与零假设检验有关的统计分析流程

人们熟知的零假设显著性检验,受到一次次质疑与辩护,地位并未动摇,报告检验结果仍然是统计分析的习惯做法。不过,其局限性促使研究者探寻更多的统计方法如区间估计、效应量分析、检验力分析等。本文先介绍假设检验与置信区间的关系;然后讨论检验力与两类错误率和效应量的关系;最后在理顺上述统计方法的基础上,提供一个可操作的统计分析流程。     

Attenuating memory distrust in a repeated checking task

Repeated checking has been demonstrated to lead to reductions in memory confidence in several previous studies using student and clinical samples. This process of reduced confidence in memory and detail for memory, are thought to arise from the inhibition of perceptual processing that develops during repeated checking. Our research investigated whether reduced memory confidence from repeated checking could be attenuated through the use of novel stimuli during the repeated checking task. Three groups were generated through random assignment of 65 undergraduate students. As seen in previous research, individuals who repeatedly checked a stimulus (a virtual stovetop) showed reduced memory confidence, vividness, and detail, when compared with individuals who repeatedly checked a different stimulus. A third group in which the colour of the repeatedly-checked stovetop changed every five trials showed no significant decline in memory confidence between the pre-test and post-test. Results suggest that increased memory distrust can be ameliorated through the use of stimuli with characteristics that are novel and distinctive. Findings are discussed in the context of the existing model of repeated checking and memory confidence, and implications for treatment methods are presented.     

Reminders supporting spontaneous remembering in prospective memory tasks1

Abstract: The objective of this study was to identify the reminders that facilitate prospective memory. A total of 64 participants were either provided with three types of reminders for a prospective memory task (to circle a question number when prescribed target words appear) during the retention interval, or not reminded of the task. In addition, an articulatory suppression task was imposed halfway through the test phase to examine whether verbal rehearsal is related to the effects of the reminders. The reminders that referred to the intended activity (to circle the number) did not improve prospective memory. However, both of the reminders that notified that there was something to do and the reminders that referred to the target events (the prescribed words) improved prospective memory. Imposing the suppression task abolished the benefit of the former, but not that of the latter. This result suggests that only the effect of the reminders that referred to the existence of intention is mediated by verbal rehearsal. These data imply that there are some differences in the functions of these reminders.     

Bias in Estimation of Misclassification Rates

When a simple random sample of size n is employed to establish a classification rule for prediction of a polytomous variable by an independent variable, the best achievable rate of misclassification is higher than the corresponding best achievable rate if the conditional probability distribution is known for the predicted variable given the independent variable. In typical cases, this increased misclassification rate due to sampling is remarkably small relative to other increases in expected measures of prediction accuracy due to samplings that are typically encountered in statistical analysis. This issue is particularly striking if a polytomous variable predicts a polytomous variable, for the excess misclassification rate due to estimation approaches 0 at an exponential rate as n increases. Even with a continuous real predictor and with simple nonparametric methods, it is typically not difficult to achieve an excess misclassification rate on the order of n ⁻¹. Although reduced excess error is normally desirable, it may reasonably be argued that, in the case of classification, the reduction in bias is related to a more fundamental lack of sensitivity of misclassification error to the quality of the prediction. This lack of sensitivity is not an issue if criteria based on probability prediction such as logarithmic penalty or least squares are employed, but the latter measures typically involve more substantial issues of bias. With polytomous predictors, excess expected errors due to sampling are typically of order n ⁻¹. For a continuous real predictor, the increase in expected error is typically of order n ^−2/3     

Examining the trade‐off between confidence and optimism in future forecasts

Confident business forecasters are seen as more credible and competent (“confidence heuristic”). We explored a boundary condition of this effect by examining how individuals react to the trade‐off between confidence and optimism. Using hypothetical scenarios, we examined this trade‐off from the perspectives of judges (i.e., business owners who hired analysts to make sales predictions) and forecasters (i.e., the analysts hired to make predictions). Participants were assigned to the role of either judges or forecasters and were asked to rate 2 potential forecasts. In the “no trade‐off” condition, the 2 forecasts were aligned in optimism and confidence (the more confident forecast was also more optimistic); in the “trade‐off” condition, the more confident forecast was less optimistic. In Experiment 1, judges were more likely to positively evaluate confident forecasters when confident forecasters were the more (vs. less) optimistic ones. Experiment 2 demonstrated that forecasters were aware of judges' preferences for optimism and strategically relied on methods that resulted in more optimistic (but less reliable) predictions. Experiment 3 directly compared the perspectives of judges and forecasters, revealing that forecasters overestimated judges' preferences for optimism over confidence. The present studies show that forecasters and judges have different views of the trade‐off between confidence and optimism and that forecasters may unnecessarily sacrifice accuracy for optimism.     

Forensic voice discrimination by lay listeners: The effect of speech type and background noise on performance

In forensic settings, lay (nonexpert) listeners may be required to compare voice samples for identity. In two experiments we investigated the effect of background noise and variations in speaking style on performance. In each trial, participants heard two recordings, responded whether the voices belonged to the same person, and provided a confidence rating. In Experiment 1, the first recording featured read speech and the second featured read or spontaneous speech. Both recordings were presented in quiet, or with background noise. Accuracy was highest when recordings featured the same speaking style. In Experiment 2, background noise either occurred in the first or second recording. Accuracy was higher when it occurred in the second. The overall results reveal that both speaking style and background noise can disrupt accuracy. Although there is a relationship between confidence and accuracy in all conditions, it is variable. The forensic implications of these findings are discussed.     

Professional Autonomy in Belgium

The Belgian health care system has a few features that may havecontributed to the rising costs of health care: patients' freechoice of physicians, large clinical freedom of physicians, essentiallya fee-for-service remuneration for medical specialists in which the feesare agreed between insurance funds and physicians. The increased medicalconsumption and costs have prompted the state and insurance companies totake measures that limit the professional autonomy of the physicians.Access to medical education, free until 1997, is now restricted. Themedical profession is organized in the Order of Physicians that hasestablished a code of professional ethics that has moral but not legalforce. So far, there is no special legislation for thepatient–physician relationship, though laws on specific issueslike organ transplantation contain duties for physicians. In recentyears a debate is taking place on patients' rights, of which informedconsent is central and gaining importance in medico-legal publications.An analysis of (ethical and legal) regulations concerning thewithholding or withdrawal of treatment by physicians demonstrate thatthe profession still enjoys a large clinical autonomy, though duediscussion with the patient has become more explicitly required. Therespect for professional autonomy is not primarily due to any formalpower that the Order of Physicians would have, but is rather grounded inthe generally high quality of the patient–physician relationshipthat in ethical terms is considered essentially as a confidencerelationship rather than a contractual relationship.     

Exact distributions of intraclass correlation and Cronbach's alpha with Gaussian data and general covariance

Intraclass correlation and Cronbach's alpha are widely used to describe reliability of tests and measurements. Even with Gaussian data, exact distributions are known only for compound symmetric covariance (equal variances and equal correlations). Recently, large sample Gaussian approximations were derived for the distribution functions. New exact results allow calculating the exact distribution function and other properties of intraclass correlation and Cronbach's alpha, for Gaussian data with any covariance pattern, not just compound symmetry. Probabilities are computed in terms of the distribution function of a weighted sum of independent chi-square random variables. NewF approximations for the distribution functions of intraclass correlation and Cronbach's alpha are much simpler and faster to compute than the exact forms. Assuming the covariance matrix is known, the approximations typically provide sufficient accuracy, even with as few as ten observations. Either the exact or approximate distributions may be used to create confidence intervals around an estimate of reliability. Monte Carlo simulations led to a number of conclusions. Correctly assuming that the covariance matrix is compound symmetric leads to accurate confidence intervals, as was expected from previously known results. However, assuming and estimating a general covariance matrix produces somewhat optimistically narrow confidence intervals with 10 observations. Increasing sample size to 100 gives essentially unbiased coverage. Incorrectly assuming compound symmetry leads to pessimistically large confidence intervals, with pessimism increasing with sample size. In contrast, incorrectly assuming general covariance introduces only a modest optimistic bias in small samples. Hence the new methods seem preferable for creating confidence intervals, except when compound symmetry definitely holds. An earlier version of this paper was submitted in partial fulfillment of the requirements for the M.S. in Biostatistics, and also summarized in a presentation at the meetings of the Eastern North American Region of the International Biometric Society in March, 2001. Kistner's work was supported in part by NIEHS training grant ES07018-24 and NCI program project grant P01 CA47 982-04. She gratefully acknowledges the inspiration of A. Calandra's “Scoring formulas and probability considerations” (Psychometrika, 6, 1–9). Muller's work supported in part by NCI program project grant P01 CA47 982-04.     

An a priori solution to the replication crisis

Possibly, the replication crisis constitutes the most important problem in psychology. It calls into question whether psychology is a science. Existing conceptualizations of replicability depend on effect sizes; the larger the population effect size, the greater the probability of replication. This is problematic and contributes to the replication crisis. A different conceptualization, not dependent on population effect sizes, is desirable. The proposed solution features the closeness of sample means to their corresponding population means, in both the original and replication experiments. If the researcher has specified the sampling precision desired, it is possible to calculate the probability of replication, prior to data collection, and without dependence on the population effect size or expected population effect size. In addition, it is not necessary to know population means or standard deviations, nor sample means or standard deviations, to employ the proposed a priori way of thinking about replicability.