In four experiments human subjects pressed the space bar of a computer keyboard to make a triangle displayed on the monitor of the computer flash. The subjects then rated the causal effectiveness of the response under a number of conditions. In Experiments 1 and 2, subjects experienced four conditions: response immediately followed by outcome; outcome unrelated to response, outcome 5 s after response, and outcome 5 s after response, but in which a 2-s stimulus filled part of the delay. In all experiments responses emitted in the uncorrelated and unsignaled delay conditions were rated as less effective than those in the immediate condition. In Experiment 1, when the stimulus filled the first part of the delay, ratings of causal effectiveness of the response were elevated relative to a condition with no stimulus in the outcome delay. In Experiments 2a and 2b, when the delay stimulus filled the latter part of the delay, ratings of the causal effectiveness of responses in this condition were no different than in the unsignaled delay condition. In Experiments 3 and 4, responses during the delay but prior to the stimulus reset the delay period (ensuring a 3-s delay between response and delay stimulus), and ratings were higher in a condition with a delay stimulus prior to the outcome than without a stimulus. These results are similar to those noted for response rates in non-human conditioning experiments and further extend the parallel between the results from such studies and those from experiments on human causality judgment.