Factors affecting accuracy in the Developmental Eye Movement Test measurement for Cantonese-speaking children.

Background:  This study aims to determine the factors affecting the accuracy in the Developmental Eye Movement (DEM) test measurement for Cantonese-speaking children. Method:  The voices of 10 Cantonese-speaking children (aged between six and eight years) undertaking the vertical test (40 numbers) of the DEM test were recorded on a digital audio recorder. These voice clips were assessed by an experienced examiner to give the DEM test times. The examiner repeated the measurement five times for each voice clip and completed all 10 voice clips from the 10 children. The average of a single measurement, the mean of two, the mean of three and the mean of five repeated measurements were then compared. Five experienced and five inexperienced examiners of the DEM test were asked to record the results from five Cantonese-speaking children by listening to the playback of prerecorded audio clips. The deviation of the results from the preset values between the two groups was compared. Results:  There is no difference between the single measurement and those obtained by mean of two, three, four or five. Both experienced and inexperienced examiners obtained a higher deviation from the preset values in the adjusted compared with the non-adjusted times in both vertical and horizontal times. Experienced examiners measured the vertical times and adjusted vertical times significantly closer to the preset values than the inexperienced examiners (paired t-test, p < 0.05). Conclusion:  The DEM test needs to be measured only once for an accurate time measurement. Inaccurate assessment of ‘reading errors’ in the DEM test can increase the inaccuracy of the adjusted times. It is suggested that an audio recording of the test be made to allow reassessment of reading errors. Experience in using the DEM test yields a more accurate DEM measurement as errors are detected more easily.