The Effects of Anxiety on Self-Perceived Proficiency and Actual Achievement ()
1. Introduction
Foreign language learning is influenced by multiple factors, such as attitudes (Petrides, 2006), emotions, and motivation (Khodadady & Khajavy, 2013). About the various factors that impact the process of learning, emotions are believed to play an essential role. The fundamental role of those factors has been well-documented (Vygotsky, 2000; Swain, 2013). Evidence indicates a strong correlation between emotions and achievement (Pekrun et al., 2006). Specifically, positive emotions have been shown to bolster learning outcomes, whereas negative emotions hinder them (MacIntyre, 1999). Therefore, anxiety is the commonly studied emotion in foreign language (FL) research due to its significant influence on second language (L2) achievement, as highlighted by MacIntyre & Mercer (2014).
Strong anxiety is often perceived as a disruptive influence on behavior, impairing interpersonal interactions, cognitive processes and learning procedure (Gregersen & Horwitz, 2002; MacIntyre & Gardner, 1994). Therefore, the effect of anxiety has been the widely studied topic (e.g. Dewaele, 2007; Horwitz, 2001). Regarding second language anxiety, current research mainly concentrates on two distinct types of assessments: self-perceived proficiency (Shao et al., 2013) and actual achievement (e.g. Botes et al., 2020). Although Wang & Li (2022) discovered that FLCA has a significant negative correlation with both self-perceived and actual achievement, few studies explore whether FLCA has greater predictive power for self-perceived proficiency or actual achievement. To bridge this gap, current study compares influence of anxiety on self-perceived and actual achievement, aiming to ascertain which holds greater predictive power.
2. Literature Review
2.1. Foreign Language Classroom Anxiety
“Achievement emotions” is an omnibus term encompassing a wide array of diverse feelings, such as enjoyment, anxiety, shame, and boredom, which arise surrounding achievement outcomes. About those achievement emotions, a prominent theory is “Control-Value Theory” (CVT, Pekrun, 2006). Within CVT model’s framework, taxonomy involves three distinct factors: valence (positive and negative), activation level (extent of physiological arousal), and the object focus (whether related to activities or outcomes). According to this, anxiety is classified as a negative, high arousal result emotion. Horwitz et al. (1986: p. 128) points that the definition of foreign language anxiety (FLA) is a unique constellation of self-perceptions, beliefs, feelings, and behaviors linked to classroom second language acquisition, stemming from the distinctive characteristics of the language acquisition process. Consistent with CVT, anxiety can be ascribed to a perceived insufficiency in control and negative subjective evaluation of the outcomes. This means that individuals are more prone to experience anxiety when they perceive outcomes as uncertain and highly significant to them.
Notably, the pioneering research conducted by Horwitz et al. (1986) ushered in a new era in the realm of anxiety research. It is the creation of the comprehensive Foreign Language Classroom Anxiety Scale (FLCAS). Comprising three taxonomies—communication apprehension, test anxiety and fear of negative evaluation with 33 items, it has been widely used in published articles. However, the advent of the Short Form FLCAS (S-FLCAS) measurement by Dewaele & MacIntyre (2014) redirected research attention. In order to create a more concise and practical tool, the researchers carefully selected eight items from the original questionnaire to formulate the Short Form Foreign Language Classroom Anxiety (S-FLCA). This abbreviated version maintained the reliability of the measurement, as evidenced by MacIntyre’s (1992) research. This concise scale offers a significant advantage by streamlining administrative time, thereby enabling researchers to incorporate a wider array of assessments in their studies (Heene et al., 2014). Consequently, numerous recent research has favoured this version (e.g., Dewaele et al., 2019). Moreover, the validity of this version for future investigation has already been confirmed by Botes (2022).
2.2. Self-Perceived Proficiency and Actual Achievement
Data can broadly be classified into two categories: attitudinal and factual (Tourangeau et al., 2000). Verification of factual data can be accomplished by cross-referencing with external sources, such as institutional records. These data sources usually range from Grade Point Average (GPA) (Cabrera et al., 1990) to The College English Test 4 (CET-4). On the other hand, attitudinal data are collected based on individuals’ self-reported beliefs or opinions, which, by their nature, are generally not amenable to external verification. This type of data can be identified as Self-perceived communicative competence (Self-perceived CC). This concept concerns an individual’s self-assessment of their communicative competence (McCroskey & McCroskey, 1988) including interactions with diverse interlocutors across a broad spectrum of contexts and situations (Canale, 1983). Self-perceived CC is typically assessed utilizing 5-point Likert scales, wherein participants are prompted to assess their perceived proficiency in speaking, comprehending, reading, and writing abilities (Dewaele, 2010). The questions are in the format: “Please evaluate your skill level in speaking, listening, reading, and writing English”.
Anaya (1999) observed a notably strong correlation between individuals’ self-perceived achievements and their actual accomplishments. But there were conflicting opinions on whether self-perceived approach can be used to accurately assess objective achievement. Some researchers (e.g., MacIntyre et al., 1997) challenged the accuracy of self-perceived language proficiency and found the inconsistency underlying in this assessment. Other studies (Cassady, 2001; Cole & Gonyea, 2010) discovered high correlation and consistent results between self-reporting measurement and objective ability assessment. Despite conflicting views, those two assessments are prevalent approaches (Phinney et al., 2001), which deserves to be researched further.
However, there is limited research comparing individuals’ self-perceived proficiency and actual academic achievement levels. This type of research is crucial because it not only enriches the theoretical framework of educational psychology, offering more scientifically grounded guidance for educational practices. It also further helps to explore the underlying causes and mechanisms of anxiety’s impact on self-perceived proficiency and actual achievement.
2.3. The Relationship between FLCA, Self-Perceived Proficiency and Actual Achievement
In terms of actual achievement, numerous studies consistently demonstrate an inverse relationship between FLCA and actual academic achievement levels (Horwitz, 2010). Regarding self-perceived ratings, research conducted by Dewaele & Ip. (2013) showed that FLCA served as a negative factor in predicting self-perceived English proficiency. Although both self-perceived proficiency and actual achievement serve as measures of academic achievement, they still yield divergent assessment outcomes. However, the majority of research has focused on either self-perceived or actual performance measurement. There remains a scarcity of studies exploring the varying degrees to which self-perceived proficiency and objective achievement reflect foreign language anxiety. To bridge the gap, the aim of the study is to further verify which assessment—self-perceived performance or actual achievement—is more greatly influenced by FLCA.
3. Methodology
3.1. Research Questions
1) What is the status of FLCA among Chinese EFL learners?
2) How do Chinese EFL learners’ FLCA, actual achievement and their self-perceived proficiency correlate?
3) How does FLCA predict achievement and self-perceived proficiency independently?
3.2. Instruments
Two established scales were utilized: the Foreign Language Classroom Anxiety Scale (FLCAS) and four self-assessment items derived from the scale developed by Gee et al. (2010). The primary instrument employed in this research was the Short Form-FLCA (S-FLCAS), a well-recognized and reliable measure constructed by Dewaele & MacIntyre (2014). This scale is designed to assess classroom anxiety among EFL learners consistently. The scale has been regularly utilized since it was developed (e.g., Bensalem, 2021). It includes 8 items on 5-point Likert scales, with the responding options ranging from “strongly disagree (1)” to “strongly agree (5).” Mean scores between 4 and 5 are indicative of high levels of anxiety, whereas mean scores ranging from 3 to 4 suggest a moderate level of anxiety. Conversely, mean scores falling within the range of 1 to 3 imply minimal anxiety. The S-FLCAS has demonstrated a strong level of internal consistency (Cronbach’s alpha > 0.8) in many studies (Dewaele & MacIntyre, 2014; Botes et al., 2022). The questionnaire’s reliability in the current study was 0.855. Several studies (e.g., Fathi & Mohammaddokht, 2021) used CFA and fit indices to approve the validity of the S-FLCAS questionnaire. The fit indices GFI (0.97), CFI (0.98), TLI (0.98) and RMSEA (0.04) showed that this model was acceptable fit. The items 4 and 5 are formulated negatively with coded-reversed responses.
Self-perceived proficiency was assessed with 4 self-evaluation items with responses ranging from “far below average (1)” to “far above average (5)”. Four items assessed self-perceived proficiency: “Please evaluate your skill level in speaking, listening, reading, and writing English.” Alternatively, the actual achievement was evaluated using their scores on the CET-4, a standardized English test conducted by the Ministry of Education of China. Prior research (Yang & Weir, 1998) has validated the reliability and validity of CET-4 scores, establishing their suitability for assessing Chinese participants’ actual English achievement.
3.3. Procedure and Data Analysis
A total of 214 undergraduate students from China (24 males and 190 females) were selected based on convenience sampling procedure. The gender imbalance in the data was not surprising, as it has frequently been observed in previous FL learning research (e.g. Dewaele et al., 2023). These participants were from a diverse range of majors, including Engineering, Pedagogy, Management, and Economics. The full version of FLCAS was developed by Horwitz et al. (1986). Dewaele & MacIntyre (2014) subsequently selected 8 items to form a shortened version (S-FLCAS), which was employed in this study. The research objectives were explained clearly to the participants, who were also informed that their names and answers would be confidential, thus ensuring adherence to the principles of anonymity and privacy protection in accordance with academic standards.
The data analysis of the questionnaires was conducted using SPSS 26.0. Three research questions are addressed. The first question is examined through descriptive analysis. To solve the second question, we carried out Pearson correlation analyses to investigate the relations between the variables. Regarding the last question, regression models were applied to investigate the FLCA and self-perceived proficiency, as well as FLCA and actual achievement so as to determine the different amount of variation explained by these variables.
4. Results
4.1. Levels of FLCA
To provide a comprehensive overview of students’ status regarding FLCA, descriptive statistical measures are employed. In Table 1, the mean score of FLCAS was 3.184, which is slightly above the median anxiety level of 3, suggesting that participants’ anxiety levels in the English class are moderate.
4.2. Correlations FLCA with English Achievement and
Self-Perceived Proficiency
The study carried out a correlation analysis to ascertain the associations between
Table 1. Descriptive statistics.
|
N |
MIN |
MAX |
M |
SD |
FLCA |
214 |
1 |
4.88 |
3.18 |
0.67 |
Self-perceived Proficiency |
214 |
1 |
4.5 |
2.74 |
0.6 |
Actual Achievement |
214 |
298 |
544 |
395.71 |
49.32 |
FLCA, actual achievement levels and self-perceived proficiency. As depicted in Table 2, a slight inverse association was identified between students’ FLCA scores and their actual achievement levels, with a correlation coefficient (r = −0.20, p < 0.01), indicating a modest inverse relationship. This indicates that individuals who have higher FLCA scores tended to exhibit lower scores in actual achievement. Similarly, students’ self-perceived proficiency was significantly and negatively correlated to FLCA. (r = −0.42, p < 0.01). Alternatively, the higher the level of anxiety, the lower the self-perceived proficiency is. After a comparative analysis, it was discovered that the Pearson correlation coefficient between FLCA and self-perceived proficiency is significantly higher than that between FLCA and actual achievement. This indicates a much closer correlation between FLCA and self-perceived proficiency. Furthermore, participants’ self-perceived proficiency demonstrated a moderate positive relationship with their achievement scores (r = −0.27, p < 0.01). In other words, individuals with a favorable self perception of language proficiency attained higher scores on actual assessments.
Table 2. Correlations among FLA, achievement, and self-perceived proficiency.
|
1 |
2 |
3 |
1) FLCA |
- |
|
|
2) Self-perceived proficiency |
−0.42** |
- |
|
3) Actual achievement |
−0.20** |
0.27** |
- |
Note. ** p < 0.01.
4.3. Regression FLCA with English Achievement and Self-Perceived Proficiency
Two regression models were established to further verify the relationship between FLCA, actual achievement and self-perceived proficiency. Table 3 & Table 4 show the results from multiple regression. None of the regression models developed in this study exhibited the issue of multicollinearity (all VIF < 3). The results indicate that FLCA is a negative predictor for both students’ CET scores and self-perceived proficiency. From comparative analysis of β coefficients, FLCA exhibits a stronger predictive power for self-perceived proficiency (β = −0.42, p < 0.01) than for CET-4 achievement (β = −0.20, p < 0.01). This indicates that anxiety has a more significant predictive power on self-perceived achievement compared to actual achievement.
5. Discussion
The present study reveals that the participants’ anxiety levels regarding English
Table 3. Predicting actual achievement.
|
Unstandardized coefficients |
Standardized coefficients |
|
Collinearity Statistics |
|
B |
Std. error |
β |
t |
Sig. |
Tolerate |
VIF |
FLCA |
−14.76 |
4.99 |
−0.2 |
−2.96 |
0.00 |
1 |
1 |
Dependent variable: actual achievement.
Table 4. Predicting self-perceived proficiency.
|
Unstandardized coefficients |
Standardized coefficients |
Collinearity Statistics |
|
B |
Std. error |
β |
t |
Sig. |
Tolerate |
VIF |
FLCA |
−0.38 |
0.05 |
−0.42 |
−6.78 |
0.00 |
1 |
1 |
Dependent variable: self-perceived proficiency.
classes are moderate, slightly above the median level (M = 3.18). These results are consistent with numerous other studies. A similar slightly above median level of FLCA (M = 3.19) in Chinese context was found in Wang & Li (2022)’s study. Besides, nearly half of the participants exhibit moderate levels of anxiety in Libya (Toubot et al., 2018). It means that the Chinese participants’ scores slightly surpass those of their Libyan peers. It may be due to the hypothesis that a positive relationship exists between linguistic distance and anxiety in language learning (Tao et al., 2024). This means that the notable differences in syntax, vocabulary and phonology between Chinese and English pose unique challenges for Chinese learners rather than Libyan students, often leading to increased anxiety levels during the language learning process.
The results of the current research indicate a modest inverse relationship (r = −0.2, p < 0.01) between FLCA and academic achievement. Specifically, participants with higher FLCA scores tended to achieve lower grades. This is consistent with previous research. For example, a meta-analysis was conducted by Teimouri et al. (2019) on anxiety and actual achievement, finding an overall small effect size of r = −0.36 (k = 105; N = 19,933) from 23 countries. Besides, according to Wang et al. (2023)’s study, FLCA was also found to have negative relationship with English achievement, with a small effect size (r = −0.210, p < 0.01) in China. One plausible explanation is that anxious students may be less inclined to communicate as much information as their less anxious counterparts, which can affect the quality of their output (MacIntyre et al., 1997).
Similarly, a medium negative correction (r = −0.42, p < 0.01) was also observed between students’ FLCA and their English self-perceived proficiency, echoing findings from earlier studies. For example, Li (2020) observed a medium to high level of correlation (r = −0.51, p < 0.001) between individuals’ self-perceived proficiency and their levels of anxiety within China. Besides, a medium inverse correction (r = −0.4, p < 0.001) was also found in Arab EFL learners (Yang, 2021). It is posited that learners experiencing affective states exhibit greater confidence and a higher self-assessment of their potential competencies in English language learning, whereas those frequently grappling with negative emotions tend to negatively appraise their academic prospects (Wang & Li, 2022). A plausible explanation is that learners who have higher levels of anxiety may view themselves as less capable, while those people with lower anxiety may demonstrate greater confidence in language proficiency. Additionally, the study delves further to uncover a positive moderate correlation between participants’ achievement and their self-perceived proficiency. This discovery aligns with previous research (e.g. Anaya, 1999) suggesting that individuals with a favorable self-perception of their abilities tend to attain higher scores in terms of actual achievement.
This study corroborates the Control-Value Theory, which indicates that emotions pose a significant influence on self-perceived language proficiency and actual achievement. The finding that self-perceived proficiency is a predictor of anxiety aligns with prior research (e.g. Arnaiz & Guillén, 2012; Sparks & Ganschow, 2007). For instance, FLCA was the slightly medium negative predictor of performance in English (β = −0.44, t = −5.3) in Chinese context (Dewaele et al., 2013). The results also indicate that although participants’ FLCA negatively predicted both self-perceived and subjective achievement, their predictive power varied significantly, with self-perceived language proficiency being notably more influenced. The current study corroborates the findings reported by Li (2020), indicating that the relationship’s effect size between anxiety and the self-perceived language proficiency exceeds that with actual achievement. This outcome aligns with prior research. Furthermore, self-perceived proficiency serves as a more potent predictor of anxiety compared to actual achievement measures (Cheng et al., 1999).
The reasons underlying this phenomenon can be elucidated from the following three aspects. To begin with, it is evident that learners’ self-perceived proficiency is more easily affected by emotions, particularly, the negative high arousal emotion—anxiety. For instance, language learners with high anxiety may concentrate more on their perceived inadequacies and risks associated with potential failure. This was further interpreted by MacIntyre et al. (1997). They discovered that students with high levels of anxiety prone to underestimate their ability levels and for more calm students to overestimate theirs. Secondly, Eysenck et al. (2007) argued that anxiety may not invariably impede the quality of performance. For example, if moderate level of anxiety triggers the adoption of compensatory measures, such as, heightened effort and intensified utilization of cognitive resources, the quality of performance does not deteriorate; instead, it is enhanced. This further explains why anxiety sometimes exhibit less significant influence on actual achievement. The third reason may involve combined effect of indirect minor factors, for example, self-esteem and impression management and other social desirability factors. In the study conducted by Nancarrow & Brace’s (2000), participants who are subject to social desirability pressure tend to engage in impression management or ego-defensive behaviors thereby purposely distorting their self-perceived proficiency. In addition, the language distance between target language and source language intensifies students’ FLCA. Simultaneously, the widespread emphasis on English proficiency in China strengthens learners’ desire to acquire social recognition, thus forming a vicious cycle, with many external forces exacerbating the impact of FLCA on self-perceived proficiency.
6. Conclusion
Overall, this study yields two key findings. Firstly, Chinese EFL learners’ Foreign Language Anxiety (FLA), self-perceived proficiency and academic achievement exhibit inter-correlations. Specifically, FLCA negatively correlates with both self-perceived proficiency and actual achievement, while there exists a positive correlation between individuals’ self-perceived proficiency and their achievements. Second, Chinese EFL learners’ FLCA serves as a negative predictor for both self-perceived and objective performance, with self-perceived one being significantly predicted and influenced. The series of findings can be ascribed to a multiple interacting psychological and cognitive factors that collectively co-determine the complex relations among FLA, self-perceived proficiency and actual achievement.
Important implications can be drawn from these findings. Pedagogically, teachers can intensify students’ positive bias, such as self-enhancement, to facilitate language learning to some extent (MacIntyre & Charos, 1996), provided that this positive image do not distress students. This underscores the importance of incorporating anxiety-reduction strategies into FL pedagogy, as emphasized in Reid (2009). In addition, concerning specific classroom interventions, for example, offering systematic scaffolded feedback, is conducive. Specifically, indirect and positive feedback language should be favored rather than direct humiliating feedback. Moreover, more detailed dimensions in explaining mistakes are proved to increase self-confidence in oral performance (Öztürk & Öztürk, 2021). Besides, taking individual differences into consideration is also crucial, so flexibility in anxiety reduction strategies should also be developed (Pashler et al., 2008). For example, incorporating visual aids and collaborative activities could ease communication anxiety in visual or auditory learners. Generally, integrating relaxation techniques, mindfulness exercises, and positive reinforcement throughout the curriculum can reduce all students’ FLCA (Firissa & Gebremariam, 2024).
From a research perspective, further research should also delve into the mechanisms through which FLCA influences self-perceived and objective performance differently. With a deeper understanding of these mechanisms, interventions would be more targeted to address the specific needs of students. Additionally, this study can lay a foundation for future longitudinal studies to examine the long-term effects of FLCA on language learning. Such studies can help determine whether the negative impact of FLCA persists over time and how it evolves in response to various interventions.