significant p = .52. The value of Cramer’s V was .05 showing that only 2.5% of the variance in language anxiety can be explained by ought-to/other and actual self discrepancy. Again, there is little evidence to suggest that ought-to/other and actual discrepancy is associated with language anxiety.
These results go against what was expected. Strauman and Higgins (1988) found in their causal model that discrepancy between the ought-to and actual selves is significantly associated with social anxiety (R = .33, p < .01). The analogue of this association in the language learning context did not come to light, most probably due to the lack of instruments that could reliably and in a valid way measure incongruities between learners ought-to and actual L2 selves.
6.5.1 Language Anxiety and Self-efficacy Beliefs
The following variable analyzed relative to language anxiety was learners’
self-efficacy beliefs that is their beliefs of having the ability and tools to perfrom language learning tasks and ultimately successfully learn a foreign language.
Spearman’s rho was calculated to depict the relationship between self-efficacy and language class anxiety. However, in the cases of comparing language use anxiety and the FLCAS scores to self-efficacy beliefs, Pearson’s coefficient was also meaningful as these three variables met the normality criteria of parametric tests (see Table 24 above).
All three indicators of language anxiety were found to have significant moderate correlations with self-efficacy beliefs (rs = -.52, r = -.53, r = -.67) at the p <
.01 level of significance. The strongest relationship was depicted by the correlation coefficient r = -.67 between the summative score of language classroom anxiety and self-efficacy beliefs.
Cheng (2001) in the Taiwanese EFL context also found a moderately strong and significant (r = -.68) relationship between learners’ self-efficacy beliefs and their language anxiety based on scores from an adapted version of the FLCAS and a newly devised instrument to measure language learners’ self-efficacy beliefs.
Mills, Pajares and Herron (2007) measured reading and listening language anxiety and self-efficacy beliefs. The hypothesis that the researchers set out to test was whether “French grade self-efficacy or students’ confidence beliefs in their ability to attain an intermediate French grade will make an independent contribution to the prediction of the French grade of college intermediate-level French students when French learning anxiety, French learning concept, efficacy for
self-regulation, and value of French language and culture are controlled” (p.425). Their study found self-efficacy beliefs of grades to be the best predictor of language learning achievement. The explanation that the authors offer for this result is that learners with stronger self-efficacy have positive expectations of overcoming stressful situations, which generally leads them to success and positive experiences that further enhance their self-efficacy beliefs.
The above findings are in line with Bandura’s (1997) claim that anxiety correlates negatively with efficacy. This was further investigated using data from the present sample by testing whether there is a significant difference between the low and high language anxiety groups with respect to their self-efficacy. The Mann-Whitney U = 559.50 (z = -6.23) results depicted a significant (p < .01) difference between the less anxious and the more anxious groups based on their self-efficacy beliefs. This means that the difference of the two groups along the self-efficacy dimension is not likely to be the consequence of sampling error.
From these results we can infer that those learners who do not feel apt at completing language learning tasks tend to feel more anxious. In contrast, those learners who judge themselves capable of approaching language tasks and succeeding feel less anxious. This is also what Lazarus’ (1993) concept of the cognitive appraisal implies: the feeling of anxiety is reciprocal to the self-efficacy beliefs about whether one has the resources to cope with the language learning task and the accompanying stressors.
6.5.2 Language Anxiety and Learners’ General Beliefs about Language Learning
The data analysis of many studies conducted using BALLI are based on the analysis of frequencies of the responses to each item (for an overview of such BALLI studies see Horwitz, 1999). Following in these footsteps, learners’ answers to each item were correlated with their levels of anxiety (see Table 27) using Spearman’s rho.
Table 27
Significant Correlations between BALLI Items Language Anxiety Using Spearman’s Rho
*p < .05. ** p < .01.
Generally low but significant correlations were found in case of 16 BALLI items and language anxiety. The most noteworthy are items 5, 13, 16, and 21with correlation
FLCAS Language use anxiety
Language class anxiety
balli1 -.28** -.17* -.24**
balli3 -.26** -.20** -.25**
balli4 .40** .27** .28**
balli5 -.59** -.37** -.42**
balli9 .18** .12 .16*
balli11 .23** .22** .24**
balli13 -.43** -.31** -.29**
balli14 -.20** -.05 -.17*
balli15 -.23** -.22** -.16*
balli16 -.48** -.27** -.40**
balli19 .20** .07 .08
balli21 .60** .58** .58**
balli28 .20** .13 .20**
balli31 -.23** -.22** -.18**
balli32 -.29** -.29** -.20**
balli34 .18** .18** .18**
coefficients rs flcas= -.59, rs flcas = -.43, rs flcas = -.48, rs flcas = .60 respectively, which show a statistically significant moderate relationship between learners’ level of anxiety and their beliefs about the difficulty of language learning (item 5), learning and communication strategies (item 13 and item 21), and aptitude (item 16) (based on Horwitz’s (1987, 1988, 1999) theoretical construct of learners’ beliefs about language learning).
Inspecting the above items more closely, it seems that the higher levels of language anxiety learners experience the less likely they are to believe that they will eventually succeed in learning to speak English well (item 5). The more anxious they are the less they enjoy meeting native speakers and communicate with them using English (item 13). Item 21, originally associated with communication apprehension (see Horwitz, 1987) depicted the highest correlations with anxiety. Upon a closer look, this item seems to be more related to learners’ feelings of anxiety rather than communication strategies: “Félénk vagyok, amikor angolul kell beszélnem másokkal”
(“I feel timid speaking English with other people.”). In this view, it is not surprising that this item’s relations to other anxiety measures could be characterized by a stronger correlation coefficient than any other items’.
A notable finding is that five of the items that were found to have significant correlations with anxiety were theorized by Horwitz (1987, 1988) to measure language learning difficulty, and the second largest group of items found to be in a statistically significant relationship with anxiety are the four items grouped under the flag of beliefs about language learning aptitude. In other words, the more anxious learners are, the more difficult they find language learning and the lower their perceived level of aptitude is, or perhaps the reverse is true: the more difficult learners perceive language learning to be and the lower their levels of aptitude is, the higher
their levels of language anxiety. This parallels Sparks and Ganschow’s (1993, 1995) LCDH which states that those learners who encounter language learning difficulties are the ones likely to experience language anxiety. Thus, beliefs about language learning difficulty can be taken as indicative of learners possessing language learning anxiety.
Examining the differences between less anxious and more anxious learners, the Mann-Whitney U (see Table 28) and z statistics were found to be significant at p <
.05 level in the case of BALLI items 1, 11, 16 (referring to beliefs about aptitude);
items 3, 4, 5, 34 (in connection with beliefs of language learning difficulty); items 13, 21 (alluding to beliefs about strategy use); item 28 (beliefs related to the nature of language learning); and item 32 (referring to motivation). Again, it seems that more anxious learners hold significantly different beliefs about difficulty of language learning and about their aptitude to learn a foreign language successfully than less anxious learners.
Table 28
Differences between Less Anxious and More Anxious Learners’ Responses on BALLI Items Using the Mann-Whitney U statistic
balli1 balli3 balli4 balli5 balli11 balli13 balli16 balli21 balli28 balli32 balli34
Mann-Whitney U
195.50 2068.50 1765.00 1256.00 1978.00 1707.50 1536.50 720.00 1994.00 2125.50 2048.00
Z -2.94 -2.75 -3.68 -5.29 -2.98 -3.70 -4.27 -6.87 -2.77 -2.34 -2.57
Note: Grouping variable: anxiety level