Investigator Effects

Investigator effects are those sources of artifact or error in scientific inquiry that derive from the investigator. It is useful to think of two major types of effects, usually unintentional, that scientists can have upon the results of their research. The first type operates, so to speak, in the mind, eye, or hand of the scientist. It operates without affecting the actual response of the human participants or animal subjects of the research. It is not interactional. The second type of investigator effect is interactional. It operates by affecting the actual response of the subject of the research.

Noninteractional effects include (a) observer effects, (b) interpreter effects, and (c) intentional effects.

• (a)
  Observer effects refer to errors of observation made by scientists in the perception or recording of the events they are investigating. The analysis of a series of 21 studies involving 314 observers who recorded a total of 139,000 observations revealed that about 1% of the observations were in error, and that when errors were made, they occurred two thirds of the time in the direction of the observer's hypothesis.
• (b)
  Interpreter effects refer to differences in the theoretical interpretations that different scientists give to the same set of observations. For many years, for example, investigators of the effectiveness of psychotherapy disagreed strongly with one another in the interpretation of the available studies. It was not until systematic, quantitative summaries of the hundreds of studies on the effectiveness of psychotherapy became available that this particular issue of interpreter effects became fairly well resolved.
• (c)
  Intentional effects refer to instances of outright dishonesty in science. Perhaps the most common example is simple data fabrication, in which [Page 528]nonoccurring but desired observations are recorded instead of the honest observations of the events purported to be under investigation.

Interactional investigator effects include (a) biosocial, (b) psychosocial, (c) situational, (d) modeling, and (e) expectancy effects.

• (a)
  Biosocial effects refer to those differences in participants' responses associated with, for example, the sex, age, or ethnicity of the investigator.
• (b)
  Psychosocial effects refer to those differences in participants' responses associated with, for example, the personality or social status of the investigator.
• (c)
  Situational effects refer to those differences in participants' responses associated with investigator differences in such situational variables as research experience, prior acquaintance with participants, and the responses obtained from earlier-contacted participants.
• (d)
  Modeling effects refer to those differences in participants' responses associated with investigator differences in how they themselves responded to the task they administer to their participants.
• (e)
  Expectancy effects refer to those differences in participants' responses associated with the investigator's expectation for the type of response to be obtained from the participant. Expectations or hypotheses held by investigators in the social and behavioral sciences have been shown to affect the behavior of the investigator in such a way as to bring about the response the investigator expects from the participant. This effect, referred to most commonly as the EXPERIMENTER EXPECTANCY EFFECT, is a special case of interpersonal expectancy effects that has been found to occur in many situations beyond the laboratory.