Inhibitory effects of omission training

In Experiment 1, pigeons were trained to peck red or blue keys for food reinforcement at variable intervals, while food was contingent on withholding key pecks in the presence of a vertical line (omission training). When the line was briefly superimposed on red or blue in a compound test, responding was reduced. When the orientation of the line was varied during extinction, generalization gradients were variable but often had most responding at or near vertical. In Experiment 2, pigeons were trained in a discrete trials procedure that made food contingent upon pecking in the presence of triangle, and upon the absence of pecking in the presence of red (omission training). Food was never given on green-key trials (extinction). When red or green backgrounds were presented with the triangle in a compound test, responding was reduced similarly in the presence of both key colors. Subsequent resistance to auto-shaping was also similar for red and green. These data, taken together with reports in the literature, suggest that the inhibitory effects of omission training are quite similar to those of extinction. Thus, the crucial condition for obtaining inhibitory effects is not a negative stimulus-reinforcer correlation, as in extinction, but simply the establishment of low rates of responding to the inhibitory stimulus.     

Sudden Infant Death or Murder? A Royal Confusion About Probabilities

In this article I criticize the recommendations of some prominentstatisticians about how to estimate and compare probabilitiesof the repeated sudden infant death and repeated murder. Theissue has drawn considerable public attention in connectionwith several recent court cases in the UK. I try to show thatwhen the three components of the Bayesian inference are carefullyanalyzed in this context, the advice of the statisticians turnsout to be problematic in each of the steps.

1 Introduction

2Setting the Stage: Bayes's Theorem

3 Prior Probabilities ofSingle SIDS and Single Homicide

4 Prior Probabilities of theRecurrence of SIDS and Homicide

5 Likelihoods of Double SIDSand Double Homicide

6 Posterior Probabilities of Double SIDSand Double Homicide

7 Conclusion