Download Le trouble déficitaire de l`attention - University of Ottawa

A Meta-Analysis of
Stress in Teachers
Cameron Montgomery & André A. Rupp
Faculty of Education
University of Ottawa
Organization of Presentation







Background
Definition of Stress
Theoretical Model of Stress
Theoretical Relationships with other Constructs
Methods for Meta-analysis
Preliminary Results
Conclusion and Discussion
Definition of Stress
Stress is an interaction between a person
and the environment, appraised by the
person as taxing or exceeding his or her
personal resources, and, as a consequence,
disrupting daily routines.
(e.g., Lazarus, 1966; Lazarus & Folkman,
1984; Derogatis, 1987)
Background
High teacher stress is associated with
psychological distress, which is mediated
through coping strategies and personality
characteristics (Chan, 1998).
Specifically, poor active coping skills may lead
to negative emotional responses and,
subsequently, to burnout.
Background (cont.)
Researchers have used varying methods
for explaining the relationships between
psychological stress and other variables
such as coping, personality mediators,
emotional responses, environmental
factors and burnout.
Stress & Coping
According to the transactional model of stress
and coping, problematic events act as triggers for
stressful episodes.
The different stages in coping with stressful
situations are
-Primary Appraisal
-Secondary Appraisal
-Coping (problem-focused, emotion-focused)
Stress & Burnout
Three components of burnout
1. Emotional exhaustion
2. Depersonalization
3. Personal accomplishment
Teachers may experience burnout as a result of
stress itself, as a result of their coping mechanisms
breaking down, or as a result of their coping
mechanisms not being effective over a long period of
time.
Stress & Personality Mediators
Personality characteristics (e.g., Type-A
personality, willpower) influence the degree to
which an individual seeks social support.
Thus, along with demographic characteristics
(e.g., age, gender, ethnicity), they mediate
the ability to establish and maintain
supportive social networks and facilitative
cognitive appraisal of the stressor.
Stress & Personality Mediators (cont.)
Therefore, stress is not viewed as being
exclusively due to situational or personal
characteristics but to the interaction between
them.
This postulate reveals that the seeking of social
support and the engagement in successful
coping strategies can create a virtuous circle
whereby the same objective situation can
begin to appear to be less demanding to the
teacher.
Graphical Model of Stress
The different theoretical models of
teacher stress and related constructs
were used as a basis for mapping out
the empirical relationships that were
observed in the studies included in this
meta-analysis.
Background
Characteristics
Environmental Structure
Personal Support
Personality Mediators
Emotional Coping
Cognitive Coping
Health Posture
Behavioral Coping
Satisfaction
Active
Coping
Positively Oriented
Student
Student Behavior
Behavior
EE
School Structure
Workload
External
Emotional
Stressors
Response
Burnout
DP
Colleagues
Administration
PA
Passive
Personal Life
Negatively Oriented
Coping
Dissatisfaction
Resignation
Wishful Thinking
Avoidance
Personality Mediators
Personal Support
Environmental Structure
Theoretical-empirical Model of Construct Relationships
Objectives
To investigate and summarize the correlational evidence of the
relationships between teacher stress and other related
constructs such as
(a) active coping
(b) passive coping
(c) types of emotional responses
(d) burnout
(e) personality mediators
(f) support systems
(g) environmental structures, and
(h) background characteristics.
Rationale
Examining recent research (i.e., 1998 and later)
through a meta-analytic framework answers
Guglielmi and Tatrow’s (1998) call for a shift
toward more theory-based investigations that
test causal models of teacher stress and
health-related outcomes with more
sophisticated research designs and
measurement strategies.
Sampling Frame
International research databases
PSYINFO, SOCIOFILE.
such
as
ERIC,
Varied keywords such as stress, teacher, coping,
burnout.
Quantitative studies either published (i.e., article,
proceeding) or unpublished (i.e., master’s thesis,
doctoral dissertation).
Sample
59 refereed journal articles
18 dissertations
2 conference proceedings
Research was written in English, French, German,
Chinese, and Arabic and, if needed, professionally
translated.
Number of Studies
Country
Breakdown of
studies by country
of primary data
collection
Australia
4
Australia & Scotland
1
Belgium
1
Canada
5
China
2
Finland
1
Georgia
1
Germany
10
Greece
2
Hong Kong
4
Israel
1
Italy
1
Japan
4
Netherlands
3
New Zealand
2
Pennsylvania
1
Scotland
1
Singapore
2
Slovenia
1
South Africa
1
Spain
1
Sweden
2
UK
4
USA
22
USA & Jordan
1
Descriptive Statistics
Median sample size = 184
(reported by n = 75 studies)
Median % of males = 39.9%
(reported by n = 41 studies)
Median average age = 41.3 years
(reported by n = 30 studies)
Median average number of years of experience = 16 years
(reported by n = 27 studies)
Coding
After entering all bivariate relationships into a data file,
variable names were recoded and regrouped using
theoretically-driven labels.
This ensured a mapping of the empirically observed
variable labels onto the construct labels represented in the
graphical model shown earlier.
Coding (cont.)
After eliminating entries with non-missing effect-size information that
could not be computed by hand and eliminating partial coefficients
(e.g., regression coefficients, interaction effect coefficients), we obtained
2,527 distinct effect sizes for bivariate relationships, which could be
broken down as follows:
- 2,061 Pearson correlation coefficients
- 62 Spearman correlation coefficents
- 6 Pearson point-biseral correlation coefficients
- 134 independent-samples t-test statistics
- 264 F-test statistics (mostly representing t-tests)
Statistical Analysis
1.
Effect-size statistics were transformed onto a correlational
metric
2.
Correlation coefficients so obtained were subsequently
Fisher-tranformed to the standard normal metric, which
possesses more robust statistical properties.
3.
Standard errors, weights, and 95% confidence intervals were
computed on that metric.
4.
Confidence interval boundaries were transformed back to the
original correlation metric.
Notes. Values in the lower off-diagonal are average effect size statistics on a correlation metric with 95% confidence intervals
underneath.
Values in the upper-off diagonal are the total sample size of all independent samples (N)
and the number of effect sizes on which the averages are based (K).
Stress
Active
Coping
Passive
Coping
Emotional
Responses
Burnout
Personality
Mediators
Support
Environment
Structure
Background
Stress
1
N = 6280
K = 179
N = 1761
K = 25
N = 6254
K = 128
N = 4453
K = 124
N = 7941
K = 99
N = 5798
K = 61
N = 5166
K = 92
N = 15360
K = 133
Active Coping
.2025
(.1942,
.2107)
1
N = 1761
K = 41
N = 3274
K = 92
N = 4217
K = 123
N = 5562
K = 77
N = 4966
K = 55
N = 2491
K = 24
N = 13358
K = 93
Passive
Coping
.0751
(.0475,
.1025)
.1544
(.1327,
.2759)
1
N = 1626
K = 17
N = 1626
K = 27
N = 1540
K=7
N = 1540
K=8
N = n/a
K = n/a
N = 6950
K = 16
Emotional
Responses
.2512
(.2413,
.2611)
.0500
(-.0628,
.1615)
.0950
(.0611,
.1287)
1
N = 2447
K = 69
N = 5402
K = 79
N = 3910
K = 42
N = 1446
K = 18
N = 10267
K = 57
Burnout
.2673
(.2565,
.2780)
.2685
(.2575,
.2793)
.0858
(.0583,
.1131)
.3977
(.3846, .4106)
1
N = 4821
K = 49
N = 3700
K = 39
N = 3112
K = 33
N = 12464
K = 116
Personality
Mediators
.2535
(.2433,
.2635)
.1624
(.1489,
.1760)
.1031
(.0531,
.1527)
.3033
(.2080, .3930)
.2746
(.2600,
.2890)
1
N = 4837
K = 24
N = 2967
K = 19
N = 11667
K = 18
Support
.2604
(.2475,
.2731)
.1503
(.1355,
.1650)
.0608
(.0105,
.1108)
.2671
(.2495, .2844)
.2357
(.2188,
.2525)
.2302
(.2095,
.2507)
1
N = 549
K=8
N = 9683
K = 28
Environment
Structure
.1914
(.1806,
.2021)
.2067
(.1867,
.2265)
n/a
.2772
(.2488, .3052)
.1781
(.1609,
.1951)
.2225
(.2047,
.2401)
.3342
(.2991,
.3683)
1
N = 448
K=3
Background
.1120
(.1035,
.1205)
.0909
(.0820,
.0999)
.1292
(.1085,
.1497)
.0631
(.0506, .0756)
.0728
(.0642,
.0813)
.1199
(.1037,
.1361)
.0616
(.0438,
.0794)
.1400
(-.0634,
.3322)
1
Results
All average correlations are statistically
significant (p < .05) except for the average
correlations between active coping and
emotional responses, as well as between
background characteristics and environmental
structures.
Results: External stressors
External stressors are moderately
correlated with burnout (r = .26),
support variables (r = .26), personality
mediators (r = .25), and emotional
responses (r = .25).
Results: Active coping
Active coping was most strongly
correlated with burnout (r = .26).
Weaker correlations were found
between active coping and
environmental structure (r = .20)
Results: Passive coping
In terms of passive coping, average
correlations were generally low.
The construct that was most strongly
correlated with passive coping was
background characteristics (r = .12).
Results: Emotional responses
A high average correlation was observed
between emotional responses and burnout
(r = .39).
Moderate average correlations were
observed between emotional responses and
personality mediators (r = .30), environmental
structure (r = .27), and support (r = .26).
Results: Personality mediators
Average correlations of personality
mediators with other variables are
moderate as shown with, for example,
the average correlations between
personality mediators and support
(r = .23) and environmental structure
(r = .22).
Results: Support variables
In terms of support variables, average correlations
were more varied depending on the related construct.
For example, while a moderate average correlation
exists between support variables and environmental
structure (r = .33), there is only a low correlation
between support and background characteristics
(r = .06) showing that the perceived support available
to individual teachers does not vary systematically
with background characteristics such as sex or
educational level.
Results: Burnout
A strong average correlation exists between
burnout and emotional responses (r = .41).
A moderate average correlation was found
with personality mediators (r = .27) but a
basically nonexistent average correlation was
found with background characteristics
(r = .07) indicating that general personality
traits are more strongly affecting burnout
directly than general background
characteristics.
Background
Characteristics
Environmental Structure
Personal Support
Personality Mediators
Emotional Coping
Cognitive Coping
Health Posture
Behavioral Coping
Satisfaction
Active
Student
Student Behavior
Behavior
Coping
.2025; p < .05
(.1942, .2107)
.0500; n.s.
(-.0628, 1615)
Positively Oriented
EE
School Structure
Workload
.2926; p < .05
(.2775, .3076)
External
Emotional
.3977; p < .05
(.3846, .4106)
Burnout
Response
Stressors
DP
Colleagues
Administration
Personal Life
PA
Passive
.0751; p < .05
(.0475, .1025)
.0950; p < .05
(.0611, .1287)
Coping
Negatively Oriented
Dissatisfaction
Resignation
Wishful Thinking
Avoidance
Personality Mediators
Personal Support
Environmental Structure
Stress & Burnout
.2673; p < .05
(.2565, .2780)
Selected Average Correlations in Theoretical-empirical Model
Discussion
External stressors may be moderately influencing
burnout directly with support from family and other
colleagues playing a mediating role with similarly
moderate effects.
Emotional responses such as anxiety, depression,
and disappointment are also moderately directly
affected by stress or moderately affected by active
coping as a mediating variable; however, they are not
very much affected by passive coping mechanisms.
Discussion (cont.)
How one actively copes through exercise (i.e.,
through one’s health posture), through concrete
behavioral strategies, and cognitive planning, and
through using appropriate emotional coping
mechanisms in the face of various stressors may
moderately determine if one will indeed feel
emotionally exhausted, depersonalized, or not
accomplished.
Discussion (cont.)
The strongest association, generally, is between
external stressors (i.e. student behavior, school
structure, workload, colleagues, administration,
personal life) and negatively- and postively-oriented
emotional responses.
Discussion (cont.)
The highest average correlation observed between
two constructs was between emotional responses
and burnout (r = .39).
This shows that the degree in which teachers
emotionally respond to stressful events and how
satisfied they are as a consequence, either mediated
through coping mechanisms or not, has a strong
influence on the degree of burnout they experience.
Discussion (cont.)
Yet, at the same time, the degree to which one
engages in active coping does not vary
systematically with background characteristics and
does not systematically predict directly the emotional
responses to stressful events, even though these are,
in themselves, moderately correlated to stressful
events as a cause and burnout as a result.
Discussion (cont.)
Emotional responses and personality mediators are
closely associated suggesting that the way teachers
respond to a variety of stressful situations is closely
tied to the relatively stable personality traits that
mediate reponses.
The environmental structure and the support
structure also influence quite strongly how people
emotionally respond to stressful events, which, in
turn, influences the degree of burnout they
experience.
Discussion (cont.)
The majority of emotional responses are negativelyoriented (e.g. distress, anxiety, depression, suicidal
ideation).
This implies that the moderate to high average
correlations between external stressors and these
responses are relfective of the fact that exposure to
stressful events leads to negative experiences to
teachers, be they mediated through coping
mechanisms or not, which, in turn, may lead to
different types and magnitudes of burnout.
Discussion (cont.)
Since environmental structures and support structures
also influence teachers’ emotional responses, there is
evidence in the literature that primary importance should
be given to the precise examination of emotional
responses within the stress cycle.
A Meta-Analysis of
Stress in Teachers
Cameron Montgomery & André A. Rupp
Faculty of Education
University of Ottawa
For a copy of the paper contact
[email protected] or [email protected]