Diapositive 1

1
Faculté de médecine Bureau dévaluation
Quelques biais possibles liés aux
évaluateurs dans lutilisation des échelles
dappréciation
Serge Normand, M.A.
Octobre 2007
2
Évaluer la compétence clinique
Most patient encounters require the practitioner
to integrate and perform at least 10 separate
components capabilities. 16 hours of performance
observation across a variety of clinical
situations is necessary to achieve reproducible
estimates of clinical competence. All reports
agree that somewhere between 7 and 11 ratings are
necessary to achieve a generalizable global
estimate of competence when raters are basing
ratings on a nonsystematic sample of observations.
Williams, R.G. et al., Cognitive, Social and
Environmental Sources of Bias in Clinical
performance Ratings, Teach Learn Med 200315
(4)270-92
3
Évaluer la compétence clinique Être sensible aux
biais de lévaluation
4
Threats to the validity of clinical teaching
assessments What about rater error ?
Downing, S.M., Medical Education, 2005
In most studies, the variance of raters ( nested
in persons or students ) is the largest variance
component, typically in the 80-90 range. For
example, in a recent study by Kreiter and
Ferguson an average of 89 of the total variance
was attributable to raters (nested in persons),
while only about 10 of the total variance was
due the main effect of persons.
5
Erreur de clémence
Error of Leniency
Je  sympathise  avec lui Je  m identifie  à
lui
Je tente de corriger ce biais
6
Erreur de clémence
Error of Leniency
Correctif
Bon
Très bon
Excellent
Moyen
Faible
7
Effet centripète
The Error of Central Tendency
8
Effet de halo
The Halo Effect

9
Effet de halo
The Halo Effect
10
Erreur de logique
A Logical Error
Lexaminateur présume quil existe une relation
entre deux variables. Si la première est telle,
la seconde sera semblable.
11
Erreur par contraste
A Contrast Error
Moins que moi
Plus que moi
12
Effet de proximité
A proximity Error
Si un observateur évalue deux facteurs
différents, lévaluation de lun a tendance à
influencer lautre, dautant plus que le laps de
temps entre les deux évaluations est plus court.
13
Bibliographie

• Downing, S.M., Threats to the validity of
  clinical teaching assessments What about rater
  error ?, Medical Education 200530350-355
• Guilford, J.P., Psychometric Methods, McGraw-Hill
  Book, New York, 1954, 597p.
• Williams, R.G. et al., Cognitive, Social and
  Environmental Sources of Bias in Clinical
  performance Ratings, Teach Learn Med 200315
  (4)270-92. (pdf)

14
Merci de votre attention !
http//www.medbev.umontreal.ca/docimo
serge.normand_at_umontreal.ca