Within this study, it was investigated how university students can be encouraged to consider more counterarguments during writing argumentative texts. One hundred eighty undergraduates wrote essays on TV violence. In Experiment 1, students, which were given specific goals, generated more counterarguments and rebuttals than controls and than a group, who was stimulated to give reasons for arguments. In Experiment 2, one group of participants was provided with a text outlining arguments and counterarguments. Another group was asked to write a persuasive letter. Persuasion instructions had no or small negative effects. Text positively affected supporting reasons for primary claims and the overall quality of arguments. There were no effects of prior attitudes or gender. The danger of using persuasion goals and the advantages of specific goal instructions are discussed.

Writing texts with persuading arguments represents an important skill in academic, economic, legal, or political contexts, above all, when these contexts have potential for conflicts (e.g., Hofer & Pikowsky, 1993). There are several reasons, why "counterarguments" play a major role in argumentation. Such arguments concern propositions why a given position might not be true or not conclusive. Although counterarguments represent an important part of well-designed argumentative texts, there exists a "my-side bias", that is a tendency, not to state counterarguments (Perkins, Farady, & Bushey, 1991). High-school and university students hardly stated counterarguments, when they had to write about complex and controversial issues in daily life (Baron, 1995; Toplak & Stanovich, 2003). In recent years, research in the field of Educational Psychology focussed on goal instructions in order to reduce the problem of my-side bias (Covington, 2000; Klaczynski, Gordon, & Fauth, 1997). This kind of intervention contained instructions, in which writers of texts were informed about the goal, which they should achieve during writing.

The primary goal of this study was to replicate the remarkable findings of Nussbaum and Kardash (2005) with university students from US. Within a first experiment, it was intended, to produce a baseline for evaluating the quality of argumentative texts and for the effectiveness of goal instructions in respect to European and German speaking students. The study from Nussbaum and Kardash (2005) should, on the one hand, be replicated, because it showed strong effects of goal instructions on the production of counterarguments, although other studies reported difficulties in changing such productions. On the other hand, a baseline in respect to argumentation skills should be identified for non-US university students, which do, as a rule, not attend significant argumentation or critical thinking courses during high-school and/or university programs as it is often the case for university students from US. Overall, within this study, three types of interventions for improving the quantity of counterarguments and the quality of argumentative texts are tested: (a) specific goal instructions for stimulating the generation of certain elements of arguments (with stimulating reasons, counterarguments/rebuttals and reasons, balanced arguments, and counterarguments/rebuttals without reasons), (b) general goal instructions, which are focussing on persuading somebody, and (c) the presentation of a balanced text to an issue. It will also be examined whether there are interaction effects of these kinds of interventions with prior attitudes.

The results of the first experiment confirm the phenomenon of my-side bias within the given population. Like within the study from Nussbaum and Kardash (2005), there was a significant improvement in argumentation (frequency of elements and overall quality), when instructions were specifically related to the production of counterarguments. Negative effects were found, when no specific instructions were given at all and when instructions focussed on the production of reasons.

A second experiment was conducted for two reasons. First, within a second experiment, the effects of the general goal instruction, to persuade, was tested. In the first experiment, participants were stimulated to produce arguments without focussing on a certain intention. In the study from Ferretti et al. (2000), students had the task to write a persuading letter, in order to motivate a teacher-parents-association for a certain point of view. In the first experiment, the instruction of persuading others was omitted because it was expected that such an instruction could increase my-side bias. Especially, a person, who has to persuade others, could exclude arguments (i.e., counterarguments), which might weaken a certain position. In the second experiment, it was tested, whether the instruction to persuade somebody had an effect on my-side bias.

The second experiment was also undertaken, because there were no differences between the experimental conditions in respect to the frequency of primary claims and of supporting reasons within the first experiment. It is expected that knowledge about a certain subject will increase the production of argumentative elements and their reasons. Prior subject knowledge showed to have an effect on many cognitive phenomenons, also for argumentation skills (e.g., Allen, Berkowitz, Hunt & Louden, 1999; Stein & Miller, 1993). It can be expected, that the presentation of additional information about a certain issue will stimulate elaborations. These elaborations build a comprehensive basis of hints and background knowledge for the generation of arguments.

The results of Experiment 2 replicated the results of Experiment 1 concerning the problem of my-side bias. Once again university students delivered only very few counterarguments and rebuttals. In addition, they did not produce supporting reasons for these elements of arguments. The instruction to persuade and the presentation of a balanced text showed less strong effects than the specific goal instructions, which were tested in Experiment 1. Both, balanced texts and the instruction to persuade only weakly affected the frequency of counterarguments and rebuttals. The instruction to persuade reduced nearly all argumentative elements, especially also the frequency of supporting reasons for counterarguments. However, the presentation of a balanced text led to a significant improvement of the holistic quality of argumentation. Also, there were more primary claims and supporting reasons, when such texts were presented. It was also shown that the instruction to persuade led to a reduction in supporting reasons for primary claims, but only within the no-text condition. In contradiction to Experiment 1, prior attitudes were found to have an effect on argumentation, especially on the supporting reasons for primary claims. There were no gender differences on all outcome variables.

Finally, unexpected or inconsistent results of this study are discussed. These issues concern the goal instruction to persuade, the usage of balanced texts, and the effects of prior attitudes.

The results of this study could be used to formulate tasks in courses on argumentation and critical thinking. The balance in argumentation represents an important condition for identifying and evaluating the point of view of other people. That again is a core competence within a pluralistic knowledge society.