e-Tandem jitters: a study of online learners’ foreign language anxiety

Due to the increasing use of technology to enhance Foreign Language (FL) education, research on learners’ emotions in new learning environments is calling for more attention (Beirne, Mac Lochlainn, Nic Giolla, & Mhichíl, 2018). In this study, we focus on Foreign Language Anxiety (FLA), a debilitating emotion; and e-Tandem learning, a telecollaborative Foreign Language Learning (FLL) practice. e-Tandem has a vast potential to foster learners’ FL skills (Cziko, 2013), although it might trigger learners’ FLA as well. Since little research has been carried out, hitherto, on FLA in e-tandem learners, this investigation aims to gain new insights into this topic. First, we want to analyze the appropriateness of the Foreign Language Classroom Anxiety Scale (FLCAS) ( Horwitz, Horwitz, & Cope, 1986) to assess e-tandem learners’ FLA. Second, we want to observe to what extent e-tandem contributes to reduce learners’ FLA over time. Descriptive statistics are carried out and results are discussed.


Introduction
FLA is described as "a distinct complex of self-perceptions, beliefs, feelings, and behaviors related to classroom language learning arising from the uniqueness of the language learning process" (Horwitz et al., 1986;p. 128). It is a negative and debilitating emotion (Gregersen, MacIntyre, & Meza, 2014), considered both a threat (Dewaele, 2011) and a predictor for FL achievement (Onwuegbuzie, Bailey, & Daley, 1999 Due to the increasing use of technology to enhance FL education, language learning scenarios are changing, although learners' emotions remain the same. In this study, we focus on learners' FLA when taking part in e-tandem, a technology-based FLL practice. Within e-tandem, two learners who are physically distant and who have a different native language communicate in order to learn each other's language (Brammerts, 2001).
e-Tandem contributes, among others, to the improvement of learners' FL skills (Cziko, 2013) and self-confidence when performing in the FL (Appel, 2012). Nevertheless, e-tandem might potentially trigger learners' FLA (El-Hariri, 2017). Despite the fact that FLA has been widely investigated (MacIntyre & Gregersen, 2012), little research has been conducted, hitherto, on this emotion in e-tandem contexts.
In this work, we aim to analyze the applicability of the FLCAS (Horwitz et al., 1986), a broadly used instrument to assess FLA in e-tandem contexts. Moreover, we want to observe to what extent this practice helps to reduce learners' FLA over time. Our Research Questions (RQs) are: • RQ1: Does the FLCAS predict learners' FLA in e-tandem contexts?
• RQ2: Does practice in e-tandem help to reduce learners' FLA?

Participants
In total, 101 adult FL learners enrolled in an online FL course based on e-tandem activities form the purposive sample of the study. Participants are Native Speakers (NSs) or proficient speakers of English who are learning Spanish as a FL (SFL), or NSs or proficient speakers of Spanish who are learning English as a FL (EFL). They have an intermediate or upper level of the FL learned and have completed, at least, one e-tandem activity within the course in pairs. Partnerships could be assigned randomly by the system, or arranged by the learners themselves.

Research context
The research context is TandemMOOC, a free and open English-Spanish FL speaking course organized by the Universitat Oberta de Catalunya in Spain. The course has a duration of six weeks and is based on e-tandem activities, designed to be carried out via videoconference. Each activity counts with an introductory task to be completed in both English and Spanish; one English task and one Spanish task.

Research instruments
The employed research instruments are: (1) a socio-demographic questionnaire, Before enrolling into the course, learners answer the socio-demographic questionnaire. Afterwards, they voluntarily complete the short version of the FLCAS, which allows us to classify them as High Anxiety Participants (HAPs) and Low Anxiety Participants (LAPs). Finally, they self-rate their FLA through the Anxometer which appears automatically on their screen at the end of each task.

Analysis
In the first place, data was anonymized. Afterwards, according to the FLCAS scores, participants were classified as HAPs and LAPs. Subsequently, correlation analyses were carried out in order to study the link between participants' FLCAS scores and the Anxometer mean scores (RQ1), as well as the link between the number of tandems done by each subject and the Anxometer mean scores (RQ2).

RQ1
According to participants' FLCAS scores, the sample counts with 54 HAPs and 47 LAPs. Analysis Of Variance (ANOVA) showed no significant gender (F(1.99)=.435, p=.511) nor age (F(5.96)=1.42, p=.223) differences in this classification. Concerning the correlation between the FLCAS scores and the Anxometer mean scores, the average Anxometer mean score was calculated for HAPs and LAPs. The average Anxometer mean score is slightly lower for LAPs (see Table 1). Finally, correlation analysis revealed no significant correlation between these variables (see Table 2).

RQ2
The average number of e-tandem activities carried out per participant was 5.53 (SD=6.29). The minimum amount of e-tandem activities per participant was 1 and the maximum amount was 35. The average number of e-tandem activities per participant was similar for HAPs and LAPs, being slightly higher for LAPs (see Table 3). Concerning the link between the amount of e-tandem activities completed by participant and the Anxometer mean scores, significant negative correlation was found between these two variables (see Table 4).

Conclusion
First, results show no correlation between the FLCAS scores and the Anxometer mean scores, meaning that participants' prior classification as HAPs and LAPs did not predict the FLA they experienced while taking part in e-tandem activities. Hence, although the FLCAS is a widely used research instrument with acknowledged reliability, its applicability in certain online learning contexts seems to be limited. Recent studies have developed new research instruments for telecollaborative contexts (e.g. Chametzky, 2019;Fondo & Jacobetty, 2020), although their usage is still very narrow.
Second, significant negative correlation was found between the number of e-tandem activities carried out per participant and the Anxometer mean scores. This result suggests that practice in e-tandem helps to reduce FLA over time. Indeed, previous studies reported similar findings (e.g. El-Hariri, 2017;Melchor-Couto, 2017).
Although the results of this investigation allowed us to answer our RQs, further research with larger and less heterogeneous samples would provide us with deeper knowledge on this topic.