Transcript iaps_18_2004_619.slides

Reducing Use of Toxic Household
Products Through Guided Group
Discussion
Carol M. Werner, Sari Byerly,
& Carol Sansone
University of Utah USA
Paper presented at iaps2004, Vienna
Environmental Behavior Change


No Silver Bullet
Holistic approach
individual & supportive context
social milieu (friends, society)
political/economic system
physical environment
Supportive Context


Political-economic system: Are there
mechanisms to support the new behavior?
(nontoxic alternatives; health department
education program; HHW)
Physical environment. Does the physical
environment support the new behavior? (making
it easy to use nontoxics and hard to use toxics)
Individual


Strong Attitudes Predict Behavior.
Strength of attitude related to depth of
processing and attitude accessibility.
Social Milieu

Perceived opinions of:
Immediate Friends/Family
Larger Social milieu
TV, radio, print:
advertising, commentary

Social Processes:
“False consensus” (believe others agree w/them)
“Pluralistic ignorance” (disagree, but fear rejection)
Hearing others endorse new behavior opens the individual
to change.
Creating positive social milieu:
A route to individual attitude
change


Guided group discussions (Lewin)
Not a lecture:




Group members endorse new idea
Group members discuss problems and solutions
Leader guides discussion in support of nontoxics
DV=attitudes & intended behaviors (11-pt
scales, >6 is positive)
Results: Community Groups

Table 1. Attitudes and Behaviors Since the Meeting

ALL ORGANIZERS
(n = 46)










PERSONAL ATTITUDES AND BEHAVIORSa
Took things to HHW facility?
Shared leftovers?
Begin/continue sharing?
Important to reduce use
Plan to use nontoxicsb
33%35%
36%
6.6
9.7
8.4
MATCHED SAMPLE
Organizer
35%
5.3
9.7
8.7
10%*
n
Control
20
12%*
3.6*
9.0*
7.2*
17
18
23
23
* matched groups differ at p < .05, 1-tailed dependent t-tests
Column one shows responses of all 46 organizers (for comparison to the reduced sample). Columns two and three show the subgroup of organizers with their matched
controls; n’s for the subgroups are in parentheses; statistical tests compare the subgroup of 23 organizers with their matched controls.
a Percentages indicate percent of respondents saying “yes” to that item. Other items were rated on 1-11 scales, with ends labeled “Extremely Unlikely/Extremely Likely,”
“Extremely Unimportant/Extremely Important,” or “Extremely Unsatisfied/Extremely Satisfied.”
b Mean of three items: 1) likely to use more nontoxics around the home; 2) likely to use nontoxics to care for landscaping; and 3) satisfaction with nontoxic alternatives.
Results: Community Groups

Table 1. Attitudes and Behaviors Since the Meeting

ALL ORGANIZERS
(n = 46)


ESTIMATES OF GROUP’S
ATTITUDES AND BEHAVIOR

Group valued meeting
9.4

Group shared leftovers
24%

Group begin/continue sharing? 5.7




MATCHED SAMPLE
Organizers
9.2
Control
Not asked
n
23
27%
No answers
22
6.0
4.2*
12
* matched groups differ at p < .05, 1-tailed dependent t-tests
Column one shows responses of all 46 organizers (for comparison to the reduced sample). Columns two and three show the subgroup of organizers with their matched controls; n’s for the
subgroups are in parentheses; statistical tests compare the subgroup of 23 organizers with their matched controls.
a Percentages indicate percent of respondents saying “yes” to that item. Other items were rated on 1-11 scales, with ends labeled “Extremely Unlikely/Extremely Likely,” “Extremely
Unimportant/Extremely Important,” or “Extremely Unsatisfied/Extremely Satisfied.”
b Mean of three items: 1) likely to use more nontoxics around the home; 2) likely to use nontoxics to care for landscaping; and 3) satisfaction with nontoxic alternatives.
Replication: High School Classes

True experiment:
Lecture vs. Guided Discussion
Random assignment to treatment

Is guided group discussion more
effective than a lecture format?
Mediation Analysis
Relevant
Discussion
?
Attitude
change
WHY?
WHAT PSYCHOLOGICAL PROCESSES MIGHT BE ACTIVATED?
INCREASED PROCESSING OF STRONG MESSAGE?
MORE LEARNING? ACTIVE LEARNING?
PERCEIVED GROUP ENDORSEMENT?
Strategy for showing “Why”
Mediation analysis
What psychological process occurred?
Do perceptions that group agrees with
message mediate attitude change?
Three Steps to Mediation:
1. Does the treatment affect outcome? (is there an
effect to be mediated?)
2. Does the treatment affect proposed mediator?
(did the treatment activate the mediator?)
3. Is the treatment effect reduced or eliminated
when the mediator is added to the analysis?
Design: 2(lecture/guided group discussion) by
2(relevance: low/high)
PREDICTORS
Discussion vs. lecture
Topic Relevance:
How many products do you choose (vs. parents choose for you)
2 groups, “little choice” vs. “some/complete choice”
Initial attitude:
How favorable are you towards nontoxic alternatives?
Single item, 7-pt. scale
Preliminary results (22 classes, 300 students)
DV
Post-meeting Attitude towards Nontoxics:
6-item scale: effectiveness of nontoxics, importance of using
nontoxics, likelihood of using a nontoxic, interest in learning more,
no problem using nontoxics, concerns re: toxics and health; alpha
= .76
PROPOSED MEDIATORS:
“Perceived group endorsement”
5-item scale, similar to above, “what would your classmates say?”;
alpha = .77.
Cognitive Elaboration (positive-negative comments)
“what were you thinking about during presentation?” (inter-rater r =
.86).
Step 1. Something to be mediated:
Discussion increased attitude change, when topic relevant
Initial attitude
Discussion
vs. lecture
Topic relevance (.00)
Discussion x relevance
Class was not significant.
F(5, 294) = 9.70, p < .001.
*p < .05 +p < .10
Attitude towards
nontoxics
PREDICTED “ATTITUDE TOWARDS
NONTOXICS”
Non Relevant
Relevant
Lecture
-.03
-.02
Discussion
.02
.26
Step 2a. Potential Mediator “perceived endorsement” is activated by Treatment,
especially when topic is relevant:
After Discussion, students said “group endorsed nontoxics”
Initial attitude
Discussion
vs. lecture
Topic relevance (-.04)
Discussion x
relevance
.17*
Perceived group
endorsement
Predicting “perceived group endorsement”
Class was not significant.
F(5, 294) = 6.01, p < .001
*p < .05 +p < .10
Step 2b. Potential Mediator “elaboration” is not clearly activated by Treatment,
even when relevant: Small differences for “positive minus negative” thoughts
Initial attitude
Discussion
vs. lecture
Topic relevance
Discussion x
relevance
-.04
Predicting cognitive elaboration.
Class was not significant.
F(5, 294) = 1.70, p > .10
*p < .05 +p < .10
Cognitive elaboration
Step 3. Adding mediators to predictors of attitudes.
Discussion of relevant information leads to attitude change
because students believe others endorse new information
Initial attitude
Discussion
vs. lecture
.08
Attitude towards
nontoxics
Topic relevance
(.03)
Discussion x
relevance
Perceived group
endorsement
Cognitive
elaboration
Coefficients (ßs) in red are from the final analysis.
Class was not significant
F(7, 292) = 17.14, p < .001.
*p < .05
+p < .10
Step 3. “Mediation” and “Partial Mediation” (smaller ß’s)
Discussion leads to attitude change because
students believe others endorse new information
Initial attitude
Discussion
vs. lecture
(.13*)
.08
Attitude towards
nontoxics
Topic relevance
(.02)
Discussion x
relevance
Perceived group
endorsement
Cognitive
elaboration
Coefficients (ßs) in red are from the final analysis,
those in green are from Step 2, and those in
parentheses are from Step 1. Class not sig.
F(7, 292) = 17.14, p < .001.
*p < .05
+p < .10
Discussion
PGE
Attitude Change
PERCEIVED GROUP ENDORSEMENT (PGE)
MEDIATED ATTITUDE CHANGE WHEN TOPIC
RELEVANT.
Implications/Discussion



Environmental behaviors are social
behaviors.
Attitude and behavior change require
social support, e.g., Staats’ EPT program.
Results suggest cognitive elaboration was
activated equally for all groups.