Attention and estimated line length

Whether attention affects the estimated length of a line has been debated for a long time. Some authors have found estimated length to increase with attention; others have found that it decreased. The present study further investigated this problem with two experiments. The first confirmed that estimated length decreased with attention; however, this result had low reliability. The second experiment indicated that estimated length significantly decreased with attention for some participants and significantly increased for others. This finding accounts for the low reliability of the first experiment and for the conflicting results of previous studies. Implications of opposite effects of attention for models of sensory intensity are discussed. An interpretation of these effects in terms of response preferences is proposed.