Vision Rex Vol. 22, pp. 1025 to 1031. 1982 0042-6989# 82jO81025-07603.00~ 0 zyxwvutsrqponm Printed in Great Britain Pergamon Press Ltd A FAILURE TO OBSERVE NEGATIVE PREFERENCE IN INFANT ACUITY TESTING MARTIN S. BANKS,* BENJAMIN R. STEPHENS and JAMESL. DANNEMILLER Department of Psychology, Mezes Hall 330, The University of Texas at Austin, Austin, TX 78712. U.S.A. (Receiued 10 April 1981; in reuisedform 7 December 1981) Abstract-Held et al. (1979) Vision Rex 19, 1377-1379, reported that infants’ psychometric functions in preferential looking experiments exhibit a region of below chance performance or “negative preference”. They argued from this that previous preferential looking experiments may have systematically underesti- mated acuity because they ignored this negative preference. We present data from an experiment designed to reveal negative preference in the version of the preferential looking paradigm used in our laboratory. The results do not exhibit negative preference apart from random variations about SO”,;. Gwiazda et al. (1980) Am. J. Ogtom. physiol. Opt . 57, 420427; 428-432, developed a psychophysical procedure for infant testing that is designed to yield accurate threshold estimates in a short period of time. We present the results of computer simulations of more conventional psychophysical procedures. The simulations demonstrate that an updown staircase procedure is more efficient than the Gwiazda et al. procedure. INTRODUCTION The preferential looking technique has become widely used to test infant visual acuity. Held et al. (1979) recently reported data which call one of the tech- nique’s basic assumptions into question. This paper and the adjoining one (Teller zyxwvutsrqponmlkjihgfedcbaZYXWVUTSRQPONMLKJIHGFEDCBA et ul., 1982) concern the criticisms raised in the Held et al. report. Held et al. reported that infants’ psychometric func- tions in preferential looking experiments are non- monotonic; that is, with increasing spatial frequency such functions drop from high levels of preference for the grating (approximately 100%) to below 50% before finally asymptoting at 50% preference. We will call this region of below chance performance the negative preference dip. Held et al. argued correctly that per- formance significantly below chance implies discrimi- nation of the grating and blank fields, and hence de- tection of the grating. Because they have ignored pref- erences significantly below SO%, they concluded that they have “systematically underestimated acuity”. Moreover, they “suspect this conclusion is also true of other preference procedures.. .” (p. 1378). Held’s group has developed a psychophysical pro- cedure intended to yield accurate threshold estimates in a short period of time (Gwiazda et (II., 1980b). They claimed that the procedure is quite efficient in studies of normal and clinical populations. The procedure’s reliability is increased significantly by the presence of a negative dip in the infant’s psychometric function (Gwiazda et al., 1980b). This paper makes two specific points concerning the results and procedures described by Held er a/. (1979) and Gwiazda et al. (1980b). First, we present data from an experiment specifically designed to reveal negative preference dips in a conventional *To whom requests [or reprints should be addressed. preferential looking experiment. We used the method- ology currently employed in our laboratory. It is simi- lar but not identical to that of Held et nl. Our motiva- tion was simply to determine if negative preference occurs with the methodology, apparatus, and age ranges we typically use in our laboratory. The results do not exhibit a negative preference dip apart from random variation about 50%. Second, we present the results of computer simulations of conventional up- down, staircase procedures and of the Gwiazda et al. procedure. The simulations indicate that up-down procedures are more efficient; that is, for a given number of trials, the threshold estimates of updown procedures are less variable and biased than those of the Gwiazda et al. procedure. The simulations also indicate that any procedure using small numbers of trials yields relatively unreliable estimates of threshold. Consequently, only substantial differences in acuity can be detected with confidence. METHODS Infants were recruited by letter and telephone. We were primarily concerned with young infants, so only l- to 3-month olds were tested (mean age = 71.0 days, SD = 19.3, range = 38-108). Eleven of the 17 infants who participated provided complete data sets. The others did not complete the experiment due to fussi- ness, sleepiness, or scheduling difficulty. The forced-choice preferential looking technique was employed (Teller, 1979). Stimuli were generated on a large-screen CRT (Hewlett-Packard 1317A) using the method of Campbell and Green (1965). At the 4Ocm viewing distance, the display subtended 48” x 37”. Space-average luminance was 10.6cd/m2 and the surround was dark. Simultaneous presen- tation of a’ squarewave grating and a uniform field was accomplished by splitting the display screen elec- 102s