Attentional Processing of Visual Stimuli: What Makes a Face a Face?

Reaction time (RT) and accuracy have been used to investigate the manner in which humans process visual stimuli for many years. Some researchers have found evidence of serial processing (e.g., Sternberg, 1969a) while others have shown that some visual stimuli are processed in parallel (e.g., Neisser, 1963). There have been very few studies investigating the processing of schematic faces with respect to RT and accuracy. The present study investigates the mechanisms associated with the recognition of schematic faces with respect to RT and accuracy. Forty-two Old Dominion University students were shown a target face for ten seconds and were then shown a series of display sets varying on 4, 5, or 6 features. The display set contained either 4, 6, 8, or 10 faces, and half of the trials were conditions where the target face was in the display set and half of the trails did not contain the target face. It was hypothesized that subjects would show serial search strategies for increases in set size, parallel strategies if the faces were ecologically valid, and learning effects over blocks of trials. Two ANOVAs were performed on the data. First, 3 levels of features x 4 levels of display set size x 2 levels of trials x 2 levels of display set type x 2 levels of block were analyzed with the respect to the dependent measures RT and accuracy. For RT, 4 main effects and 8 interactions were found to be significant. For accuracy, 2 main effects and 4 interactions were found to be significant. Second, 2 levels of display set type x 3 levels of minimal difference (DMIN), a variable reflecting the closest non-match to the target face, were analyzed with respect to RT and accuracy. For both RT and accuracy, the main effect for DMIN and the interaction was found to be significant. The results supported the serial processing and learning effects hypotheses. Although there was an indication that parallel processing might have occurred, a RT-accuracy trade-off negated the support for the parallel processing hypothesis. Future research could include investigations into the salience of features, an increase in the ecological validity of the faces, and variations in the presentation of stimuli.