Analysis of the main variables

I used descriptive statistics and various statistical analyses to examine the relationship between the variables and the Finnish FLCAS scores. The variables included age, sex, level of degree (bachelor’s or master’s degree), starting year of studies, major subject, language(s) as a minor, first language (or languages, if bi- or multilingual), the experience of living in an English-speaking country, foreign language competence(s), level of English language competence, and anxious personality. The statistical analyses included the Pearson correlation coefficient, independent samples t-test, analysis of variance (ANOVA), and chi-squares. The Pearson correlation coefficient or Pearson’s r “measures the linearity of the relationship of two variables” (Lavrakas, 2008: 155).

I used this correlation test to measure, for example, the correlation between the participants anxious personality self-evaluations and their FLCAS scores. According to Lavrakas (2008: 911) the t-test

“assess[es] the probability that a particular characteristic (the mean) of two populations is different”. I used independent samples t-tests to detect whether variables, such as major, resulted

to significant differences in the FLCAS score means between the two learner groups. Similarly as the t-test, the ANOVA “compare[s] groups on possible differences in the average (mean) of a quantitative (interval or ratio, continuous) measure” (Lavrakas, 2008: 26). I used one-way ANOVA (instead of the t-test) to compare the possible differences caused by language competence(s), because this analysis included more than two comparable groups. Moreover, the chi-square “is a test of significance for categorical variables” which ”let[s] the researcher know what the probability is that a given sample estimate actually mirrors the entire population” (Lavrakas, 2008: 95). I used this test, for instance, to examine the differences between the Finnish FLCAS responses of the English and non-English majors. These analyses suggest that English language as a major and the self-evaluated English language competence level significantly affected the learners’ FLA levels.

The other variables, age, sex, level of degree, starting year of university studies, language(s) as a minor, first language(s), experience of living in an English-speaking country, foreign language competence(s), and anxious personality, were found not to significantly affect degree of FLA.

Figure 3. Correlation between age and FLCAS scores

33

Figure 3 above shows the correlation between the average FLCAS score and age of each participant. The average age of the 125 participants was 25; the youngest participants being 19 and the oldest participant 47. When looking at the English majors and non-English majors separately, the youngest students in both groups were 19. The oldest English major participant was 47, and the oldest non-English major participants were 34. Correlation between age and the FLCAS score means was calculated using Pearson’s r. This test was performed separately with the group of English majors and then with the group of non-English majors. As Figure 3 indicates, the results showed weak negative and weak positive correlation (English majors r=-0.15, r²=0.02; non-English majors r=0.02, r²=0.01) between age and the FLCAS scores. The value of r² indicates that 2% of the English majors followed the negative correlation and 1% of the non-English majors followed the positive correlation. To conclude, age did not correlate significantly with the FLCAS scores.

Figure 4. Sex and FLCAS scores

Figure 4 presents the average FLCAS scores of English and non-English major men and women.

In Figure 4, and in similar figures that follow, the colored bars represent the average FLCAS scores, and the error bars represent the standard deviation. A vast majority, 71% (N=89), reported being

75 79

female and 26% (N=32) reported being male. 2% (N=2) percent of the participants reported being non-binary and the sex of the remaining 2% (N=2) is unknown. The latter two groups were not included in the analysis due to the small comparison group size. Independent samples t-test for the average FLCAS scores of men and women were performed separately on the English majors and on the non-English majors. The results suggest that sex was not a determining variable with the studied groups of English learners with regards to FLA. To elaborate, the results of these analyzes indicated that the p values (English majors p=0.31; non-English majors p=0.60) were higher than alpha (alpha=0.05) and thus sex did not significantly affect the FLCAS scores.

Figure 5. Level of degree and FLCAS scores

Figure 5 illustrates the average FLCAS scores and the standard deviation of the English and non-English major bachelor’s and master’s degree students. Most of the participants, 72% (N=90), were completing a bachelor’s degree, in other words, were in an earlier phase of their university studies, whereas a minority, 28% (N=35), was already in the master’s degree phase of their studies. 68%

(N=50) of the English majors and 77% (N=40) of the non-English majors were bachelor’s degree students, whereas 32% (N=23) of the English majors and 23% (N=12) of the non-English majors

79 78

were master’s degree students. Significance of the level of degree to FLA was examined by independent samples t-tests performed separately on the average FLCAS scores of the afore mentioned groups. These t-tests indicated that the values of p (English majors p=0.97; non-English majors p=0.65) were higher than alpha 0.05, and therefore the results suggested that the level of degree did not significantly affect the participants’ FLCAS scores.

Figure 6. Correlation between starting year of university studies and FLCAS scores

33 66 99 132 165

FLCAS scores

Starting year of university studies Non-English majors

33 66 99 132 165

FLCAS scores

Starting year of university studies English majors

Figure 6 presents the correlation between the starting year of university studies and the average FLCAS scores. The earliest starting year of university studies within the group of English majors was 2009 and within the group of non-English majors 2004. The latest starting year within both groups was 2019. The correlation between starting year of university and the average FLCAS scores was examined using Pearson’s r. As is seen in Figure 6, the analysis showed a weak positive correlation with the starting year of studies and level of FLA within both of the groups (English majors r=0.15, r²=0.02; non-English majors r=0.13, r²=0.02). This indicates that the starting year of studies and the FLCAS scores had a positive correlation for only 2% of the participants in both of the groups, and thus no significant correlation was found.

Figure 7. Language minor and FLCAS scores

The participants were asked to disclose their minor subjects in the background section of the online questionnaire (see Appendix 3). Those with one or more languages as their minor were taken into a closer inspection and compared to those who did not have a language as their minor. The aim was to examine whether a broader experience in language studies would affect the FLA levels of the studied groups of English learners. Figure 7 shows the average FLCAS scores and the standard deviation of participants with and without a language minor. Out of all of the 125 participants 44%

77 80

No language minor Language as a minor

FLCAS score

No language minor Language as a minor

FLCAS score

FLCAS score average and standard deviation Non-English majors

(N=55) had a language as their minor subject. Of the 73 English majors 55% (N=40) had a language as their minor, and of the 52 non-English majors 29% (N=15) had a language as their minor.

Independent samples t-tests comparing the FLCAS scores of those with one or more language(s) as a minor to those without a language minor were conducted separately with the English majors and with the English majors, and in both cases the values of p (English majors p=0.63; non-English majors p=0.14) were higher than alpha 0.05. This indicates that the presence or absence of a language minor did not significantly affect the FLA of the studied group of learners. However, as Figure 7 indicates, the two non-English major groups differ more from each other than the two English major groups, which suggests that there is a greater chance that language as a minor affects the level of FLA with the group of non-English majors than there is with the group of English majors.

As mentioned earlier, all 125 selected participants had Finnish as their first (or native) language.

Of the 125 participants four were bilingual i.e. had two first languages: two of the participants had Finnish and Swedish as their first languages, and the other two had Finnish and Russian as their first languages. Due to the small size of the comparison group (N=4) I did not analyze the relationship between the first languages and the FLCAS scores for this would have not produced generalizable results.

Figure 8. Experience of staying or living in an English-speaking country and FLCAS scores

Figure 8 presents the average FLCAS scores and the standard deviation of English and non-English major participants who had spent less than six months in an English-speaking country and more than six months in an English-speaking country Altogether, 39% (N=49) had stayed or lived in an English-speaking country. These English-speaking countries included Australia, Canada, Ireland, the UK, and the USA. Of these 49 participants, 11% (N=14) had lived more than 6 months in that English-speaking country. Half of these (N=7) were English majors and the other half (N=7) non-English majors. The average FLCAS scores of these 14 participants were compared to the FLCAS scores of those who had spent less than 6 months in an English-speaking country using independent samples t-tests. The t-tests performed separately on the English majors and non-English majors indicated that those who had lived more than 6 months in an English-speaking country did not score significantly lower or higher on the FLCAS than those who had spent less than 6 months in an English-speaking country. To elaborate, the p-values of the t-tests were higher than alpha 0.05 (English majors p=0.60; non-English majors p=0.34). Thus, the experience of staying or living in an English-speaking country was not a determining variable with regards to the levels of FLA.

79 74

The participants were also asked to disclose all foreign languages in which they had at least a B1 competence level according to the CEFR levels (A1 being the lowest and C2 being the highest) (Common European Framework of Reference for Languages. Global scale - Table 1 (CEFR 3.3):

Common Reference levels: 2020). The instructions in the questionnaire implied that the participants should also mention Swedish language here (even though, it is seldom studied as a foreign language in Finland, because Swedish is the second official language in Finland) if they had, for instance, studied Swedish in school. In other words, Swedish was categorized as a foreign language in the following analyses. The participants were divided into three groups based on the number of foreign languages they were competent in. These groups were: competence in one foreign language, competence in two foreign languages, and competence in three or more foreign languages. In total, 23% (N=29) reported that they were competent in one foreign language, of which 14 were English majors and 15 non-English majors. 45% (N=56), reported that they were competent in two foreign languages, of which 33 were English majors and 23 non-English majors, and 32% (N=40) reported that they were competent in three or more foreign languages, of which 26 were English majors and 14 were non-English majors. I conducted one-way ANOVA, separately on all participants, the English majors, and the non-English majors, to uncover whether the FLCAS scores of participants with different language competencies differed significantly from each other. The results indicated that the FLCAS scores between participants with different language competencies did not significantly differ from each other, as in all of the analyzes the p-values were higher than alpha 0.05 (all participants p=0.09; English majors p=0.13; non-English majors p=0.46). Therefore, we can conclude that the number of foreign languages did not significantly affect the level of FLA with the studied group of English learners.

In the background section of the questionnaire, the participants were also asked to evaluate their own English language competence level based on the CEFR levels (see section 2.2 in Appendix 1). The complete distribution of the English competence level self-evaluations among the English and the non-English major participants is presented in Table 1 in the following page. In Table 1, the column titled “CEFR” refers to the CEFR levels which range from C2 to A1 (Common European Framework of Reference for Languages. Global scale - Table 1 (CEFR 3.3): Common Reference levels: 2020). Column “N” refers to the number of participants, and column titled

“percentage” refers to the percentage of the number of participants from the total number of either English majors (N=73) or non-English majors (N=52). The values in the columns titled “FLCAS M” and “FLCAS SD” indicate the mean and standard deviation of the FLCAS scores in each group.

As can be seen from Table 1, a vast majority of the English majors, 78% (N=57), evaluated that their English competence level was at least C2, whereas 31% (N=16) of the non-English majors evaluated that they reached at least a C2 competence level. Therefore, in total 73 participants (58%

of the total 125 participants) evaluated that according to the CEFR they were “proficient users” in English language. The lowest self-evaluation among the English majors was level B2, which refers to “independent user” on the CEFR. However, only one English major had evaluated that their English skill level was at level B2. Two non-English majors evaluated that their English skill level was only at level A2, which refers to “basic user” on the CEFR. Altogether, the non-English majors, naturally, evaluated their own English language skill level much lower than the English majors.

Also, the non-English majors were more divided in their self-evaluations than the English majors were. These differences in self-evaluations are very likely a result of the fact that the English language aptitude is higher amongst those who study English language at a higher level and therefore their self-perceived English language competence is also higher.

Table 1. English language competence level self-evaluations according to the CEFR levels

English majors (N=73) Non-English majors (N=52)

CEFR N Percentage FLCAS

To explore whether or not there was a connection between the self-evaluated English language competence level and the FLCAS scores, I divided the participants into groups based on their CEFR level self-evaluations: As shown Table 1, the English majors were divided into three groups, C2, C1, and B2, while the non-English majors were divided into five groups, C2, C1, B2, B1, and A2.

I then performed one-way ANOVA separately with the FLCAS scores of the English majors and the non-English majors. The results suggested that the differences between the FLCAS scores of participants in different competence levels were in fact significant: In both instances the p values were <0.01, which was lower than alpha 0.05. This indicated that participants who evaluated their English skill higher on the CEFR scored lower on the FLCAS. In other words, English learners with higher self-evaluated English language competence experienced less FLA than those with lower self-evaluated English language competence.

To examine if the variable of anxious personality correlated with the degree of FLA, the participants were asked to evaluate whether or not they felt that nervousness and anxiousness was characteristic to their personality on a 5-point Likert-scale (see section 2.3 in Appendix 1).

Responses coded as 4=agree or 5=strongly agree indicated that the participant felt that they had a nervous or anxious personality. Therefore, the number of these responses were calculated first from the total amount of participants, and second, separately from the groups of English majors and non-English majors. Table 2 below shows the descriptive statistics derived from the anxious personality self-evaluations.

The descriptive statistics in Table 2 show that 66% (N=82) of the participants felt that they had a nervous or anxious personality. The descriptive statistics revealed an interesting difference between the English majors and non-English majors: according to the Likert-scale responses, a vast majority (77%) of the English majors felt that they had an anxious or nervous personality, but only half (50%) of the non-English majors felt the same. This indicates that as a group, the English majors were generally more anxious than their non-English major peers. To examine the relationship between anxious personality and degree of FLA I conducted Pearson’s r correlation tests with the FLCAS scores and the Likert-scale responses of the anxious personality inquiry.

Among both learner groups the values of r (English majors r=0.45; non-English majors r=0.44)

indicated weak positive correlation with anxious personality and the FLCAS scores. The values of r² (English majors r²=0.20; non-English majors r²=0.19) suggested that 20% of the English majors and 19% of the non-English majors followed this positive correlation, and consequently a vast majority (80% of the English majors and 81% of the non-English majors) of both groups did not follow this correlation. Furthermore, as Table 2 shows, the English majors with anxious personalities scored lower on average on the FLCAS (M=83) than the non-English majors with anxious personalities (M=92). Firstly, this indicates that learners with anxious personalities did not yield alarmingly high levels of FLA in the Finnish FLCAS, which means that having an anxious or nervous personality does not automatically result to experiencing considerably high levels of FLA. Secondly, this finding suggests that FL anxiety is separate from general anxiety. To conclude, for a majority, anxious personality did not correlate with the degree of FLA, which confirms that FLA is a unique type of anxiety and a result of the language learning situation.