In 1935 Dann¹³ conducted a study to assess the impact of publicity of executions on deterrence. He hypothesized that if the death penalty is a deterrent, its greatest effect should be shown through executions which are well publicized. Furthermore, the effect should be more noticeable in the community where the offence occurred, where the trial aroused wide publicity and the offender lived and had relatives, friends and acquaintances. To test the hypothesis, Dann compiled the dates of executions of Philadelphia residents for a period of several years and was able to find five cases that met the study's specifications. Three of the five cases were of great notoriety. The study found no significant difference in homicide rates for equal periods before and after the execution. There was a total of 105 days free from homicides during the 60-day periods before the executions and 74 in the periods after the executions. There were a total of 91 homicides in the "before the execution" periods and 113 in the "after" periods. Of the 204 homicides included in the study, 19 resulted in sentences for murder in the first degree. Nine of them had occurred during the 60-day periods preceding and ten in the corresponding periods following the executions. During the ten days just before the executions there were two, and during the ten days immediately following there were three such first degree murders in Philadelphia.

Another study was undertaken, also in Philadelpphia, by Savitz.³² After examining homicide trends before and after four widely publicized trials during the 1940's, Savitz concluded that no pattern emerges that would indicate deterrence and that the assumption that the deterrent effect of the imposition of the death penalty might be felt shortly after the date of sentencing is not borne out by the data. He further concluded that on the basis of the data "there was no significant decrease or increase in the murder rate following the imposition of the death penalty on four separate occasions".

Recently, Phillips²⁹ examined weekly murder statistics for the city of London, England for the period 1858-1921 and came to the conclusion that the homicide rate drops during the week of a highly publicized execution and during the following week then it begins rising again.

"Within five or six weeks of a publicized execution, the drop in homicides is cancelled by an equally large rise in homicides".

Phillips results were in contrast to those of Bowers and Pierce⁷ who used monthly rather than weekly murder rates. They found an increase of two murders during the month following a highly publicized execution.