Valence and Saliency of commonly used smoking-related words

advertisement
Valence and Saliency
of commonly used
smoking-related words
Louise Hopper & Dr Michael Gormley
Trinity College, Dublin
1
Presentation outline

Attentional Bias and Addiction
• Addiction-Stroop Task
• Importance of valence and saliency

Research aims and design

Results

Conclusions
PSI Annual Conference, 2011
2
Attentional Bias in Addiction

Extensive evidence supporting the influence of
attentional bias has been found across a range
of addictions, including
•
•
•

Alcohol (Cox et al., 2002)
Nicotine (Gross, Jarvik & Rosenblatt, 1993)
Opiates (Franken, Kroon, Wiers & Jansen, 2000)
Has been suggested that these biases are
•
•
•
A good predictor of clinical outcome (Waters et al., 2003)
An important factor in drug relapse (Shiffman et al., 1996)
And amenable to therapeutic intervention (Munafo et al.,
2003)
PSI Annual Conference, 2011
3
Addiction-Stroop Task

Stroop Task (Stroop, 1935)
•
•
Requires the colour of a word to be identified, and
responded to, while ignoring its semantic content
When the content is especially salient to the individual,
colour naming is slower than when the content is neutral.
This has been termed a modified-Stroop effect and it is
common in emotional disorders and in addiction.

Historically word tasks predominately based on
lists of words from a small number of studies

Recent increase in the number of studies based
on addiction-related images
PSI Annual Conference, 2011
4
Addiction-Stroop Task

Now a widely used measure of attentional bias
and can differentiate users from controls
Alcohol (Bauer & Cox, 1998; Lusher, Chandler &
Ball, 2004)
 Cannabis (Field et al., 2004)
 Opiate (Marissen et al., 2006)
 Cocaine (Hester & Garavan, 2006)


BUT in some instances both users and controls
can demonstrate a Stroop effect
•
Smoking (Fehr et al., 2006; Johnson et al., 1997)
PSI Annual Conference, 2011
5
Valence in Addiction Cues

Drug users have been found to react
differently to positive and negative related
stimuli

Alcohol: (Armstrong, 2001)
• Stronger associations found with positive, in
•
comparison to negative, alcohol-related words
These differences differentiate addicts from nonaddicts but also heavy levels of alcohol
dependence from light.
PSI Annual Conference, 2011
Smoking-related Images

(van Hanswick de Jonge & Gormley, 2005)
• Smokers rate positively valenced smoking-related
•

images more positively than both never- and exsmokers
No differences were found for negatively
valenced images
No studies to date have examined the
valence of smoking-related words to see if a
similar effect occurs with this task
PSI Annual Conference, 2011
Smoking-related Words
Zack
(1999)
Ash, carton, butt, exhale, drag, lighter, pack,
matches, puff, menthol, smoke, tobacco, filter,
nicotine, inhale, cigarette
Waters & Smell, coffee, inhalation, bronchitis, habit, cancer
Feyerabend
(2000)
Irate, urge, craving, angry, upset, uneasy, hungry,
restless, anxious, tense, stressed, depressed,
withdrawal bored, tired, irritated, agitated, awkward,
related
desperate
Mogg &
Bradley
(2002)
Fags, ashtray, cigar
PSI Annual Conference, 2011
8
Research Aims

Determine if smokers show the same
positivity bias for smoking-related words as
has been seen for smoking-related images

Investigate if this positivity bias differs with
age?

Create a set of smoking words, validated on
perceptual salience and smoking valence,
with different age groups
PSI Annual Conference, 2011
9
Research Design
Participants
Between
Group
(Adolescents)
(Young Adults)
(Older Adults)
Repeated
Measures
n=177
Smoker
Non-Smoker
n=84
n=93
Rating Type:
(1) Valence, (2) Saliency
Age 15-18
Age 15-18
n=28
n=35
Age 19-25
Age 19-25
n=27
n=29
Age 26-40
Age 26-40
n=29
n=29
PSI Annual Conference, 2011
Valence:
(1) Positive (2) Negative
Smoking Words
36 smoking-related words from previous studies
Fags
Weight
Nicotine
Irate
Pack
Cigarette
Addiction
Lung
Smell
Ashtray
Inhale
Anxious
Bored
Habit
Tense
Puff
Drag
Tired
Filter
Menthol
Cancer
Butt
Urge
Exhale
Tobacco
Craving
Desperate
Carton
Ash
Slim
Bronchitis
Stressed
Upset
Smoke
Cigar
Matches
PSI Annual Conference, 2011
11
Word Ratings
Participants were asked to:
‘Think about each of the words listed in relation to
smoking’

Valence : assessed using the question
PSI Annual Conference, 2011
12
Word Ratings

Salience : assessed using the question

Different formats were used for the two response
scales so as to discourage automatic responding
•
The lower the score on this scale, the more closely the
word related to smoking.
PSI Annual Conference, 2011
13
Participant Characteristics
Adolescent
(15-18)
Young Adult
(19-25)
Adult
(26-40)
Smoker
Non-S
Smoker
28
35
27
29
29
29
Mean Age
17.47
16.68
23.26
22.80
33.30
32.42
Age Started Smoking
14.11
15.63
15
Age Daily Smoker
13.18
17.54
16.91
Cig per Day
6.78
8.39
11.74
Cig per Week
49.82
58.83
79.07
Dependence (FTND)
2.07
2.04
2.41
n
PSI Annual Conference, 2011
Non-S Smoker Non-S
14
Analysis 1: Valence & Saliency

Standardised ratings (z-scores) were used
in this analysis to facilitate comparison

Ratings were then analysed by
• Smoking status (smoker, non-smoker)
• Age (adolescent, young adult, older adult) and
smoking status
PSI Annual Conference, 2011
15
Results : Smoking Status
2 (smoking status) X 2 (rating type) MF ANOVA
Smoker
Non-smoker
Mean standardised rating
0.5

Significant smoking
status by rating type
interaction (p<.001)

Smokers gave
significantly more
positive feeling
ratings than nonsmokers (p<.001)

There was no
significant difference
in saliency ratings
0.4
0.3
0.2
0.1
0
-0.1
-0.2
-0.3
Valence
Saliency
Rating Type
PSI Annual Conference, 2011
16
Results : Smoking Status & Age
2 (smoking status) X 3 (Age) X 2 (rating type) MF ANOVA
Valence
Saliency

Significant smoking
status by rating
interaction as before
(p<.001)

No other significant
interactions

Age-related decline
in positivity and in
saliency ratings BUT
these differences
were not significant
in post-hoc analysis
Mean standardised rating
0.25
0.2
0.15
0.1
0.05
0
-0.05
-0.1
Adolescent Young Adult Older Adult
Rating Type
PSI Annual Conference, 2011
17
Analysis 2: Valence ratings

For each participant, means were
calculated for the 10 most positive and the
10 most negative words

Top 10/Bottom 10 scores were analysed
• By smoking status
• By age and smoking status
PSI Annual Conference, 2011
18
Results : Smoking Status
2 (smoking status) X 2 (valence) MF ANOVA
Smoker
Significant smoking
status by valence
interaction (p<.001)

Smokers gave
significantly higher
ratings for both types
of words than nonsmokers (p<.001)

Both groups gave
higher ratings for
positive words than
for negative (p<.001)

Size of the difference
(Top10 – Bot10) was
significantly larger in
smokers (p<.001)
Non-smoker
7
6
Mean word rating

5
4
3
2
1
0
-1
Top 10
Bot 10
Words
PSI Annual Conference, 2011
19
Results : Smoking Status & Age
2 (smoking status) X 3 (age) ANOVA run for each measure
Significant smoking status X age X valence interaction (p = .001)
Top 10 (Positive)
8
7.5
7
6.5
6
5.5
5
4.5
Non-smoker
2.75
Mean word rating
Mean word rating
Smoker
Bottom 10 (Negative)
Smoker
Non-smoker
15-18
19-25
Age Groups
PSI Annual Conference, 2011
26-40
2.5
2.25
2
1.75
1.5
1.25
15-18
19-25
Age Groups
26-40
20
Results : Smoking Status & Age
Top 10 (Positive)
Smoker

Non-smoker
8
Mean word rating
7.5
Smokers are significantly
more positive than nonsmokers at each age
level
•
7
6.5
•
6
5.5
5
4.5
•
15-18
19-25
Age Groups
PSI Annual Conference, 2011
26-40
No main effect of age
among non-smokers
Adolescent smokers are
more positive than both
young (p<.05) and older
adult smokers (p<.005)
There is no significant
difference between the two
adult groups
21
Results : Smoking Status & Age
Bottom 10 (Negative)
Smoker
Non-smoker

2.75
•
•
2.5
Mean word rating
Main effect of smoking
status only (p<.001)
2.25
No main effect of age
No Interaction
2
1.75

1.5
1.25
15-18
19-25
Age Groups
PSI Annual Conference, 2011
26-40
So, age only influences
ratings for the positive
smoking-related words
22
Conclusion: Smoking Status

Distinct differences between smokers and
non-smokers emerged when considering
positive smoking-related stimuli
• Suggesting that positive associations with
smoking are especially strong in smokers
• And that both smokers and non-smokers are
aware of the negative connotations of smoking
PSI Annual Conference, 2011
23
Positive and Negative Words
Zack
(1999)
Ash, carton, butt, exhale, drag, lighter, pack,
matches, puff, menthol, smoke, tobacco, filter,
nicotine, inhale, cigarette
Waters & Smell, coffee, inhalation, bronchitis, habit, cancer
Feyerabend
(2000)
Irate, urge, craving, angry, upset, uneasy, hungry,
restless, anxious, tense, stressed, depressed,
withdrawal bored, tired, irritated, agitated, awkward,
related
desperate
Mogg &
Bradley
(2002)
Fags, ashtray, cigar
PSI Annual Conference, 2011
24
Conclusion: Status and Age

No effect of age for non-smokers

Adolescent smokers rated smokingpositive but not smoking-negative words
significantly more positively than both
adult groups

Smokers demonstrate a positivity bias
towards smoking-related words, notably
positive words, which declines with age
PSI Annual Conference, 2011
25
Future Research

Consider and control for the valence of
addiction-related stimuli

Investigate the impact of differentially
valenced stimuli on responding in implicit
addiction tasks

Overall aim is to determine how positivity bias
contributes to the development and
maintenance of a smoking-related addiction
PSI Annual Conference, 2011
26
Thank you for your
attention
Louise Hopper
School of Psychology, TCD
hopperl@tcd.ie
Questions
27
Download