TEST ADMINISTRATION

THE EXAMINER AND THE SUBJECT

 The Relationship Between Examiner and Test Taker - Both the behavior of the examiner and his
or her relationship to the test taker can aff ect test scores. Examiner's rapport with test takers
can influence the results.
 The Race of the Tester - Because of concern about bias, the effects of the tester’s race have
generated considerable attention. Some groups feel that their children should not be tested by
anyone except a member of their own race.
 Language of Test Taker - Some of the new standards concern testing individuals with different
linguistic backgrounds. The standards emphasize that some tests are inappropriate for people
whose knowledge of the language is questionable.
 Training of Test Administrators - Different assessment procedures require different levels of
training. Many behavioral assessment procedures require training and evaluation but not a
formal degree or diploma.
Structured Clinical Interview for DSM-IV (SCID) Typical SCID users are licensed
psychiatrists or psychologists with additional training on the test. There are no
standardized protocols for training people to administer complicated tests such as the
Wechsler Adult Intelligence Scale-Revised (WAIS-R).
 Expectancy Effects - A data sometimes can be affected by what an experimenter expects to
find. Robert Rosenthal and his colleagues at Harvard University conducted many experiments
on such expectancy effects, often called Rosenthal effects. The experimenter’s influence is
not limited to human subjects.
 Effects of Reinforcing Responses . Sattler and Theye (1967) reviewed the literature on
procedural and situational variables in testing and found that an inconsistent use of feedback
can damage the reliability and validity of test scores.
Because reinforcement affects behavior, testers should always administer tests under
controlled conditions. Reinforcement and feedback guide the examinee toward a
preferred response.

Testing also requires standardized conditions because situational variables can aff ect test
scores. Th e book Standards for Educational and Psychological Testing, published by the APA and
other professional groups (AERA, APA, & NCME, 1999), emphasizes that a test manual should
 Computer-Assisted Test Administration - Computer technology aff ects many fields, including
clearly spell out the directions for administration. Th ese directions should be suffi ciently
testing and test administration.
detailed to be duplicated in all situations in which the test is given. A good test manual gives
Hereinstructions
the test examiner are some of thatthe advantages
include that
the exact computers
words offer:
to be read to the test takers. It also
• excellence of standardization,
includes questions that testers will likely ask and instructions on how administrators should
• individually tailored sequential administration,
answer them.
 • precision of timing responses,
• release of human testers for other duties,
• patience (test taker not rushed), and
• control of bias.
 Mode of Administration - A variety of studies have considered the diff erence between self-
administered measures and those that are administered by a tester or a trained interviewer. In
educational testing, it is less clear that mode of test administration has a strong impact
( pencil- paper test, self-administered measures, etc.)
 Subject Variables - A final variable that may be a serious source of error is the state of the
subject. Motivation and anxiety can greatly aff ect test scores.

BEHAVIORAL ASSESSMENT METHODOLOGY

In behavioral observation studies, the observer plays a more active role in recording the data
and, therefore, is much more likely to make errors. Some of the problems include reactivity, drift, and
expectancies.

 Reactivity - The reliability of observers in behavioral observation studies is usually assessed in

selected sessions during which an experimenter “observes the observers.” In other words,
someone looks over the observer’s shoulder to determine whether he or she is recording
properly. Studies have shown that reliability and accuracy are highest when someone is
checking on the observers. Th is increase in reliability is called reactivity because it is a
reaction to being checked.
To deal with this problem, some experimenters resort to covert operations. For example,
the experimenter might randomly check on the performance of the observers without their
knowledge. In general, you should always use caution in interpreting reports on interrater
reliability.
 Drift - When trained in behavioral observation methods, observers receive extensive feedback
and coaching. After they leave the training sessions, though, observers have a tendency to
drift away from the strict rules they followed in training and to adopt idiosyncratic
definitions of behavior.
 Expectancies - Some studies show that administrator expectancies can aff ect scores on
individual IQ tests, whereas other studies do not . Some studies have shown that behavioral
observers will notice the behavior they expect. To avoid this sort of bias, observers should not
know what behavior to expect.
 Deception
 Statistical Control of Rating Errors - Many eff orts to improve the accuracy of raters have
produced discouraging results. Attempts to increase rater reliability through extended
training have been particularly frustrating for many researchers and applied psychologists
because training is expensive and time-consuming.
 Halo effect - Is the tendency to ascribe positive attributes independently of the observed
behavior. Some psychologists have argued that this effect can be controlled through partial
correlation in which the correlation between two variables is found while variability in a third
variable is controlled.