Géraldine Rouxel

Cognitive–affective determinants of performance in mathematics and verbal domains: Gender differences

Abstract Reviewed in this article are the few previous studies that addressed the application of structural modeling to specify the systemic organization of the relations between self-efficacy, anxiety, gender, and performance. According to an interactionist approach, one can expect rather different systems of relations when people are confronted with different academic areas and when gender is taken into account within the same area. For this study, 476 students in the fourth and fifth grades were asked to take examinations in mathematics and verbal domains and to complete several questionnaires relative to these specific domains. Quantitative, rather than qualitative, differences were observed between genders (i.e., multisample analyses). In accordance with an interactionist approach, there were structural differences between the models tested in the two domains. Finally, the results demonstrated the value of repeated measures of state anxiety to understand the evolution of the relationships with the other variables over time.

Path Analyses of the Relations between Self-Efficacy, Anxiety and Academic Performance.

Differences in academic performance can surely be explained by differences in cognitive aptitudes, but cognitive-affective variables like self-efficacy or anxiety may also intervene. This paper is partly based on Bandura’s social cognitive theory and several assumptions relative to the systemic organisation of the cognitive-affective variables just mentioned with academic performance are tested. 505 students in the 4th and 5th grades were invited to participate in the study. Anxiety and self-efficacy are measured by self-report questionnaires at two levels of generality: dispositional (trait-like variables) and situational (state variables). Performance refers to the scores obtained at an usual French exam. Path analyses replicate the same functional relation between self-efficacy and anxiety. But when one aims at explaining performance, Bandura’s general hypothesis of an effect of self-efficacy on performance directly and indirectly via anxiety doesn’t hold anymore when individual differences in the level of knowledge are taken into account. Interesting enough is the finding of rather different models when the level of knowledge is considered. Indeed, it seems that cognitive affective variables play a more important role to explain performance when the students’ level of French knowledge is low.RésuméOn peut évidemment expliquer les différences de performance scolaire par des différences d’aptitudes cognitives, mais des variables cognitivo-affectives comme l’auto-efficacité ou l’anxiété peuvent également être évoquées. En se basant en partie sur la théorie sociocognitive de Bandura, on testera plusieurs hypothèses relatives à l’organisation systémique des variables mentionnées ci-dessus avec la performance. 505 élèves de CM1-CM2 ont participé à létude. L’anxiété et l’auto-efficacité sont mesurées par des questionnaires d’auto-évaluation à deux niveaux de généralité: dispositionnel et situationnel. Le niveau de performance réfère ici aux résultats obtenus lors d’un banal examen de français. Dans les différentes analyses de parcours effectuées, on réplique la même relation fonctionnelle entre auto-efficacité et anxiété. Cependant, quand on cherche à expliquer le niveau de performance, l’hypothèse générale de Bandura d’un effet direct et indirect (via l’anxiété) de l’auto-efficacité sur la performance, ne tient plus quand on prend en compte les différences individuelles dans le niveau de connaissances en français des élèves. Un autre résultat intéressant renvoie à la découverte de modèles assez différents en fonction du niveau de connaissance. En effet, les variables cognitivo-affectives semblent jouer un rôle plus important dans l’explication de la performance quand le niveau de connaissance des élèves est faible.

Quel outil d'évaluation de l'acceptabilité des nouvelles technologies pour des études francophones?

This paper describes the first step of a larger research concerning the establishment of a new technology acceptance model. Our final purpose aims at creating a French-speaking tool to evaluate intentions to use new systems. We wanted this tool to be constructed on the basis of an existing model, wherein we planed to incorporate an emotional dimension. Referring to the literature, we considered the TAM (Technology Acceptance Model, [3, 4]) as a reliable groundwork. We translated items into French and checked we recovered the factorial structure of TAM. We did this by assessing the intentions to use the digital work environment of the university. 192 students, registered for the first time in the university, constituted our sample. We ran the statistical analysis and then noted that we did not recover the original factorial structure of TAM. This result is surprising at first sight and is discussed in the last part. We conclude this paper by presenting the next steps of our research.

From Acceptability to Acceptance: Does Experience with the Product Influence User Initial Representations?

Adopting a product can be seen as a process that implies some repeated use influencing either positively or negatively judgment over time. Indeed, product judgement is likely to evolve over time and influence the intention to continue using. At yet, little research is available on the evolution of product judgement with use. In this paper, we examined the evolution of User eXperience (UX) and affective-motivational factors between evaluation before and after use. Results showed that interaction with the product significantly influences initial representations with, in our study, more negative judgments after use. Moreover, using Expectation-Disconfirmation Theory, significant influences of initial evaluations (i.e. judgement before use) and disconfirmation (i.e. difference between judgement after and before use) were found on behavioral intention after use.

What links between user experience and acceptability

This publication describes the initial phase of work: selection of the most relevant indicators for measuring the UX and acceptability in order to develop a new assessment tool. Then, the aim will be to study the links between user experience (UX), acceptability and affective-motivational variables using structural equation modeling.

Échanges conversationnels avec un partenaire familier : étude comparative entre enfants sourds, implantés cochléaires depuis 2 ans et enfants entendants

Resume Les enfants sourds affichent tres souvent des difficultes a s’approprier les conventions et regles implicites regissant l’utilisation du langage. De ce fait la pose d’un implant cochleaire devrait permettre a des enfants atteints de surdite profonde de presenter des profils de production langagiere se rapprochant de ceux des enfants entendants du meme âge. Dans cette perspective, 18 enfants implantes depuis 24 mois et 18 enfants entendants ont ete filmes en situation de jeu, avec un partenaire familier. Les indices structurels et pragmatiques des productions indiquent, malgre un lexique moins etendu, une predominance de la modalite orale chez les enfants implantes. La repartition des differents actes de langage apparait comme comparable a celle observee chez les enfants entendants. La quantite de tours de parole reste cependant moindre chez les enfants implantes. En conclusion, malgre les differences observees, le developpement des echanges conversationnels permis par l’implant se rapproche de celui de l’enfant entendant.

Induction and Evaluation of Affects for Facial Motion Capture

In this study, we are interested in capturing the facial configuration of Affectsin order to use them for Embodied Conversational Agents. In order to create a believable eca , it is necessary to capture natural Affectsthat can be learnt and replayed. However, until now, animation data are extracted from videos and their description is far from being sufficient to generate realistic facial expressions. It seems that believable results cannot be obtained without using 3D motion capture. This is why in this study we tried to set up a protocol for Affectsinduction in a motion capture situation with manipulated subjects who are unaware of the real goals. Similarly from [1], we induce natural Affectsin order to capture the related facial expressions.