Search

To correctly interpret the results of a randomised controlled trial (RCT), practitioners have to spot bias and other potential problems present in the trial. Internal as well as external validity of the trial are linked to the presence of such bias. The internal validity is ensured by a clear definition of the objectives of the trial. The number of patients to be included in the trial is calculated on the basis of the main objective of the trial and more precisely on the basis of the primary endpoint selected to assess the efficacy of treatment. This is the best way to ensure that the statistical significance of the result may have a clinical relevance. Internal validity depends also on the process of patients selection, the methods used to ensure comparability of groups and treatments, the criteria employed to assess efficacy, and the methods for the analysis of data. External validity refers to subjects that have been excluded from the trial, limitations of RCTs, as well as the coherence and clinical relevance of the trial. Internal validity has to be fueled by external validity.

International audience; Many studies examining the effect of information on consumer food product evaluation use a pre-test/post-test quasi-experimental design. Generally, one group of subjects is placed in successive tasting situations. However, sensory peerceptions are not stable overtime, which may lead to internal validity and statistical power problems. A review of the literature develops the advantages of introducing of control groups for the detection of the effects of the information on consumers responses. An empirical study measuring the effect of the region of origin image on the perceived quality of olive oil illustrates the use of control groups and exposes the various statistical means that can used to limit the effect of external uncontrolled variables.; La plupart des études examinant l’effet de l’information sur l’évaluation d’un produit alimentaire chez les consommateurs utilisent un plan quasi-expérimental de type pré-test/post-test avec un seul groupe. Or l’évaluation par les sujets de leurs perceptions sensorielles est instable dans le temps, ce qui peut poser problème au niveau de la validité interne et de la puissance statistique des résultats expérimentaux. Le recours à un groupe témoin peut pallier cette insuffisance. Une revue de la littérature est proposée afin de mettre en avant l’importance des groupes témoins dans la détection de l’effet de l’information. Une étude empirique mesurant l’effet de l’image de la région d’origine sur la qualité perçue de l’huile d’olive illustre l’utilisation de groupes témoins et expose les différents moyens statistiques susceptibles d’éliminer l’effet des variables externes.