第122章
Jot down
any such behaviors and code them as “S” for a speech disturbance. Does the suspect shift
in his seat, cross his legs, twist, wring his hands or put them in his pockets, fidget with
cigarettes, paper clips, etc.? Note these behaviors and code them as “P”, for a physical
movement. Finally, focusing on the suspect’s face, does he smile, frown, wet his lips, close
his eyes, etc? These behaviors can be noted and coded in a general facial expression
category as “F”. Code behaviors not fitting these categories as “O”. If time allows, give
students a chance to practice their scoring, using a volunteer “suspect” from the class.
Have your reaction-timer write the times on the data sheet.
6. When the first suspect returns and knocks on the door, bring him in and seat him in front of the
class with his back to the timekeeper (on a high, backless stool if you have one) and have a student
experimenter give him the following instructions: “I will call out a word and you are to reply
quickly with the first word that comes to mind. We will repeat this for each of 30 words. That is all
there is to it. Is that clear?” (Minimize questions.)
7. If time is a problem, 20 of the 30 words should suffice, but pick half neutral and half critical ones.
Have your timekeeper erase the times before the second suspect comes in.
8. If the first subject is allowed to remain in class while the second is being tested, he should sit
behind the class so as not to give any telling reactions.
9. An excellent extension of this demonstration, proposed by Mikkel Hansen of Stanford University, is
to have students (or associates) film the two suspects completing their tasks. At the end of the class,
after votes have been cast, the videos can be played to the section to dramatically reveal who is
414
innocent and who is guilty.
PITFALLS TO AVOID
Do not get too involved in the initial discussion; this demonstration requires a lot of time, so hold all but the
necessary setting of the context for afterward. Pick a safe place for the burning to take place. Do not select
subjects who are very expressive–the guilty one might give it all away with the first blush.
DATA ANALYSIS AND INTERPRETATION
1. Half of the stimulus words are neutral, in the sense that they are not associated with any aspect of
the crime, whereas half are emotionally loaded, in that they relate to some details of the crime of
which you and the “criminal,” but not the innocent suspect, are aware. Have the class discuss
which words should be counted as “critical”, based on their knowledge of the crime. To help them
with this, you may want to read a description of the crime (the instructions given to the suspect).
Then have them calculate mean reaction times separately for both types of words for each suspect.
2. Mention the use of premeasured RTs for neutral words and its function.
3. Was there a difference in the reaction time of the two suspects to the critical words they had in
common? Repeat this analysis for the other measures. How would you explain the differences you
observed?
4. There may be two complex effects of guilt or emotion on RT: a perseverance effect that carries over to
the next word in the sequence, and a heightened variability effect of giving either much faster or
slower RTs to the critical words. This would yield a mean comparable to the innocent victim, so
different statistical methods might have to be used to assess the significance of this bimodal
reaction tendency (should it occur).
5. Ask students to suggest other ways of analyzing the data to detect guilt. Have them discuss which
measures seem to be doing the best job of predicting guilt. Using the measures they agree on, have
the students predict which of the two suspects is guilty. Each student should make a private verdict
and give an estimate of his or her confidence in the verdict. These data should be tallied and
presented to the class. (Once the verdicts are in, have the two suspects return to the room for a
discussion of the experiment.)
6. Are the three response measures correlated?