机器理解古汉语的实验和若干问题

本文介绍了机器理解古汉语的一个实验系统(ACLUS)。该系统由一个知识库和三个子程序组成。知识库包含有一部机器词典和若干集用来加工句子的规则。三个子程序是:句法分析子程序、翻译子程序和问答子程序,它们分别执行三种不同的功能。本文还讨论了有关机器理解古汉语的若干重要问题。     

THE IMPACT OF JOB ANALYSIS ON EMPLOYMENT TEST VALIDATION FOR MINORITY AND NONMINORITY ACCOUNTING PERSONNEL

The purposes of the present study were (a) to examine the comparative validity of a written job knowledge test constructed on the basis of a systematic job analysis with that of a commercial employment test selected in the absence of a prior job analysis for accounting positions and (b) to determine the fairness of each test for minority and nonminority job applicants. Results indicated that the job knowledge test was a valid and unbiased predictor of relevant criteria of job performance while the commercial employment test produced adverse impact and lacked validity. Implications of the results for future research studies and test validation efforts involving differential prediction are discussed.     

THE FEMALE SEXUAL SUBJECTIVITY INVENTORY: DEVELOPMENT AND VALIDATION OF A MULTIDIMENSIONAL INVENTORY FOR LATE ADOLESCENTS AND EMERGING ADULTS

Three studies were conducted to develop and validate a theoretically derived multidimensional inventory of females' sexual self-conceptions ( sexual subjectivity ). Study 1 revealed five factors on the Female Sexual Subjectivity Inventory (FSSI): sexual body-esteem, three factors of conceptions and expectations of sexual desire and pleasure (self, partner, and self-efficacy), and sexual self-reflection. A shorter revised version of the FSSI was tested in Study 2. In Study 3, a confirmatory factor analysis indicated a good fit to the data. The FSSI had a sound factorial structure and high reliability. Significant associations between the five scales in the FSSI and sexual self-awareness, safe sex self-efficacy, and sexual anxiety provided evidence of validity. Some FSSI scales were also associated with self-silencing in close relationships, resistance to sexual double standards, and self-esteem. Sexual subjectivity is a complex intraindividual phenomenon that includes cognitive and emotional components, some of which can be assessed with the FSSI scales.     

HOME SAFETY: DEVELOPMENT AND VALIDATION OF ONE COMPONENT OF AN ECOBEHAVIORAL TREATMENT PROGRAM FOR ABUSED AND NEGLECTED CHILDREN

Parents may be charged with child abuse or neglect or both on the basis of a variety of circumstances. Child neglect, for example, is often documented when caseworkers observe that the family's home itself is so poorly kept that it presents an environment in which young children have ready access to lethal hazards such as poisons, uncovered wall outlets, and firearms. In this study, we describe the development of a Home Accident Prevention Inventory (HAPI) which was validated and used to assess hazards in the homes of several families under state protective service for child abuse and neglect. The HAPI included five categories of hazards: fire and electrical, mechanical-suffocation, ingested object suffocation, firearms, and solid/liquid poisons. Following the collection of baseline data, parents were presented with a treatment package that included instructions and demonstrations on making hazards inaccessible to children, plus feedback regarding the number and location of hazards in the home. The multiple-baseline design across hazardous categories in each family's home showed that the package resulted in decreases in the number of these accessible hazards. These improvements were maintained over an extended period of unannounced follow-up checks. This research provides a model for the development and assessment of an area previously unexamined in the child abuse and neglect literature.     

DEVELOPMENT OF A JOB ANALYSIS-BASED PROCEDURE FOR WEIGHTING AND COMBINING CONTENT RELATED TESTS INTO A SINGLE TEST BATTERY SCORE

By definition, content-related approaches to test validation do not rely on criterion data. As a consequence, regression and other statistical procedures for weighting and generating a composite score from a test battery are not applicable when a content-related validation strategy is used. This paper presents a procedure for determining the component weights for a test battery that has been developed on the basis of a content-related validity strategy. The Relative Content Contribution (RCC) weighting procedure is a logical extension of the conceptual basis underlying the rational developmental process used to demonstrate the validity of content-related tests. Results from field implementations of the procedure in the development of two promotional test batteries (fire and police) and an entry-level test battery (police) in two large metropolitan cities are presented to illustrate the procedure.     

TOWARD THE DEVELOPMENT OF STRUCTURED CRITERIA FOR INTERPRETATION OF FUNCTIONAL ANALYSIS DATA

Using functional analysis results to prescribe treatments is the preferred method for developing behavioral interventions. Little is known, however, about the reliability and validity of visual inspection for the interpretation of functional analysis data. The purpose of this investigation was to develop a set of structured criteria for visual inspection of multielement functional analyses that, when applied correctly, would increase interrater agreement and agreement with interpretations reached by expert consensus. In Study 1, 3 predoctoral interns interpreted functional analysis graphs, and interrater agreement was low (M = .46). In Study 2, 64 functional analysis graphs were interpreted by a panel of experts, and then a set of structured criteria were developed that yielded interpretive results similar to those of the panel (exact agreement = .94). In Study 3, the 3 predoctoral interns from Study 1 were trained to use the structured criteria, and the mean interrater agreement coefficient increased to .81. The results suggest that (a) the interpretation of functional analysis data may be less reliable than is generally assumed, (b) decision-making rules used by experts in the interpretation of functional analysis data can be operationalized, and (c) individuals can be trained to apply these rules accurately to increase interrater agreement. Potential uses of the criteria are discussed.     

A META-ANALYSIS OF THE VALIDITY OF METHODS FOR RATING TRAINING AND EXPERIENCE IN PERSONNEL SELECTION

This paper reviews the validity evidence for systematic methods used to evaluate training and experience (T&E) ratings in personnel selection. Meta-analytic summaries of the data indicate that validity varies with the type of T&E evaluation procedure used. The lllinois job element and behavioral consistency methods each demonstrated useful levels of validity (.20 and .45, respectively) with small corrected standard deviations, thus supporting validity generalization. Both the point and task methods yielded low mean validities (.11 and .15, respectively) with larger variability. The authors hypothesized that both the point and task methods were affected by a job experience moderator. Partial support for this hypothesis was found. Moderator analyses suggested that the point method was most valid when the applicant pool had low mean levels of job experience and was least valid with an experienced applicant pool. Additional research is desirable on all T&E methods to decrease the potential impact of second-order sampling error in the meta-analytic results. Further research is also needed to explicate the constructs measured by T&E evaluations.     

PROJECT A: SPECIFICATION OF THE PREDICTOR DOMAIN AND DEVELOPMENT OF NEW SELECTION/ CLASSIFICATION TESTS

A major goal of the Army Selection and Classification Project was to develop an experimental predictor battery that would best supplement the Armed Forces Vocational Aptitude Battery for making selection and classification decisions for entry-level enlisted personnel. That is, what predictor measures would best serve the needs of all the jobs in an entire selection/classification system? This paper describes the characteristics of the new test battery and the procedures that were used to develop it. The major steps in the procedure were a structured literature search using a standard protocol, an extensive expert judgment study of expected true validities for a population of predictor variables against a population of performance components, fabrication of modularized software and a special response pedestal for computerized measurement of perceptual and psychomotor abilities, evaluations of experimental measures in three iterative pilot tests and one major field test, and a series of reviews by a panel of scientific advisers. The test battery that resulted from this 2 1/2-year development effort is described. The basic psychometric properties of each measure, as determined in a large concurrent validation sample, are also described.     

HONESTY TESTING FOR PERSONNEL SELECTION: A REVIEW AND CRITIQUE

Paper and pencil predictors of employee theft are described and studies of validity, reliability, and adverse impact of these tests are examined. Validity studies for 10 tests were grouped into 5 categories: comparisons with polygraph examination results, correlations with admissions of past theft, predictive studies using future job behaviors as criteria, comparisons of shrinkage rates before and after the introduction of a testing program, and comparisons of test scores of groups known to be dishonest with groups representing the general population. While positive correlations were consistently found, a variety of methodological differences between studies were identified which make the direct comparison of test validities suspect. High reliabilities are consistently reported, and test score comparisons by race and sex generally report no differences. Ethical issues in honesty test usage are considered and future research needs are identified.     

A COMPARISON OF CRITERIA FOR TEST VALIDATION: A META-ANALYTIC INVESTIGATION

Meta-analyses of validity coefficients from tests of clerical abilities for five criteria—supervisor ratings, supervisor rankings, work samples, production quantity, and production quality—were conducted, and the resulting expected true validities were compared. Ratings, rankings, work samples, and production quantity all resulted in high test validities. Validities resulting from ratings and quantity-of-production criteria were highly similar across tests. Validities resulting from rankings and work samples were on the average higher than those from ratings and quantity of production. The fifth criterion, quality of production, had low predictability and did not generalize across situations.     

FURTHER WITHIN-SETTING EMPIRICAL TESTS OF THE SITUATIONAL SPECIFICITY HYPOTHESIS IN PERSONNEL SELECTION

The situational specificity hypothesis in personnel selection holds that variation in observed validity coefficients across studies for the same test and job is due to subtle variations from setting to setting in what constitutes job performance. This hypothesis therefore predicts that, if the setting does not vary, validity will not vary. Using data from a single large-sample validity study ( N = 1,455), this research generated numerous small-sample studies for which the setting (organization, job, test, criterion measure, applicant pool, time period, and sample size) was held constant. It was found that even under these circumstances there was substantial variability across studies in (a) observed validity coefficients, (b) significance levels, and (c) (using traditional data analytic methods) conclusions about the presence or absence of validity. These findings disconfirm the situational specificity hypothesis and argue strongly against traditional data-analytic procedures and the practice of reliance on single small-sample studies. In contrast to the erroneous conclusions produced by traditional data-analytic procedures, meta-analytic methods correctly estimated the population observed validity at .22 and correctly indicated that all between-study variance in observed validities was due to sampling error alone.