Anchoring effects on prospective and retrospective metacomprehension judgments as a function of peer performance information

Abstract

Two experiments were conducted to investigate anchoring effects on metacomprehension judgments as a function of fictitious information participants received about past peer performance. In Experiment 1 participants were randomly assigned to one of the three anchor groups that, in some cases, provided past peer performance averages in terms of a percentage: a high anchor (85%) a low anchor (55%), and no anchor. Results showed that relative to the no anchor group, the low anchor group made significantly lower prospective judgments whereas the high anchor group did not make significantly higher prospective judgments. With a high anchor of greater magnitude (95%), Experiment 2 demonstrated more pronounced anchoring effects on prospective judgments: Relative to the no anchor group, the low anchor group made significantly lower judgments and the high anchor group made significantly higher judgments. In addition, Experiment 2 showed that anchoring effects occurred even on retrospective judgments. Overall, the results showed that peer performance information can serve as an anchoring point for metacomprehension judgments and the anchoring effects are tenacious given that they can even affect retrospective judgments of performance.

Appendices

Appendix I

Sample text and test

Affirmative action

Reverse discrimination, minority recruitment, racial quotas, and, more generally, affirmative action are phrases that carry powerful emotional charges. But why should affirmative action, of all government policies, be so controversial? In a sense, affirmative action is like other governmental programs, e.g., defense, conservation, and public schools. Affirmative action programs are designed to achieve legitimate government objectives such as improved economic efficiency, reduced social tension, and general betterment of the public welfare. While it cannot be denied that there is no guarantee that affirmative action will achieve these results, neither can it be denied that there are plausible, even powerful, sociological and economic arguments pointing to its likely success.

Government programs, however, entail a cost, that is, the expenditure of social or economic resources. Setting aside cases in which the specific user is charged a fee for service (toll roads and tuition at state institutions), the burdens and benefits of publicly funded or mandated programs are widely shared. When an individual benefits personally from a government program, it is only because she or he is one member of a larger beneficiary class, e.g., a farmer; and most government revenue is obtained through a scheme of general taxation to which all are subject.

Affirmative action programs are exceptions to this general rule, though not, as might at first seem, because the beneficiaries of the programs are specific individuals. It is still the case that those who ultimately benefit from affirmative action do so only by virtue of their status as members of a larger group, a particular minority. Rather, the difference is the location of the burden. In affirmative action, the burden of “funding” the program is not shared universally, and that is inherent in the nature of the case, as can be seen clearly in the case of affirmative action in employment. Often job promotions are allocated along a single dimension, seniority; and when an employer promotes a less senior worker from a minority group, the person disadvantaged by the move is easily identified: the worker with greatest seniority on a combined minority—non-minority list passed over for promotion.

Now we are confronted with two competing moral sentiments. On the one hand, there is the idea that those who have been unfairly disadvantaged by past discriminatory practices are entitled to some kind of assistance. On the other, there is the feeling that no person ought to be deprived of what is rightfully his or hers, even for the worthwhile service of fellow humans. In this respect, disability due to past racial discrimination, at least insofar as there is no connection to the passed-over worker, is like a natural evil. When a villainous man willfully and without provocation strikes and injures another, there is not only the feeling that the injured person ought to be compensated but there is consensus that the appropriate party to bear the cost is the one who inflicted the injury. Yet, if the same innocent man stumbled and injured himself, it would be surprising to hear someone argue that the villainous man ought to be taxed for the injury simply because he might have tripped the victim had he been given the opportunity. There may very well be agreement that he should be aided in his recovery with money and personal assistance, and many will give willingly; but there is also agreement that no one individual ought to be singled out and forced to do what must ultimately be considered an act of charity.

Affirmative action

• 1. The passage is primarily concerned with

  • a. comparing affirmative action programs to other government programs

  • b. arguing that affirmative action programs are morally justified

  • c. analyzing the basis for moral judgments about affirmative action programs

  • d. introducing the reader to the importance of affirmative action as a social issue

  • e. describing the benefits that can be obtained through affirmative action programs

• 2. The author mentions toll roads and tuition at state institutions in order to

  • a. anticipate a possible objection based on counterexamples

  • b. avoid a contradiction between moral sentiments

  • c. provide illustrations of common government programs

  • d. voice doubts about the social and economic value of affirmative action

  • e. offer examples of government programs which are too costly

• 3. With which of the following statements would the author most likely agree?

  • a. affirmative action programs should be discontinued because they place an unfair burden on non-minority persons who bear the cost of the programs.

  • b. affirmative action programs may be able to achieve legitimate social and economic goals such as improved efficiency.

  • c. affirmative action programs are justified because they are the only way of correcting injustices created by past discrimination.

  • d. affirmative action programs must be redesigned so that society as a whole rather than particular individuals bears the cost of the programs.

  • e. affirmative action programs should be abandoned because they serve no useful social function and place unfair burdens on particular individuals.

• 4. According to the passage, affirmative action programs are different from most other government programs in which of the following ways?

  • I. the goals the programs are designed to achieve

  • II. the ways in which costs of the programs are distributed

  • III. the ways in which benefits of the programs are allocated

    a. I only b. II only c. III only d. II and III only e. I, II, and III

• 5. It can be inferred that the author believes the reader will regard affirmative action programs as

  • a. posing a moral dilemma

  • b. based on unsound premises

  • c. containing self-contradictions

  • d. creating needless suffering

  • e. offering a panacea

• 6. The primary purpose of the passage is to

  • a. reconcile two conflicting points of view

  • b. describe and refute a point of view

  • c. provide a historical context for a problem

  • d. suggest a new method for studying social problems

  • e. analyze the structure of an institution

Appendix II

Supplemental information

Although the objective of this investigation was not to understand how text difficulty interacted with peer performance information to influence metacomprehension judgments, the following analyses were performed using text difficulty as a factor to provide supplemental information to interested readers.

Study 1

A General Linear Model (GLM) repeated measures procedure was used to analyze how prospective metacomprehension judgments were affected by anchor information (between-subjects variable) and text difficulty (within-subjects variable). Results revealed that anchor information significantly affected the magnitude of judgments, F (2, 87) = 17.29, p < .001. The magnitude of judgments significantly changed with text difficulty, F (2, 87) = 17.59, p < .001, but there was no significant interaction between anchor and text difficulty, p > .05.

Paired-sample t-tests further showed that prospective judgments on the Literature text which has an Flesch-Kincaid (F-K) readability level of 11.2 (M = 78.91%, SE = 1.3%) were higher than those on Affirmative Action which has an F-K level of 14.7 (M = 75.64%, SE = 1.5%), t (89) = 2.41, p < .05, which in turn were higher than those on Inventions which has an F-K level of 17.5 (M = 70.39%, SE = 1.6%), t (89) = 3.56, p < .01. Thus, the higher the text difficulty level, the lower participants’ prospective metacomprehension judgments.

Study 2

The same analyses were conducted in study 2 for both prospective and retrospective judgments. The findings were the same as the results of Study 1 except that the paired-sample t-test results showed that participants’ prospective or retrospective judgments did not significantly differ for Affirmation Action and Inventions. Specifically, paired-sample t-tests showed that prospective judgments on the Literature text (M = 78.17%, SE = 1.4%) were higher than those on Affirmative Action (M = 73.02%, SE = 1.6%), t (98) = 4.08, p < .001 or Inventions (M = 71.66%, SE = 1.7%), t (98) = 4.64, p < .001. But there was no significant difference between the judgments on Affirmation Action and on Inventions, p > .05. The same results were observed for retrospective judgments. Thus, although Affirmation Action and Inventions differ in difficulty based on their F-K levels (14.7 vs. 17.5), participants did not judge their performance to be significantly different on these two texts before or after test.