Emily Hall and Dr. Laura Smith, German and Slavic Languages Department

Study abroad programs have long enjoyed the prestige of being considered an exceptionally effective way to become fluent in a second language (Freed 1998). This conclusion has been supported by a multitude of studies. Smith, Aten, and Baker (2009) showed that German students on an 11-week study abroad in Berlin and on a similar study abroad in Vienna improved their accentedness, fluency, and comprehensibility in the course of the program. Moreover, Freed (1998) notes that many large-scale studies tend to find overall improvement in the listening and speaking skills of students participating in a study abroad. However, not all studies corroborate this claim. Freed herself found in the same study that in small-scale studies, some students do not progress. Martinsen (2008) found that 20-40% of students on study abroad make no measurable improvement over the duration of the program. The purpose of this study is twofold: first, to examine whether a study abroad has a positive effect on the pronunciation of students learning German, thereby providing a valuable piece of information for assessing whether or not a study abroad is helpful in achieving desired proficiency; second, to compare the improvement of three groups (beginning, intermediate, and advanced pre-program proficiencies), into which the students are divided by an assessment taken in the first week of the program. To guide the study, we asked the following questions:

1. Does study abroad have a positive effect on the students’ pronunciation of the target language?

2. Does the level of proficiency determined by the Goethe Institute correspond with the proficiency as determined by our native-speaker judges?

3. At what level of initial proficiency is study abroad most effective in improving pronunciation? The subjects of the study are 33 native English speakers from the US and Canada—23 women and 10 men. All subjects were students between the ages of 19 and 27 at Brigham Young University taking part in a short-term study abroad program in Berlin. Their experience with speaking German ranged from a few months to 7 years. Some had never visited Germany before this study abroad experience, while others had previously spent over two years in Germany. Subjects were divided into three groups based on German language ability: A (beginning), B (intermediate), and C (advanced). These groups were determined by a test administered at the Goethe Institute in Berlin at the beginning of the study abroad.

Subjects were asked to read the passage “Der Nordwind und die Sonne” twice aloud. They were recorded in a quiet room using a TalkTune microphone and an iPod. The process was repeated after an eight-week course of study at the Goethe Institute in Berlin. Using Praat, each recording was spliced into individual sentences as separate sound files with all names removed to preserve anonymity. For each sentence, we included the recordings from before the study abroad and from afterward.

Four sentences from the text were analyzed using a native speaker ratings task, administered via the online survey delivery program Qualtrics. The native speaker judgment task was broken into two shorter tasks to avoid burnout. The first task contained two of the sentences, with the 66 recordings of the first sentence (33 students x 2 recordings (pre- and post-test)) making up the first half of the survey, and the 66 recordings of the third sentence making up the second half. All recordings of sentence 1, including pre- and post-tests, were presented together, randomized by the computer, in a single block, followed by a second block containing all recordings for sentence 3. This pattern was followed in the second judgment task with sentences 2 and 6. For each sound clip, the sentence being read was displayed on the screen above the sound file. Underneath were three 7-point Likert scales. Each Likert scale was for rating a different aspect of the pronunciation: native-like accented, fluency, and comprehensibility. These surveys were then distributed to native speakers of German. These native speakers were asked to judge each clip on its nearness to a native accent using the 7-point Likert scales, where 1 is native-like pronunciation and 7 is very un-native-like pronunciation. After receiving the results of the surveys, the ratings for the pretest were compared to the ratings of the posttest to determine if there was an overall pattern of improvement among the students.

Statistical analysis of the results of the native-speaker judgment task strongly suggests that each group (A, B, and C) improved their pronunciation overall by the end of the study abroad. Moreover, the ratings of our native-speaker judges corresponded with the proficiency levels assigned by the Goethe Institute. The final question, whether initial proficiency levels played a role in the amount of progress made during study abroad, yielded interesting results. While all groups improved pronunciation, the progress varied from group to group. Group A, the least proficient, improved the most by far. The improvements in accent and comprehensibility for group A were significantly greater than both group B and group C. However, improvement in fluency did not vary significantly from group to group, though the difference between A and C suggested that A improved more.

The greatest limitations on this research were related to the native speaker judgment task. Due to time constraints, we were only able to have four sentences judged out of the eight that were recorded. It was also difficult to find native speakers who had grown up using the language, still spoke it as their primary language, and were willing to complete both parts of the native speaker judgment task. Because the tasks were time-consuming (each part took about 30-40 minutes), many qualified judges were not willing to complete the tasks. It is also recognized that having the speakers read aloud to gather the tokens may have had an impact on their pronunciation. Many people are more conscious of their pronunciation when they are reading a text rather than speaking freely. Therefore, the results may have been different if the tokens were taken from free speech than from read materials. However, this would also cause inconsistencies in that the tokens would potentially no longer use the same aspects of the German language.

Works Cited

• Freed, Barbara F. (1998). An overview of issues and research in language learning in a study abroad setting. Frontiers: The Interdisciplinary Journal of Study Abroad, 4(2), 31-60.
• Martinsen, R. A. (2008). Short-term study abroad: Predicting changes in oral skills. Foreign Language Annals, 43(3), 504-530.
• Smith, Laura Catharine, Hillary Aten, and Wendy Baker. 2009. “The effects of short term study abroad on the second language pronunciation of German.” Paper presented at GLAC (Germanic Linguistics Annual Conference), at Banff, Alberta, May 2009.