The literature concerning subjects who have reading difficulties has repeatedly noted their abnormal eye movements. The Developmental Eye Movement (DEM) test was developed on the assumption that poor eye movement control is a major cause of reading difficulties. The hypothesis tested by this study was that practice in fluent reading trains the eye movements that result in a good DEM score, whilst poor readers will exhibit low DEM scores due to insufficient training. English readers (43 children, 20 adults), and Arabic readers (six children, five adults) were recruited. The DEM test was administered twice, performed once reading the horizontal section in the habitual reading direction and secondly in the opposite direction, thus enabling the subjects’ eye movements to be compared when reading in their habitual direction and when reading in a direction which is relatively unpracticed. Paired t-tests showed that the difference in eye movements (quantified via the DEM test ratio) between the two opposing reading directions was significant in English reading adults, English reading children and Arabic reading children, but not significant in the Arabic adults, who were equally practised in reading in the two directions. The results support the hypothesis that abnormal eye movements are more likely to be an effect and not the cause of reading difficulties. The DEM test should not be used to diagnose eye movement difficulties in a patient with poor reading ability.