Tutoring Second Language Writers (22 page)

A case study is more than a story of what happened with a particular tutee, which would be an anecdote or a narrative. In a case study, you start with
research questions
and then gather data to help answer these questions. The research questions can be generated by your experience, as explained above. Once you have your questions, determine what would be the best kinds of data to answer them. Some types of data you might look at are actual tutoring sessions, written work, and interviews with relevant stakeholders. Data collection is bounded (which means you set limits on it; you don’t observe and collect everything) and continues until you have satisfactory data to answer your question. The write-up of a case study is usually in the form of a narrative and includes ample examples and quotations to illustrate your points. A famous example of a case study in the writing center literature is “Whispers of Coming and Going: Lessons from Fannie” (
DiPardo 1992
). Although not an actual study of writing center tutoring (the context was a writing group connected to a basic writing class), this was one of the first published studies to include actual data of student and tutor dialogue. It appeared in the
Writing Center Journal
, the flagship publication of the International (then National) Writing Centers Association. It has also been widely anthologized.

Once you have collected data you will need to analyze it.

When observing and recording tutoring sessions, tutors may want to use
close vertical transcription
as described by
Gilewicz and Thonus (2003)
. In this system, tutoring sessions are transcribed not as a “play” in which the participants nicely take turns but rather as a real dialogue with pauses and overlaps. Close vertical transcription allows you to more realistically represent tutorial dialogue for analysis.

Here is an example from Gilewicz and Thonus’s article in regular transcription:

In the above example, although there are four speakers, the dynamic nature of the conversation is not evident. The same conversation in close vertical transcription would look like this:

In the second transcript, you can see the overlap of conversation through the brackets. When two lines are bracketed it means that two people are speaking at once for as long as the lines are vertically aligned. In addition, this second transcript gives information on pauses. A short pause is symbolized by a period inside parentheses (.) and longer pauses are indicated by the number of seconds of the pause (2s). This exchange is lively, so it does not contain longer pauses. The series of three asterisks indicates an indecipherable word or phrase. You can see
the rest of these transcription conventions by consulting
Gilewicz and Thonus’s (2003)
article. To understand the difference in the methods, try reading each version of the transcript aloud in a group.

Close vertical transcription also allows researchers to look at backchannels (
uh-huh
,
yeah
,
OK
), which are different from overlaps or interruptions in that the speaker maintains the floor while the backchannel is simply affiliative. One investigator has found that L2 tutees tend to backchannel more than their native-speaking peers (
Thonus 2002
). Through the use of close vertical transcription, tutors could test this finding with their own data.

Once you have transcribed an interesting portion of the transcript, you have a choice of several methods of coding and analysis. Two common methods are
conversation analysis
and
discourse analysis.
Close vertical transcription more closely resembles transcriptions used in conversation analysis, which is concerned more with speed, intonation, overlap, and repair, while discourse analysis looks more closely at
what
is said rather than
how
it’s said. For instance, if you were to code the above excerpt for pronoun use (part of discourse analysis), you would note that all pronouns used are first-person singular except for the pronoun in line ten. This observation would lead you to ask about the relationship of the speaker to both the topic and the interlocutors. Why does everyone use
I
until A uses
you
in line ten? My analysis would be that here, A is invoking an audience, while in the other exchanges, the
I
is truly the persona of the speaker. I could look later on in the session and note whether the word
you
is used not to address the interlocutor but to invoke audience and would thus be a signal to the listener that audience was being invoked. If I were to code this segment for conversation analysis, I would look at overlap, latching, and interruption. In this segment I could look at lines four, five, and six with the echoing and overlap of “really good” and determine that this is an affiliative overlap, not meant to wrest the floor away from the speaker but to positively enhance and encourage the writer. Another coding example is F’s attempts to take the floor in lines nine and eleven. F’s attempts (twice saying “but like”) were unsuccessful. In this case F had something to say but was not able to enter the conversation. Although traditional conversation analysis eschews the introduction of social, racial, and cultural elements into the analysis, the use of
critical discourse analysis
enables the researcher to point out instances of dominance and social oppression. For instance, if F were a woman, critical discourse analysis
may point out gender bias as a reason she was unable to gain the floor. An understanding of these conversational moves can contribute ultimately to the evaluation of the overall effectiveness and success of a session.

In this section I will discuss a couple of studies of L2 writers so you can see how they were put together and how to construct your own project. The first study I will discuss is “Learning More from the Students” by
Kiedaisch and Dinitz (1991)
. Through
postconference surveys
, they learned that L2 tutees were less satisfied with their tutoring sessions than were their L1 peers. The initial research questions related to satisfaction and demographic details. In other words, the researchers wanted to know how satisfaction with the session related to factors like gender, linguistic status, and learning disability. After each tutoring session, researchers gave the tutee a satisfaction survey with Likert-scale items. For instance, one question read, “Were your goals for the session met?” and possible answers were on a scale of one to five with one being “not met at all” and five being “fully met.” After one year of data collection, the data set consisted of 376 surveys, which had to be hand entered into a database. Jean Kiedaisch and Sue Dinitz did what many humanists would do: they turned to a statistician to analyze the data rather than attempting their own analysis. Although Cindy
Johanek (2000)
exhorted writing center and composition scholars to engage in statistical analysis, my personal comfort zone has not yet extended to performing with SPSS.

The authors found a significant difference in satisfaction among groups. Learning-disabled students gave their tutoring sessions the lowest ratings. ESL students also rated sessions lower than did other groups. From these results, Kiedaisch and Dinitz planned to focus on improving the training for tutors in working with ESL students. The strength of this type of survey research that gathers data directly from participants, is that it can be used to confirm hunches. If ESL students seem less satisfied, we can design a survey to confirm the phenomenon before formulating a plan of action. Plus, it’s convenient to ask people to take surveys in the writing center after or before their tutorials, and it shouldn’t take up too much of their time.

Two decades ago, Janet
Moser (1993)
noticed that ESL students visited her writing center once or twice and then did not return. This
observation caused her to wonder if the writing center was not meeting their needs and what could be done to improve the situation. She designed a study lasting two semesters in which she observed and videotaped tutoring sessions between Haitian undergraduates and US peer tutors and also interviewed participants about their experiences and suggestions. Moser’s study participants included five tutees and two tutors, and each tutee was observed in one tutoring session. Moser used a technique called
stimulated recall
in which she showed participants the videotapes of the session and recorded their reactions, thoughts, and suggestions. We see this technique used sometimes in reality television shows when the participants react to events in “talking-head” segments interspersed with the action. The tutors noted that it was difficult for them to elicit answers from tutees, and they ended up answering their own questions after what they perceived as uncomfortable periods of silence. When tutees did speak, the tutors had difficulty understanding them. The tutees said they wished the tutors were more knowledgeable about grammar and preferred to conference with their professors rather than peer tutors. The Haitian students also reported forming study groups with their same-culture friends.

Moser suggested that this study could be expanded to look at tutees from various cultures and to explore possibilities in tutor training. For example, she suggested that a training program could be implemented to better prepare tutors to work with ESL writers. She also suggested that the writing center could offer a place for these study groups to meet, with a tutor available if questions arose. Moser acknowledged that without doing this study, she would not have been aware these study groups even existed.

In all these studies, the research questions were determined by moments of curiosity or unease. The methods were determined by the problem the researcher wanted to solve. In several cases, the researcher used
triangulation
to look at the data from various angles. For instance, Moser observed sessions and then asked tutors and tutees to go over their ideas and reactions with her. Keidaisch and Dinitz used the survey method, but they could also have expanded their data collection to include observations of tutoring sessions and interviews with participants.

It’s important to remember that before doing any research with human participants, you should contact the office or person on your campus
in charge of research that involves human subjects. Your research will likely be exempt from review since you will be studying normal educational practices, but you will probably be required to fill out paperwork concerning your study since you cannot declare your own study exempt. Some institutions such as community colleges may not have an Internal Review Board (IRB), but you still must gain informed consent from your study participants and also obtain permission if you plan to publish your research later. You also should ask them for their permission to share video and audio recordings of them or their written work in publications or at scholarly meetings. Most study participants are given pseudonyms by the researcher or choose their own. Some may ask that you use their real names. In any case, respect your research participants and do not do anything that may expose them to harm or embarrassment. Some researchers use
member checking
as a way to respect the ideas and opinions of participants. When I did my dissertation study, I submitted all transcripts of tutoring sessions and interviews to participants for vetting. In some cases I removed instances of
um
or
you know
at participants’ request. Since I was not studying pause fillers, these changes were appropriate. Of course I could not have done that if my research focus had been on verbal hesitations. Once I had written drafts of my analysis, I submitted chapters to participants for their comments, and where their understandings of what was happening differed from mine, I included their analyses as well in order to give a more balanced treatment. Finally, guard your data very carefully, especially in cases in which you have not yet changed the real names to pseudonyms. Password protect your computer and any removable media you plan to use. Keep written transcripts in a locked desk drawer or locked office. Shred printouts once you are done working with them. All this is not to scare you but just to make you aware that laws protect human subjects from coercion and breaches of confidentiality. You will not be able to share your research in a public forum if you do not follow the policies of your school’s IRB. For example, journal and book editors may ask to see your written consent forms before publishing your study.