Atkinson has shown that early correction of hypermetropia reduces the incidence of esotropia. If esotropia is reduced by prescribing glasses early, the rate of esotropia-induced amblyopia can be similarly reduced; this would have important economic consequences. We have studied (1) how costs compare to benefits in early visual screening, (2) how videorefraction as used by Atkinson compares to retinoscopy, and (3) whether esotropia is more likely to occur in children who have increasing as opposed to decreasing hypermetropia. The costs of the study so far have been high. It was exceedingly difficult to get all infants invited, come to the clinic and examined. Videorefraction did not compare favourably with retinoscopy in terms of costs and precision, whereas the amount of skill and time needed was approximately equal. The third question, whether esotropia is more likely to occur in children who have increasing as opposed to decreasing hypermetropia, arose from the controversy whether, in the general population, refraction increases or decreases during the first years of life. We found that papers reporting a decrease of hypermetropia in early childhood were studies of large cross-sections of the general population, whereas papers that reported an initial increase originated from ophthalmological practices or strabismus departments. These conflicting results could be reconciled by assuming a population bias: if esotropia is more likely to occur in children with increasing hypermetropia, children with increasing hypermetropia will preferentially be seen by ophthalmologists. It seems natural that children with increasing hypermetropia are more likely to squint, because additional accommodation, needed to overcome increasing hypermetropia, will inevitably confer additional convergence. This relationship has meanwhile been confirmed by others.