I chose four different task types in order to provide test-takers with the possibility to demonstrate their abilities (see Appendices 3 and 4). The task types include:
1. a narrative task, 2. a comparison task, 3. an instruction task and 4. a role play / simulation task.
The first task type, the narrative task, is typically based on picture sequences, where the picture content largely determines what will be said (Luoma 2004: 144). In both of the tests, this task consists of a cartoon strip which the test-takers are asked to observe. They’re then required to narrate the story it depicts. The comic strip in the face-to-face test has two characters, a man and his son, who are out on a ride when the car breaks down. In addition to recounting the events, test-takers are asked to describe the characters and to mention what they think is wrong with the car. In the computer-based test, the comic strip shows the same man and son as well as a third character, the mother, who has prepared dinner. She asks her husband to get their son, who is in his room reading a book. The boy joins his mother at the table, but the father has stayed behind to read his son’s book. In addition to describing the plot and the characters, the examinees were asked to reflect on the type of book the father and son are reading.
The second task consists of two pictures that the test-takers are asked to compare. This type of task was chosen because, as Luoma (2004: 147–148) claims, compare and contrast tasks entail a greater cognitive load than description tasks for their completion due to the fact that test-takers are required to analyze and discuss similarities and differences, which in turn requires the use of comparative forms and complex grammatical structures. In the face-to-face task, the two pictures show a male student working alone and a group of students working together. The informants were asked to name advantages and disadvantages of working alone vs. in a group and to justify their opinion regarding their preference of working style. In the computer-based test, the pictures depict two different school settings: one is located
in a developing country and is highly teacher-centered, while the other one is in a developed country where technology is used in the classroom. In addition to again pointing out the differences, the informants were asked to list pros and cons of using technology in teaching vs. not using it. Therefore, in addition to comparing and contrasting, they had to justify their opinion(s). I had originally embedded two short videos to be compared in the computer-based test, but when piloting the task, it became obvious that it was too demanding. The students claimed that it was difficult to recall the information provided in the videos and to then compare and contrast them. Testing short-term memory was considered to result in construct-irrelevant variance, so I opted for pictures instead.
The third task is an instruction task, the main purpose of which is “getting the message across and making sure that it has been understood” (Luoma 2004: 146). In the face-to-face test, test-takers were given a map of Helsinki and they were supposed to instruct their friend, who is visiting from England, on how to get from place A to place B. I pretended to be the friend in this task. In the computer-based version of the test, the examinees were provided with the same map, but this time their friend had sent them a text message claiming to be lost and needing help getting from place to place. The test-takers were meant to leave him/her a voice message with instructions. In the version that was piloted, the computer version of the task required the examinees to give dog care instructions. However, most of the students claimed that this was difficult because they did not own a dog. The task was left out on the grounds that it was unfair towards some of the test-takers due to differences in topical knowledge.
The fourth and last task is a simulation task, in which students pretend to take part in a job interview. According to Luoma (2004: 151) “[r]ole-plays simulate different kinds of communication situations that the target group of the test could plausibly meet outside the test”. A job interview seemed like a natural choice, given that the test-takers were at an age where they start looking for their first summer or part-time jobs. The examinees were first asked to choose between three different job advertisements, all of which I designed based on what I thought would interest adolescents. I had additionally planned the questions I would ask during the job interview simulation task so as to make sure that the task was as standardized as
possible. In the face-to-face test, I then acted as the interviewer, whereas in the computer-mediated test, an avatar, Kate, did the interviewing. The avatar was designed through a web-based educational tool, Voki, that allows users to create and customize a character, which can then be embedded into Moodle. Unfortunately, the avatar cannot be made interactive.
From the above task overview, one can see that the first three tasks are largely monologic. In the face-to-face test, I tried to intervene as little as possible and mostly resorted to backchannelling. However, I did help the test-takers if they were facing difficulties moving on. The last task in both tests is dialogic because I wanted to capture the interactive nature of speech in as far as it is possible to realize with both formats. The reason I did not want to design the face-to-face test exclusively in the form of an interview is that, in test situations, interviews typically lead to an imbalance in the amount of speech produced by the interviewer and the interviewee (Young and Milanovic 1992; Johnson 2001).