Estimating and testing mediation and moderation in within-subject designs.

Analyses designed to detect mediation and moderation of treatment effects are increasingly prevalent in research in psychology. The mediation question concerns the processes that produce a treatment effect. The moderation question concerns factors that affect the magnitude of that effect. Although analytic procedures have been reasonably well worked out in the case in which the treatment varies between participants, no systematic procedures for examining mediation and moderation have been developed in the case in which the treatment varies within participants. The authors present an analytic approach to these issues using ordinary least squares estimation.

The effects of perceptions of hazardous waste on migration: a laboratory experimental approach.

"This study develops a methodology that allows migration decision-making to be studied in a laboratory experimental setting. Moreover, this methodology permits an examination of the importance of natural and man-made hazards in migration decisions--factors that have not been extensively studied as determinants of migration. The specific application is to the location of the U.S. nuclear waste repository at Yucca Mountain, Nevada. Empirical results suggest that the repository may influence employment-related migration, but probably not retirement migration."

Statistical difficulties of detecting interactions and moderator effects.

Although interaction effects are frequently found in experimental studies, field researchers report considerable difficulty in finding theorized moderator effects. Previous discussions of this discrepancy have considered responsible factors including differences in measurement error and use of nonlinear scales. In this article we demonstrate that the differential efficiency of experimental and field tests of interactions is also attributable to the differential residual variances of such interactions once the component main effects have been partialed out. We derive an expression for this residual variance in terms of the joint distribution of the component variables and explore how properties of the distribution affect the efficiency of tests of moderator effects. We show that tests of interactions in field studies will often have less than 20% of the efficiency of optimal experimental tests, and we discuss implications for the design of field studies.

Communicating the risk from radon.

A prominent television station developed a special series of newscasts and public service announcements about radon. This was combined with their advertising of the availability of reduced-price radon test kits in a local supermarket chain. The large number of test kits sold was a success from a marketing perspective, but not from a public health perspective--especially because of the very small share of high readings that were mitigated. In contrast, a study of housing sales showed a much higher testing rate and corresponding mitigation when risk communication accompanied the housing transaction, rather than being directed toward the general public. This paper examines the relative effectiveness of these alternative approaches to radon risk communication, emphasizing the implications for developing and implementing radon programs.

Measuring contraceptive values: an alternative approach.

Previous assessments of individuals' values for various contraceptive consequences have employed one of four methodologies: free elicitation, direct ratings, multiple regression, or factor analysis. All four methodologies are flawed because they produce group rather than individual values, relying on rating scales, and fail to incorporate information regarding consequence trade-offs. Axiomatic conjoint measurement is proposed as an alternative methodology and used to determine individuals' values for a selected set of contraceptive consequences at two stages of the family-planning career.

PIP: The consideration of tradeoffs in the determination of individual values for contraceptive consequences is proposed as a superior alternative to free elicitation, direct ratings, multiple regression, and factor analysis. The technique is axiomatic conjoint measurement and requires individual rankings of a set of choice alternatives according to preference; it is different from consumer research in that it proposes and tests composition rules, and does not apply a monotonic transformation of the preference rank ordering that best satisfies an assumed composition rule. The technique was utilized in a demonstration analysis of 100 men and women (white, liberal, middle class, well educated) recruited from a Boulder, Colorado family planning clinic. among the range of possibilities, effectiveness, reversibility, side effects, and convenience were selected for expositional simplicity and frequency of mention in contraceptive literature. 3 levels of each consequence were chosen and are described, and a 3 x 3 x 3 factorial combination yielded 27 possible contraceptive methods. Individuals ranked the methods from most to least preferred as if to delay childbearing and a second time as if to stop childbearing. This was done to capture 2 stages in the family planning career. The 2 stage analysis process (proposing and testing composition rules, and deriving values) is described. The results of the analysis performed for each respondent for the 2 stages showed a clustering of responses and substantial individual differences. The most valuable delay consequence was effectiveness/reversibility, followed by side effects, and convenience. Respondents in the largest cluster (n=24) are moderately concerned about effectiveness/reversibility (median value of 60%), somewhat concerned about side effects (32%), and slightly concerned about convenience (8%). The 2nd cluster (n=18) are moderately concerned about side effects (56%), somewhat concerned about effectiveness/reversibility (33%), and slightly concerned about convenience. The 3rd cluster (n=12) are moderate concerned about effectiveness and reversibility (64%), somewhat concerned about convenience (33%), and slightly concerned about side effects (8%). The single most valued stop consequence was side effects (n=36), effectiveness/reversibility (n=25), and convenience (n=10). Delay/stop changes in values are also determined. Discussion encompasses the nature and implication of the individual differences, 2 methodological issues, and suggests an alternative methodology (functional measurement) to axiomatic conjoint measurement for assessing contraceptive values.

Use of the maximal-operant principle to motivate children's intrinsic interest.

The present research examined the hypothesis that in contrast to theory and research indicating that tangible reward decreases subsequent interest in enjoyable academic activities, rewards are perceived by adults as effective techniques to maximize long- and short-term subsequent interest for academic tasks of both high and low initial interest level. The results of our first three studies demonstrated that college students and parents view tangible reward as more effective than other less controlling techniques to enhance intrinsic motivation and value rewards more for intrinsically interesting academic behaviors in comparison with others (e.g., prosocial behaviors). Our fourth study supported the hypothesis that adults do not subscribe to the minimal-sufficiency analysis of increasing intrinsic motivation but prefer a maximal-operant principle in which the likelihood of producing long-term interest in academic tasks is assumed to vary positively with the size of a reward. Our fifth and sixth studies investigated illusory correlation as one mechanism that may perpetuate beliefs about the assumed positive relation between tangible reward and intrinsic interest in academic tasks.

Evaluation of eight methods for estimating creatinine clearance in men.

Eight methods for estimating creatinine clearance (CLcr) were compared in 65 men with serum creatinine concentrations (SCr) less than or equal to 1.5 mg/dL (group 1) and 65 men with SCr greater than 1.5 mg/dL (group 2). All patients had SCr values that did not fluctuate by more than +/- 10% for two weeks before and two weeks after measurement of CLcr. For each patient, predictions of CLcr by each of eight currently used formulas were compared with measured CLcr values; both regression analysis and predictive error analysis were used. Group 1 patients ranged in age from 32 to 64 years (mean, 53), weighed from 48 to 105 kg (mean, 73), and were from 63 to 79 inches in height (mean, 69). Group 2 patients ranged from 26 to 63 years of age (mean, 53), weighed from 34 to 141 kg (mean, 80), and were from 63 to 76 inches in height (mean, 70). Measured CLcr values ranged from 29.8 to 197 mL/min in group 1 and from 2.8 to 118 mL/min in group 2. Ranges of SCr values were 0.7-1.5 mg/dL (mean, 1.1) in group 1 and 1.6-7.1 mg/dL (mean, 2.8) in group 2; the formula of Cockcroft and Gault, which uses age, body weight, and SCr, had the highest correlation and the greatest accuracy in group 1, whereas the formula of Jelliffe, which uses body surface area and SCr, had the highest correlation and the greatest accuracy in group 2. Estimation of creatinine clearance can be improved by identification and use of the formula that is best suited to a specific patient population.

Determining the impact of sex preferences on fertility: a demonstration study.

The stopping rule measure of sex preferences represents a combination of psychological measures of preference and behavioral intentions. This study of 172 college students demonstrates that the stopping rule measure is a useful and practical method of measuring sex preferences. The results further indicate that parity progression ratio measures inherently underestimate the effect of sex preferences on individual fertility because they incorrectly assume that sex preferences (a) are homogeneous within the population and (b) can only act to increase, not to decrease, fertility. Use of the stopping rule measure to predict the possible effects of sex preselection techniques on fertility is also discussed.

Determining the impact of sex preferences on fertility: a consideration of parity progression ratio, dominance, and stopping rule measures.

The two methods commonly used to assess the effect of sex preferences on fertility are inadequate to the task. Parity progression ratio analyses suffer from logical problems stemming from the heterogeneity of sex preferences and the riskiness of fertility decisions. While conjoint measurement-dominance procedures overcome these logical problems, they cannot yield quantitative estimates of the impact of sex preferences on fertility. A stopping rule measure which overcomes these limitations is proposed and described and its potential for determining the effect of sex predetermination methods on population is discussed.

Mental search processes in problem solving.

A new approach to problem solving was applied to multisolution problems in a memory search task. Subjects memorized a list of eight four-letter foods, and then searched mentally through the list for answers to questions. The times between successive answers (IRTs) were recorded along with the answers themselves. This allowed a comparison of two possible memory search strategies: (1) sampling with replacement, and (2) sampling without replacement. The results were largely in agreement with the sampling-without-replacement strategy. However, a more detailed breakdown of the data revealed that most subjects searched through the list in a rigid serial order. Further, an analysis of questions with identical answers showed that the IRTs were very nearly additive. This led to an additive time component model based on the independent summation of (a) read-in time, (b) memory-search time, (c) decision-making time, and (d)response-output time. This approach appeared generally more satisfactory than previous attempts to account for problem-solving behavior.