A Devil’s Dictionary of Behavioral
Science Research Terms

With apologies to Ambrose Bierce
(Appeared in the Journal of Occupational Psychology)

Writing around the turn of century, Ambrose Bierce remarked that “to apology is to lay the foundation for a future offence.”   In that spirit, to those who may perceive the following as imperfectly respectful of the seriousness of our endeavors, an apology is offered in advance.

It is hoped that these definitions contain just enough truth to make us uncomfortable.   Not taking ourselves too seriously helps to retain a sense of perspective, reminds us of the fallibility always present in human endeavors, and thus may serve to strengthen our sometimes feeble attempts at science.

EVALUATION RESEARCH – Research conducted in an organization having a surplus of cash.   Acceptance of the findings is dependent upon the congruence between the reality and the dream.

EXPERIMENTER EFFECTS – All the effects in an experiment.

EX POST FACTO DESIGN – A research design growing out of a consulting contract.  

FIELD EXPERIMENT – An experiment that should have been done in the laboratory.

HOMOGENEITY OF VARIANCE – An assumption commonly made, frequently violated, and never tested.   It does not seem to matter.

HYPOTHESIS – A prediction based on theory formulated after an experiment is performed designed to account for the ludicrous series of events that have taken place.

LABORATORY EXPERIMENT – An experiment more appropriately suited for field research.

LEVEL OF SIGNIFICANCE – An imaginary dividing line between causal effects and chance.   The level of significance serves as a guide for the experimenter in terms of how many replications must be performed before chance falls his or her way.

LINEAR MODEL – An assumption concerning the nature of reality applied unquestioningly to every relationship as though God had determined that truth must always run in straight lines.  

MANIPULATION CHECK – A handy device that allows the researcher to dispose of data from subjects who stubbornly refuse to conform to the experimenter’s perception of reality.

METHODOLOGICALLY UNSOUND – using methodology with which I am unfamiliar.

NON–EQUIVALENT CONTROL GROUP – A control group.

NULL HYPOTHESIS – The type of hypothesis used by a pessimist.

ONE-SHOT CASE STUDY – The scientific equivalent of the four-leaf clover, from which it is concluded all clover possesses four leaves and is sometimes green. 

PARTICIPANT OBSERVATION – A method of gathering data somewhat analogous in degree of objectivity to taking notes while playing outside linebacker.

PLACEBO – The sugar pill of research, often used to discourage a mysterious Mr. Hawthorne from making an untimely appearance.   In the South, pronounced place-bo.

POSTTEST – A measurement made too late.

PRETEST – A measurement made too early.

QUASI - EXPERIMENTAL DESIGN – Experimental design.

 RANDOMIZATION – The assignment of subjects to conditions in an experiment according to some preconceived plan.   Randomness like chastity is more often claimed than maintained. 

RELIABLE – Sometimes capable of giving the same results.   

REPEATED MEASURES – Placing the dice in the cup for another throw. 

REPLICATION – Lightening striking twice in the same place.   Replication is a particularly hazardous undertaking for the fledging experimenter due to the undesirable consequences of failing to reproduce the results of well-known colleague.  

REVIEWER’S NOTE – A rejection slip based upon literature and theories in vogue during the period the reviewer was studying for his or her Ph.D.

SAMPLE – A unique collection of subjects having virtually no chance of being representative of the population from which it was drawn.   This shortcoming is trivial and is generally ignored.

STATISTICAL ANALYSIS – Mysterious, sometimes bizarre, manipulations performed upon the collected data of an experiment in order to obscure the fact that the results have no generalizable meaning for humanity.   Commonly, computers are used, lending an additional air of unreality to the proceedings. 

SUBJECT – Mankind’s equivalent of the white rat.   A victim of science.

TESTS OF SIGNIFICANCE – A ritual performed by worshipers of a Diety known as the “God of Significant Differences.”   The failure of this illustrious Personage to appear in the results of an experiment, even after painstaking observance of the proper rites, has been known to occasion attacks of acute temptation.

UNOBTRUSIVE MEASURES – Experimental techniques of unclear origin having something to do with worn tiles.   Observing madam in her bath without bringing forth screams.  

VALIDITY – There are many types of validity.   The distinctions among them are boring.   Suffice it to say validity issues may be summarized as being chiefly remarkable for the unfair, unrealistic constraints which they place upon the creativity and imagination of the researcher.   

VOLUNTEER SUBJECTS – A college sophomore who, of his or her own free will, is allowed to choose between participating in an experiment or failing a course.   

As reprinted from the Academy of Management Review, 1979, 4, 93-94.