Others, however, characterize the correlation as inherently artifactual because the variables being correlated have a common term (number of crimes is the numerator of the crime rate and the denominator of the objective certainty variable). Still others argue that the correlation is not inherently artifactual but is nevertheless spurious because of measurement error. This paper shows that the negative correlation is not inherently artifactual and provides evidence to support the measurement error interpretation. Unfortunately, however, there is no definitive way to demonstrate whether the negative correlation between the crime rate and the objective certainty of punishment reflects deterrence or merely measurement error. 3 tables and 26 references. (Author abstract)