Variable Analytic Research

Main Point

     Simple effects are hard to find.

I. Variable Analytic Research

     A. Strong Method, Weak Theory

     B. Pick a Characteristic (or Variable), Test Everywhere

     C. Average effect usually equals zero

     D. Effect depends on other factors

II. Source Factors (O'Keefe, 1990)

     A. Credibility

          1. Perception by receiver of source

          2. Expertise

          3. Trustworthiness

          4. Hovland & Weiss, 1951

     Four topics
          Antihistamine drugs should be sold over the counter
          atomic sub can be built now
          business caused the current steel shortage
          TV will cause a decrease in movie theatres
     Four hi-lo credible sources
          New England Journal vs. mass circulation magazine
          RJ Oppenheimer vs. Pravada
          Bulletin of Nat Res Plan vs. Right-wing columnist
          Fortune magazine vs. gossip columnist
     Opinion Change
          23   13     % moving in direction of argument
          36    0
          23   -4
          13   17

          23    7     overall average

          5. Magnitude and Direction Factors

               involvement and timing; position

          6. Why does it work?

               peripheral versus central route processing

     B. Liking

          1. Credibility stronger

          2. High liking operates under low EL

          3. Disliking has dissonance effects and high EL

          4. Important dimension of source effects

     C. Physical Attractiveness

          1. Chaiken, 1980

               Large group of undergrads photoed and rated by 
               other undergrads on attractiveness, upper and 
               lower one-third (n=68, equal M/F) selected; 
               trained to persuade undergrads to sign a petition 
               banning meat at breakfast and lunch in dorms; each 
               person had to contact 2 M and 2 F

               Results showed an attractiveness effect on 
               attitude (3.9 vs 4.4, sd=.9, d=.5) and petition 
               signing (.41 vs .32, d=.4); also attractive people 
               were more fluent and faster in speech, had higher 
               SAT scores (largely with F though), and better 
               self concepts

                  Attractive                Unattractive

    source       M          F              M             F

    receiver  M     F    M     F        M     F       M     F

Agree        4.5  3.3   4.2   3.5      4.9   4.0     4.5   4.0

Sign         .29  .53   .35   .47      .35   .38     .24   .29

Agree scored 1-7, lower score, more agree. Sign is proportion 
signing petition. Agree SD = .9.

III. Conclusions

     A. Source variables operate through crediblity and liking

     B. Source effects not simple and direct

     C. ELM offers best explanation

     D. Don't Lose Credibility

                     Message Characteristics

I. Structure

     A. Climax vs Anticlimax Order (No difference)

     B. Explicit vs Implicit Conclusions (Explicit)

II. Content

     A. One sided vs Two sided (Two sided)

     B. Discrepancy (Curvilinear family; inverted U)

     C. Fear Appeals

          1. hard to induce fear

          2. more fear, more change

     D. Examples vs Statistical Summaries (Examples)

III. Sequential Requests

     A. FITD

     B. DITF

IV. Conclusions

     A. Peripheral Route effects

     B. Not simple and direct

                   Two-Step Message Strategies

Main Point

     Order and comparative size of requests influences compliance 
and attitude.

I.  General Structure of the Two-Step

     A.  Make Two Requests of Receiver

         1. First Request is the Set-up

         2. Second Request is the real Target

     D.  Two Types

          1. Foot-in-the-door (Low, then Target)

          2. Door-in-the-face (High, then Target)


     A. First Request Accepted, then Target Request

     B. Freedman and Fraser, 1966

          TREATMENT:  Door-to-door survey of homemakers, the 
          Committee for Safe Driving, sign a petition encouraging 
          state legislature to support a law that supports safer 
          driving (most people comply).  CONTROL:  not contacted 
          by experimentors.  Several weeks later the same people 
          are surveyed by different experimentors with no 
          reference to earlier contact, asked to put up a very 
          large, unattractive sign in the front yard saying, 
          "DRIVE CAREFULLY."  RESULTS: 55% vs 17% compliance.

     C. Meta-analysis of FITD, Dillard et al., 1984.

          Summary of 28 FITD studies with 37 effect sizes. 5000 
          different subjects.  Effect size = .11 (r) 56% vs. 44%.  
          Two factors important, Unselfish appeal and incentive 
          (r = .17; 59-41).  Self-oriented appeal and performance 
          incentives kill FITD.  Also amount of delay between 
          requests is unimportant, but some delay is present.

     D. Why does FITD work?

          1. Self perception process

               observe own behavior and infer attitude; since I 
               signed the petition, I must be an environmental 
               kind of guy (aren't I cool?); Some support for 
               this in Dillard, 1990, HCR

     A. First Request Refused, then Target Request

     B. Cialdini, et al., 1975

          TREATMENT: County Youth Counseling Program reps asked 
          college adults to spend 2hrs/wk as counselors to JDs 
          for two years; CONTROL asked target request of being an 
          unpaid chaperon of JDs for a day trip to zoo (51% vs 
          17% compliance).

     C. Meta-analysis, Dillard, et al., 1984

          18 studies with 24 effect sizes. 2500 different 
          subjects.  Effect size (r = .08; 54-46).  Two factors 
          are important, Delay and Type of Request.  Any delay 
          kills DITF as does a self-oriented appeal.  No delay 
          with Unselfish appeal is best (r = .15; 58-42).

     D. Why does DITF work?

          1. reciprocal concessions (CLARCCS reciprocity rule)

               you should make concessions to someone who has 
               made a concession to you (You deny first request, 
               then the experimentor makes a concession and asks 
               for a lower request); be polite and compromise

IV. Comparing the two strategies

     A. Both produce about 20% improvement versus control

     B. Observe graph of effects

     C. DITF is clearly easier to do

     D. FITD appears to be longer lasting

V. Interesting Theoretical Problems

     A. Shouldn't same theories apply to each strategy?

     B. Self perception and DITF

          1.   I'm the kind of person who says no to strangers, 
               but now I comply with 2nd request (DITF)

     C. Reciprocal concessions and FITD

          1.   I just complied with the first request, now 
               shouldn't they do something for me?

     D. More theory work needs to be done

                        Receiver Effects

Main Point

     What does not kill you makes you strong.

I.  Inoculation Theory

     A.   Korean War preview

     B.   Stimulated new line of research McGuire

II.  Medical Metaphor

     A.   How does an inoculation work?

     B.   Shot contains mild dose of virus

     C.   Immune system mobilizes to defend

     D.   Defense actually strengthens immune system

     E.   Immune system can defend against stronger viruses

III. Inoculation Theory

     A.   Warning of impending attack

     B.   Weak attack

     C.   Active defense

IV.  Effect of Each Component

     A.   Warning starts defense activation

     B.   Weak attack does not change, but tests person

     C.   Active, self-generated defense best

V.  Why Does It Work?

     A.   Systematic Processing

     B.   Self judo

VI.  Applications

     A.   AT & T phone ads

     B.   Political Inoculation

     C.   DARE anti-drug programs

VII. Conclusions