RALPH BALEZ Dr en psychologie sociale, Ingénieur Pédagogique, formateur en simulation médicale

Home page > Informations générales / General information > Pygmalion in the laboratory: contribution to the modelling of the implicit (...)

un résumé de ma thèse en anglais / a summary of my Phd. in english

Pygmalion in the laboratory: contribution to the modelling of the implicit influence of the experimenter’s theoretical expectations on the participants

Wednesday 5 January 2011, by Ralph Balez

KEY WORDS : “Pygmalion effect”, meta-experiment, experiment, experimenter’s expectations, behavioural confirmation, expectations induction, hypothesis’ confirmation, self-fulfilling prophecies.

See online : To acquire a condensed version of my thesis (french version).


The researchers expectations, and his/her explicit and implicit behaviour combine to form a certain definition of the experimental situation. The researcher may transmit his/her theoretical expectations to the subjects of his/her research during the course of an experiment. The subjects may be susceptible to reconstructing these perceived demands and behave in ways that confirm them. This phenomenon is currently referred to as either confirmation bias, researcher’s effect, or the “Pygmalion effect”.

This thesis is a critical presentation of the theories and studies developed by some social psychologists. We focus on the relationship established between the researcher and the participant. We analyze the experimental situation through the study of its three components: context, topic and researcher (Lemaine, 1975). We subsequently use three kinds of basic precautions to limit the manner in which this artefact could be integrated into the research protocol. The investigation before/after, and the initial and formal steps method are explained and discussed here. These methods apply to all experimental situations that involve interaction between a researcher and each participant. The various elements are synthesized in an operating model called “EMIR”, which proposes to use meta-experiment in order to study “the researcher’s effect”, not as a bias but as a separate psychological phenomenon.

Next, we present five empirical studies of this phenomenon. The first meta-experimentation relates to the field of “responsibility judgement”. This field focuses on the process of evaluation of the responsibility of a suspicious actor. This research proposes to study the effect that conflicting expectations induce in different experimenters. Some researchers studying the responsibility field oppose collective situation to individual situation. Based on this, we lead delegate experimenters to think that the subjects will be more strict in collective situation than in individual situation (versus the investigated subjects will be more strict in individual situation than in collective situation). After this first study we emphasise the influence that the delegate experimenters’ expectations have upon the participants in the judgement task. Nevertheless, some of our measurements seem to be influenced by factors related to the judgement task (e.g., norm of indulgence linked to the participant population), to method (e.g., personality factors of delegate experimenters), or to the possibility of a “double Rosenthal effect” (i.e., validation of the expectations of the principal experimenter by delegate experimenters).

These limitations lead us to optimize our method by the use of video interface in order to standardize the theoretical induction onto the delegate experimenters, and maintain the principal experimenter blind to induced expectancies. More precisely, in this study, the delegate experimenters expect that one of two portraitures (« T » and « U » pre-tested equivalent) are to be evaluated more positively (« T » or « U ») by the participants. We observe an efficient but implicit influence of the expectations of the delegate experimenters on the participants. « T » and « U » are evaluated by the participants in accordance to the experimenters’ expectations. Nevertheless, the measurement of the vectors of the transmission of the expectations offers the expected results.

Therefore, we replicate this study with a record of the delegate experimenter in the aim to study para-verbal factors (i.e., vocal, proxemic, kinestesic) likely to mediate the transmission of experimenter’s expectations to the participants. But the use of video tape record seems to hinder the replication of the results. The following study proposes to use one profile « T » with two expectations (positive versus negative). Our results are convergent in this replication. Consequently, the last presented study concerns cognitive dissonance and more precisely the more labouriously task.

In this section we oppose two kinds of theoretical presentations of the impact of freedom of declaration on evaluation of attractivity of a task by participants. In the first modality we favour membership of radical theory of cognitive dissonance (French theoretical chapel). The second modality consists of questioning this theory by underlining the difficulty of reproducing classical results of radical theories. We observe an influence of the delegate experimenter’s expectancies on the evaluation of the task by the subjects in the sense of induced expectations.

We conclude this thesis by questioning the psychosocial role of the researcher in the experimental situation. To approach this question we rely on the contributions of the meta-experiment problematics.


contact : ralph @ balez.eu

— Balez, R. (2007, november 27). Web Extension to American Psychological Association Style (WEAPAS) (Rev. 2.0) [http://www.balez.eu/spip/spip.php ?article12]. URL

Follow-up of the site's activity RSS 2.0 | Site Map | Private area | SPIP | template