Researchers display confirmation bias when they persevere by revising procedures until obtaining a theory-predicted result. This strategy produces findings that are overgeneralized in avoidable ways, and this in turn hinders successful applications. (The 40-year history of an attitude-change phenomenon, the sleeper effect, stands as a case in point.) Confirmation bias is an expectable product of theorycentered research strategies, including both the puzzle-solving activity of T. S. Kuhn's "normal science" and, more surprisingly, K. R. Popper's recommended method of falsification seeking. The alternative strategies of condition seeking (identifying limiting conditions for a known finding) and design (discovering conditions that can produce a previously unobtained result) are result centered; they are directed at producing specified patterns of data rather than at the logically impossible goals of establishing either the truth or falsity of a theory. Result-centered methods are by no means atheoretical. Rather, they oblige resourcefulness in using existing theory and can stimulate novel development of theory.

Five studies examined how college students react to child eyewitnesses in criminal cases, in Study 1, subjects made predictions about a staged crime study involving eyewitnesses of varying age. The actual study found no age differences, yet subjects predicted poorer recall (but not face recognition) for children under 10 than for those 12 and over. In Studies 2-5, subjects read and reacted to written criminal cases in which the principal prosecution eyewitness was either a child or an adult. In Study 2, a 6-year-old eyewitness was judged less credible than an adult eyewitness, and fewer guilty judgments were rendered when the eyewitness was 6 or 10 years old (versus an adult) and the prosecution case was otherwise strong. In Study 3, apparent inconsistency lowered the credibility of a 6-year-old eyewitness but not of a 10-or 30-year-old eyewitness. These studies presented a narrative description of the trial and the testimony. In Study 4, a transcript of testimony was also presented. Here a 6-year-old eyewitness was judged more credible than an adult eyewitness, and more guilty judgments were rendered (before and after jury deliberation) when the case included the child eyewitness. Study 5 manipulated the description-or-transcript distinction and observed a bias against cases with a child eyewitness when mock-jurors simply knew eyewitness age (as in Study 2) and a bias favoring these cases when mock-jurors read the eyewitness testimony (as in Study 4). This pattern suggests the important roles of jurors' preconceptions, eyewitness behaviors, and whether the latter confirm or disconfirm the former.Many studies have examined adults' eyewitness testimony, while others have focused on how jurors react to adults' testimony. Only recently has the attention of psycholegal researchers turned to child eyewitnesses, perhaps, in part, as a result of the growing presence of children in the courtroom as eyewitnesses (e.g.,

The sleeper effect in persuasion is a delayed increase in the impact of a message that is accompanied by a discounting cue. Despite a long history, the sleeper effect has been notoriously difficult to obtain or to replicate, with the exception of a pair of studies by Cruder et al. (1978). We conducted a series of 16 computer-controlled experiments and a replication of the Cruder et al. study to demonstrate that a sleeper effect can be obtained reliably when subjects (a) note the important arguments in a message, (b) receive a discounting cue after the message, and (c) rate the trustworthiness of the message communicator immediately after receiving the discounting cue. These operations are sufficiently different from those used in earlier studies to justify a new differential decay interpretation of the sleeper effect, in place of the dissociation hypothesis favored by most previous sleeper effect researchers. According to the differential decay interpretation, a sleeper effect occurs when message and discounting cue have opposite and near-equal immediate impacts that are not wellintegrated in memory. The effect occurs, then, if the impact of the discounting cue decays faster than that of the message.A sleeper effect in persuasion is a delayed increase in the impact of a persuasive message. The term was first used by Hovland, Lumsdaine, and Sheffield (1949) to describe opinion change produced by the U.S. Army's Why We Fight films used during World War II. As a pattern of data, the sleeper effect is opposite to the typical finding that experimentally induced opinion change dissipates over time . As such, the sleeper effect is an "interesting quirk" that has attracted much research and textbook attention.Early in its history, the sleeper effect became identified with the dissociation hypothesis and was denned as a delayed increase in persuasive impact that occurs as a result of a persuasive message accompanied by a discounting cue. Close scrutiny of previous sleeper effect research, however, reveals that much Figure 1C), although taking the form of a delayed increase in agreement with the communication, is conceptually more similar to decay of (negative) persuasion than it is to the sleeper effect.

College students heard a strong or weak message after learning whether the message issue would have relevance to their personal lives outside the laboratory (high or low issue involvement) and whether they would later discuss the message issue (high or low response involvement). Judging from subjects' recall of message information, either high issue involvement or high response involvement was sufficient to instigate high levels of attention to the message. Issue-involved-only subjects, however, were most strongly influenced by message quality. They agreed more with and had more favorable thoughts about strong relative to weak messages, and they were most likely to engage in attitudeconsistent behavior. Response-involved-only subjects were not affected by message quality, either on public attitude and thought measures or on a private behavioral measure. Response-and-issueinvolved subjects were in between these extremes. Message quality had modest effects on their thoughts and attitudes, but not on their behavior. These results suggest that issue involvement encourages systematic processing that is sensitive to how well message arguments concur with personal standards. In contrast, response involvement encourages expression of attitudes that satisfy selfpresentational needs. This expression may be mediated by message processing that is either biased toward moderation or nonintegrative, or by outward impression management, or both.

Issues regarding the fairness of lineups used for criminal identification are discussed in the context of a distinction between nominal size and functional size. Nominal size (the number of persons in the lineup) is less important for determining the fairness of a lineup than is functional size (the number of lineup members resembling the criminal). Functional size decreases to the extent that the nonsuspect members of the lineup are easily ruled out as not being suspected by the police. The extent to which the identification of the suspect can be considered an independently derived piece of incriminating evidence is positively related to functional size. Empirical estimates of functional size can be obtained through pictures of the corporal lineup from which mock witnesses make guesses of whom they believe the police suspect. A distinction is made between a functional size approach and hypothesis testing approaches. Uses of functional size notions in the court, by police, and in research are discussed.