Results for each RQ will be discussed in turn. Quantification methods, including inferential statistical tests, will be described as appropriate for each presentation of data.
RQ 1: Are the learner variables of gender, primary motivation for language study, primary focus in class, contact with NSs, time abroad, and achievement associated with significantly different ratings of the attainability of the 19 features of German taken together?
RQ1, different from RQ3, looked at aggregate scores (taking all 19 items together). It was answered by performing oneway ANOVAs as an omnibus test, followed by post-hoc Scheffe tests for all non-bi-variate variables, that is, all variables but gender. The alpha level of statistical significance, as with all statistical analyses in this paper, was set at p < .05, although, as appropriate, instances of marginal (near) statistical significance will also be discussed.
Results showed only one variable to yield unambiguously significant results, namely achievement. Initial ANOVAS had yielded a p-value of .021* (F, 3.892) and post-hoc Scheffes pinpointed the differences to be between learners with the most recent final course grade of C or D on the one hand and, on the other, both learners with a final grade of A or AB (p= .033) and learners with a final grade of B or BC (p= .022*). The weakest students perceived the 19 features of German to be much less attainable. Contact with NSs accounted for an initial statistically significant difference, as assessed in the oneway ANOVA (F, 4.214; p= .016) but post-hoc Scheffes showed only a marginal significant difference (p=.056*) between students with minimal and those with extensive contact with NSs. Students with extensive contact with NSs had given the most optimistic assessments of their expected final attainment of the 19 features under investigation. Orientation in class was associated with only marginal statistical significance in the one-way ANOVAs (F, 2.234; p= .051*). The post-hoc Scheffes showed no statistically significant difference among the groups although there seemed to be a trend toward a divergence between students with a meaning and those with a fluency orientation in class (p= .062). Students with a meaning orientation had given the highest mean scores, indicating a lesser belief in the attainability of the 19 features; students with a focus on fluency had given the lowest mean scores, that is, were the most opimtistic group.
RQ 2: Which features of German do Year 1, 2, and 3 learners, respectively, consider to be particularly difficult to attain?
RQ 2 was approached using both descriptive (Table 2) and inferential (Table 3) statistics. Table 2 shows the mean scores and relative ranks of each of the 19 features of German as rated by Year 1, Year 2, and Year 3 students, respectively. The order progressed from forms that were considered to be the most easily attainable (and received the lowest mean scores) to those that were considered the most difficult to attain (and received the highest mean scores). In keeping with the objective of RQ 2 and different from RQs 1 & 3, results displayed in Table 2 distinguish among the three years of enrollment.
Table 3. Features of German for Which Learners in Different Years of Study Judged Attainability Differently
| Feature (Item) || Initial (omnibus) ANOVA || Post-hoc Tukey HSD || |
| ||F|| p ||Groups implicated by Tukey HSD|| p ||Mean difference|
|Case endings||3.023||.050*||NA (no between-group differences found)|| || |
|Possessives||3.021||.050*||Year 1/Year 3||.047*||.390|
|Forms of address||3.815||.023*||Year 1/Year 3||.036*||.415|
|Adjective endings||6.328||.002**||Year 1/Year 2||.031*||−.322|
| || || ||Year 1/Year 3||.002**||−.530|
All three student groups considered subjunctive, passive, and adjective endings and other nominal morphology, such as noun gender, case endings, and, with the exception of Year 1 students, plural endings, the most difficult to attain. All learner groups further determined forms of address to be the easiest to attain, followed by verb placement and subject-verb agreement. In terms of phonology, Year 1 learners considered segmentals (sounds) easier to attain than prosodic features (word stress, sentence melody) while the opposite was true for their Year 2 and 3 peers.
The already low scores assigned to forms of address and verb placement were lowest among Year 3 learners and, conversely, relatively highest among Year 1 students. However, Year 3 students assigned even higher scores (signalling a greater belief in these features’ difficulty) to forms that were generally considered more difficult to attain (i.e., held a lower rank and received overall higher scores), such as case endings, adjective endings, the passive voice, or subjunctive. As a matter of fact, Year 1 students were more optimistic than Year 2 or Year 3 students about their ability to eventually acquire these very difficult forms. In other words, as they advanced, learners began to draw greater distinctions among the relative difficulty of forms and also became more reluctant to believe they would ever acquire all—including the more difficult — forms of German. To accommodate the different spreads of mean scores across the three years of study, subsequent explanations refer to the relative rank of a feature in a given learner group (Year 1, 2, and 3 learners), not the feature's attached mean score. That is, although features may be associated with similar mean scores, they could nevertheless occupy very different ranks across two or more groups.
Table 3 provides an overview of statistical tests that compared the responses of students in the three years of study under investigation. Post-hoc Tukey HSD tests followed omnibus ANOVAs. The alpha level, again, was set at p < .05. The difference in means is phrased from the perspective of the group mentioned first. Since the between-group tests were repeated for each of the 19 language features, Type I error cannot be ruled out. Therefore, it is important to appreciate results as exploratory (not definitive) and pay attention to patterns of findings. Moreover, different from the discussion based on ranks (above), the comparison of means does not take into account the degree to which scores spread (or did not spread) across the 19 features within each learner group.
Overall, results shown in Table 3 coincide with trends revealed in the inferential statistics that were displayed in Table 2. Statistically significant between-group differences (a) occurred in a manner that set apart Year 1 students, mostly from their Year 3 peers; (b) foremost concerned grammar, particularly nominal morphology; and (c) involved the one non-grammatical feature among the 19 total features (forms of address) that is heavily emphasized in the first-year curriculum and, perhaps most importantly, is explicitly linked to the choice of certain grammatical forms. Last, perhaps in a demonstration of ignorance being bliss, Year 1 learners, who in other regards were less optimistic about their final attainment, were more certain than either Year 2 or Year 3 learners that they could eventually acquire adjective endings.
RQ 3: How do learner variables other than year of enrollment, such as learner gender, primary motivation for language (German) study, focus in language class, contact with NSs, time abroad, and achievement, account for differences in beliefs about the attainability of specific features of German?
As was done for RQ2 with regard to the learners’ year of enrollment, individual items were also compared by the variables of achievement, contact with NSs, time abroad, primary motivation for German study, and primary focus in class. For that purpose, learners from all years of enrollment were taken together and then repeatedly divided by each of the variables under investigation. As for Table 3, initial omnibus ANOVAs were followed by post-hoc Tukey HSDs and the possibility of Type I error pertains so that results need to be evaluated with caution.
Initial analyses, as discussed under RQ1, had identified achievement (last final German course grade) as a predictor of differences in beliefs about final attainment. Still, analyses for RQ1 had relied on agreggate means (taking means from scores for all 19 features together) and could have buried differences that pertained to individual items. Indeed, analyses under discussion here confirmed that achievement accounted for significant differences in a number of language features (Table 4). Apart from signficant differences attested for the variable of achievement, results for RQ1 had pinpointed contact with NSs and, to a lesser extent, orientation in class, to be associated with marginally significant differences. As Table 5 (further below) demonstrates, contact with NSs remained a productive variable in item-by-item analyses, orientation in class less so. Instead, experience abroad and primary motivation for language (German) study entered the picture. In the discussions of Tables 4 and 5 below, respectively, primary attention will go to how items associated with significant differences patterned and how contrast or overlap across learner variables manifested.
Table 4. Features of German for Which Learners with Different Final Course Grades Judged Attainability Significantly Differently
| Feature (Item) || Initial (omnibus) ANOVA || Post-hoc Tukey HSD || Difference in means |
| ||F|| p ||Groups implicated by Tukey HSD|| p || |
|Nominal word order||5.485||.005**||A&AB/C&D||.004**||−1.138|
| || || ||B&BC/C&D||.004**||−1.205|
| || || ||B&BC/C&D||.006**||−1.091|
| || || ||B&BC/C&D||.041*||−1.011|
| || || ||B&BC/C&D||.009**||−1.136|
Table 5. Features of German for Which Attainability Was Judged Significantly Differently by Learners Distinguished by …
| Feature (Item) || Initial (omnibus) ANOVA || Post-hoc Tukey HSD || Difference in means |
| ||F|| P ||Groups implicated by Tukey HSD|| p || |
| … the Extent of Their Contact with Native Speakers |
|Nominal word order||5.454||.005**||None/Extensive||.006**||.419|
| || || ||Minimal/Extensive||.039*||.326|
| || || ||Minimal/Extensive||.000***||.505|
| … the Duration of Their Experience Abroad |
|Noun gender||2.673||.047*||None/Long Time||.035*||−.625|
|Situational appropriateness||2.673||.047*||None/Short Term||.032*||.329|
| … Their Primary Focus in Class |
| … their Motivation for Studying German |
|Verb tenses||2.939||.021*||Requiremenl/Professio nal Uses||.014*||.543|
|Situational appropriateness||5.301||.000***||Requiremenl/Love of Language Learning||.000***||.655|
| || || ||Requiremenl/Communication with NSs||.009**||.528|
Table 4 shows that significant differences between student groups with different final grades concerned their beliefs about the ultimate attainment of nominal word order and verb placement. Whereas, on average, these two forms were considered relatively easy (ranks 3–4 and 2–3, respectively, as seen in Table 2), less-successful learners considered them to be relatively difficult to acquire. Significant differences also emerged for all questionnaire items that had to do with phonology (sounds, word stress, and sentence melody). With one exception, all apparent differences involved students who had earned a most recent final course grade of C or D. As a rule of thumb, C & D students were less optimistic about their ultimate attainment in these features. The most prominent exception to this pattern emerged in the context of B & BC students, who declared themselves more optimistic about their attainment of German sounds than their A & AB peers. Although a unique and perhaps aberrant result, it gains importance in light of the prominence that phonology seemed to play in the beliefs of weaker as compared to stronger students. With middling (B & BC) students apparently more optimistic than either their A & AB or their C & D colleagues about their ultimainte attainment of German sounds, it seems possible that learners framed their expectations of success in phonology both in terms of learning difficulties (C & D students) and a more acute awareness of the precise demands (A & AB students). To summarize Table 5, below, (a) learners with no contact as well as, to a lesser extent, learners with minimal contact with NSs tended to be less optimistic than learners with extensive contact with NSs about their ultimate attainment in 5 listed features of German (nominal word order, personal pronouns, possessives, verb tenses, situational appropriateness); (b) similarly, learners with no time abroad were less optimistic about 3 features of German (possessives, conjunctions, situational appropriateness) than learners who had spent some time abroad but more optimistic about their final attainment of noun gender; (c) students whose focus in class was fluency were more optimistic than students who were concerned with meaning in terms of the acquisition of 1 feature (verb placement); (d) learners who took the class to satisfy a requirement were less optimistic than learners who studied German for the love of language learning, professional uses, or communication with NSs in 2 features (verb tenses, situational appropriateness); and (e) gender, different from Henry (2009), did not account for any between-group differences at all, which is surprising considering the productivity of the variable in analyses that concerned students’ beliefs about comprehensibility and pleasantness and that are not reported on here but reported (Chavez, in press).
In sum, between-group differences overall developed in the expected direction as the following types of students expressed lesser optimism than their peers about the attainment of at least one feature of German: (a) weaker students; (b) students who had not experienced authentic contacts; (c) students who emphasized fluency; and (d) students who took German primarily to satisfy a requirement. There were no consistent patterns in the specific features of German that were particularly prone to differential assessments although greater authentic experience appeared to be associated with greater confidence about the attainment of (a) possessives and, less consistently, personal pronouns, two forms that occur frequently in natural speech; and (b) situational appropriateness of language use, an objective that is difficult to achieve within the institutional speech patterns of the classroom (Seedhouse, 2004).