Practice and some methodological problems in unidimensional information transmission

Fifty circular lights differir.g in diameter from 4 to 53 cm and increasing by 1 cm were presented 135 times each to five Ss who were asked to identify each circle. Presentations were in random order. Indices of information transmission (T), response equivocation H S(R), and response uncertainty H(R) were calculated for different phases of the experiment for individual and pooled (across Ss) responses. Results have shown systematic rising of transmission and lowering of H S(R), while H(R) remained almost constant. All transmission indices calculated from pooled responses were systematically lower than comparable indices calculated from individual responses. Individual and pooled T values calculated from all responses given during the experiment were not representative for the level of identification performance at the end of the experiment. Conclusions were reached that (1) channel capacity indices should not be calculated from all individual responses given during the experiment or from pooled responses of different Ss, and (2) indices of channel capacity should be calculated only after the performance of Ss has reached a final and stable level, which means after enough repetitions of each stimulus.