significant influences.
The major part of this dissertation describes a study performed over a period of about seven months during which a sample of 240 male and female subjects were chosen randomly from a larger group of about 2800 patients who came to have their eyes examined in a clinical practice. A questionnaire was completed by each subject and retinoscopy was performed followed by subjective refraction and then autorefraction.
Ocular health was assessed and visual acuities were measured.
The study subjects were further divided into six smaller subgroups according to their ages and the analysis of results obtained with retinoscopy, subjective refraction and autorefraction included both qualitative and quantitative multivariate methods such as stereo-pair scatter plots, trajectories of change of dioptric power, meridional profiles,
confidence and distribution ellipsoids, and testing of hypotheses of equality of variance-covariance and of equality of means. All of these methods help in understanding the
nature of the various relationships between the different refractive methods, and they also are important when considering concepts such as mean refractive state and variation of refraction in human eyes. They also are used to develop an understanding of the distribution or spread of the population of refractive states from which the 'sample was
obtained.
For the three refractive methods . (retinoscopy, subjective refraction and autorefraction) the spread or distribution of measurements in the six age groups suggests that generally the refractive behaviour of right and left eyes was similar. A possible shift towards hyperopia and more astigmatism with increasing age is apparent. For example, mean autorefraction for the right and left eyes in the youngest and oldest age groups (Groups 1 and 6 respectively) is -0.41/-0.06 x 148 and -0.32/-0.07 x 177 and 0.41/-0.57 x 102 and 0.57/-0.32 x 75.
Similarly mean subjective refraction for the right and left eyes of Group 1 is 0.10/-0.14 x 97 and 0.30/-0.17 x 81 and is 0.78/-0.70 x 95 and 0.89/-0.47 x 82 for the right and left
eyes of Group 6 respectively. Again mean retinoscopy for the iv
right and left eyes of Group 1 and Group 6 respectively is 0.69/-0.05 x 91 and 0.65/-0.21 x 82 and 1.72/-0.88 x 89