ABSTRACT
Rousseau(1992) de nes three symptoms of pathological science (the excessive loss of scienti c objectivity). The rst is an aversion to crucial experiments that could disprove a favored theory. The second is a disregard for prevailing ideas and theories. Traditional theories are given inadequate consideration as the researcher becomes more and more enamored with a new discovery. These rst two symptoms should seem familiar to the reader as they were discussed in Chapter 1 as lowrisk testing and precipitate explanation. The third symptom was mentioned only very brie³y in that chapter. Rousseau suggests that the last symptom of pathological science often begins with an effect that is “at the limits of detectability or has very low statistical signi cance…. Once the investigator [is] convinced that something new and important has been discovered, the fact that all of the parameters involved … are not under control is viewed as having little consequence” (Rousseau
1992). The improperly controlled or poorly understood measurement process acts as the seed from which increasingly biased behavior grows. Fortunately, the means are available for minimizing such misinterpretations of the results of most measurement processes. Several methods for controlling measurement dif culties in or de ning the limitations of the measurement process are outlined here at the beginning of this book because, in their absence, later methods would be useless.
