COGNITIVE EDITING OF ARGUMENTS AND REASONS FOR REQUESTS: EVIDENCE FROM THINK-ALOUD PROTOCOLS.

• Author: Hample, Dale

For several years, scholars have studied how people edit their arguments. In particular, researchers sought the standards arguers apply in deciding whether to suppress or produce an idea in discourse. These standards are thought to be a number of things at once: they are the editorial criteria, the rules people think ought to be respected, the normative background for the interaction, and part of people's implicit theories of communication.

The research to date has produced three classes of results. First of all, Hample and Dallinger determined what the editorial standards are (Hample, 1984; Hample & Dallinger, 1985). They finally settled on a list of seven substantive reasons for deciding to suppress an argument: that it won't work, that it is too negative to use, that it would damage one's own self image, that it would harm the other person, that it would injure the relationship between producer and target, that it would be false, or that it would be irrelevant. These standards are also the focal point of the present investigation. A later paper (Hample & Dallinger, 1992) reports that these fall into two or three classes: effectiveness, discourse competence, and person-centered issues. The effectiveness codes (the first two on the list of seven) often combine with the discourse codes (the last two) to form a single factor.

The second class of results has related preference for various of the codes to individual differences variables (e.g., Dallinger & Hample, 1991; Hample & Dallinger, 1987; Kosloski, Hinck & Dailey, 1991). This sort of research has dominated the research program, and has been summarized in Hample and Dallinger (1990). There are complex but sensible patterns of relationship between personality and use of the standards.

The third sort of work explores the effects of situation on editing. Researchers have found meaningful relations between editing and the valence (liked or disliked) of the message target (Dallinger & Hample, 1991), and between editing and the intimacy and extent of consequences implicated in the situation (Kosloski, Hinck & Dailey, 1991). Based on a cumulation of the data from the early studies, reasonably clear conclusions about the ways in which situational dimensions (e.g., intimacy, dominance, etc.) affect editorial choices are now available (Hample & Dallinger, in press). Hample and Dallinger (1998) have also shown how repeated rebuffs of one's arguments result in one's editorial standards becoming increasingly aggressive and rude.

All the work to date has relied mainly on two methodologies. The first two studies, those that generated the list of criteria, supplied respondents with a list of possible messages, and asked people to say which messages they would be willing to use and which they would not. In an open-ended format, respondents were then asked to give explanations for the suppression of the unendorsed arguments. These were coded into categories of editing standards, which were used in the later investigations. The more recent studies provide participants not only with lists of arguments, but also with a checklist of responses. This checklist includes the seven criteria mentioned above, as well as endorsement and residual choices.

While I am not dissatisfied with the methodological choices made during the course of the research program, I wish here to triangulate some of the most basic findings, and to explore editing in another research paradigm. This paper is an effort after those ends. Think aloud protocols (see Daly, Weber, Vangelisti, Maxwell & Neel, 1989; Ericsson & Simon, 1985; Flower & Hayes, 1984; Smagorinsky, 1989, 1994) are used here for several reasons. First, they free the respondents from the constraints imposed in the usual research design, since I will not be supplying lists of possible arguments or lists of possible judgments of them. Second, the think-aloud protocols permit us to observe (to some extent) the actual thought processes involved in editing. And third, they allow us to see revisions of messages, rather than the cruder utter-or-suppress editorial decisions that have been the sole focus of the earlier studies.

Therefore, the primary research questions motivating this study are:

RQ: Do Hample and Dallinger's editorial standards app ear in the open-ended data of think-aloud protocols? If so, how do they function in revision of people's persuasive request messages?

The qualitative method used in this study involved considerable viewing, re-viewing, reading and re-reading of the data. In the course of reflection on what people were doing during their request revisions, I noticed what I think is an interesting phenomenon concerned with how people justify their requests. I see no reason for the pretense that I had a research question on this issue, but wish here to give the reader notice that this paper will also report findings concerning what I call the "reason slot" in elaborated request messages. These are connected to the editorial criteria, as we will see.

METHOD

Respondents were 28 undergraduates at my institution. They were invited to come into a room temporarily set up as a communication laboratory so that they could be videotaped. They were seated at a typical classroom desk, and provided with paper and pen. They were told that the study concerned itself with how people compose messages.

Each person was given four written persuasive situations, one at a time, with the order of presentation systematically distributed within the sample. To ease worries about generalizability, several situations were used. These stimuli took different values on three variables: size of request (large or small), relationship with target (acquaintance, same sex friend, opposite sex friend, or romantic partner), and type of request (borrow money, borrow a car, tape record a class lecture, or deliver a term paper to a professor). This design resulted in 32 stimuli being used altogether. Each participant received each type of request, each type of relationship, and two each of the request sizes. For example, here is the borrow money/small size/opposite sex friend situation: "You need five dollars. Write a note to one of your close different sex friends (not a romantic partner), asking them to lend you five dollars for two days."

The undergraduates who served as respondents were told to write a note for the situation, and while writing, to speak aloud. Here is the oral instruction, given just after handing the respondent the written persuasive situation: "Start speaking your thoughts as soon as you begin to read the goal. Remember to say every thought that passes through your mind as you read this and write your note." Students were prompted whenever they paused noticeably.

The whole episode was videotaped, and the participants' utterances were transcribed. Analysis of the data involved both viewing the tapes and reading the transcripts. The original transcripts made use of standard conversation-analytic markings, but have been somewhat simplified for presentation here.

Results and discussion are organized according to the two main investigative foci: the research question dealing with the appearance of the editorial standards, and the nature of the "reason slot" in elaborated requests.

RESULTS, FIRST ISSUE: HOW PEOPLE EDIT PERSUASIVE MESSAGES

The point of this section of the paper is to replicate and extend the most basic of the editing research findings. I will provide evidence that people are aware of the editorial criteria (all but one of them, actually: the "too negative to use" standard proves to be too ambiguous to identify in ordinary talk, and may not be there at all). This evidence takes several forms: the criteria are mentioned as reasons for suppressing ideas, they are used as the bases for revision of ideas, or they are explicitly mentioned in appropriate contexts. I will begin with a study of the effectiveness and discourse issues, and then proceed to the person-centered standards.

THE EFFECTIVENESS STANDARD

The effectiveness standard--'I wouldn't use this one because it won't work'--is sometimes difficult to identify. In watching respondents think aloud, and considering what they mean by their words, I often suppose that they are worrying about effectiveness when another interpretation is possible. Therefore, I will confine myself here to instances that seem to be pretty straightforward. Readers should keep in mind that respondents are given clearly stated objectives (e.g., to borrow a car for a weekend), and this may have given special prominence to the instrumental goal.

For quite a few students, the effectiveness standard obviously dominated their planning from the beginning, and...