Specification of Scales in Biodata Form Development: Rational vs. Empirical and Global vs. Specific

Although empirical keying has been the most popular scoring strategy with biodata, researchers have increasingly argued that rational approaches are better for advancing theory. Higher validities and less faking with empirical keys, however, have made many reluctant to abandon them. Research in the personality field provided support for the notion that many rational biodata scales may be multi-faceted. Development of more specific scales was suggested as a technique for creating rational scales with validities that more closely approximate that of empirical keys. Three different strategies (rational, internal, and external/empirical) were used in the scoring of a biodata inventory for use with mechanical equipment franchise owners. The rational and internal approaches were investigated within two samples and at two levels - very specific constructs and global constructs. In addition, two types of criteria were used, including an objective measure of sales and supervisory performance ratings. The specific rational scales were as predictive as the empirical item key. Strategies for developing and scoring a biodata form using a rational approach are discussed.     

Scoring Biodata: Is it rational to be quasi‐rational?

Five quasi‐rational biodata scales were developed by empirically keying biodata items to predict scores on a measure of the Big Five dimensions of personality. The criterion‐related validities of the quasi‐rational scales were compared to empirical and rational biodata keying methods using supervisory ratings of job performance as the criterion. Empirical keying outperformed the quasi‐rational and rational methods (which had similar validities).     

Impact of elaboration on socially desirable responding and the validity of biodata measures

The current study investigated the impact of requiring respondents to elaborate on their answers to a biodata measure on mean scores, the validity of the biodata item composites, subgroup mean differences, and correlations with social desirability. Results of this study indicate that elaborated responses result in scores that are much lower than nonelaborated responses to the same items by an independent sample. Despite the lower mean score on elaborated items, it does not appear that elaboration affects the size of the correlation between social desirability and responses to biodata items or that it affects criterion-related validity or subgroup mean differences in a practically significant way.     

Effect of Item Level Social Desirability on Factor Structure Stability

This study examined the effect of item‐level social desirability as a threat to construct validity for both a five‐factor biodata and personality inventory. Seven hundred and five participants were randomly assigned to either an Answer Honestly or Faking condition. All participants completed a biodata inventory and a NEO‐Five‐Factor Inventory. Both inventories' items were divided by low/high social desirability value (SDV) scores and grouped together. Differences between samples' covariance matrices and factor structure fit were then analyzed. It was predicted that items grouped by low SDV compared with those grouped by high SDV, would result in smaller changes in covariance matrices across conditions, and more closely retain factor structures consistent with the Five‐Factor Model. Although the results of this study supported predictions that covariance matrices changed less for low‐SDV items, overall results indicate that item‐level social desirability did not degrade the factor structure.     

Construct-oriented Biodata Development for Selection to a Differentiated Performance Domain

This study describes the development of a multidimensional biodata form which used explicit constructs to guide item generation and rational scale development, construct validation, criterion measurement and empirical keying. These constructs were goal-orientation, teamwork, customer service, resourcefulness, learning ability and leadership. Exploratory and confirmatory factor analyses in both applicant and incumbent samples were used to identify and test the model which included the thirteen, more differentiated rational scales relating to these six, broader constructs. Empirical keying of the rationally developed scales was conducted against criterion construct scales conceptually related to each predictor construct. Empirical keying at the item level was found to result in higher validities and cross-validities than either empirical keying at the scale level, or rational keying. The item keyed instrument also demonstrated incremental validity over a test of cognitive ability for specific work performance domains as well as overall work performance.     

Control of social desirability in personality inventories: Principal-factor deletion

Rational and statistical techniques for controlling social desirability (SD) in scales derived from self-report inventories are reviewed. The principal-factor deletion technique is explained in detail. It is applied during factor analysis when derived content scales are expected to be tainted with SD. If one principal factor is found to represent social desirability, the implicated factor is dropped and the communalities adjusted before the remaining factors are rotated to a desired simple structure criterion. Subscales derived from resulting factors are free of SD bias. Principal-factor deletion is compared with other approaches including rational techniques, covariate procedures, and target rotation.     

Evaluation of the psychometric properties of two forgiveness scales

This study examined the psychometric properties of two forgiveness scales using participants enrolled at a Midwestern Catholic university (N = 328). The Forgiveness Scale is a 15—item Likert-type scale designed to measure forgiveness toward an of-fender. The Forgiveness Likelihood Scale is a 10—item Likert-type scale designed to measure tendency to forgive across situations. Factor analyses revealed that the For-giveness Scale contains two subscales (i.e., Absence of Negative, Presence of Posi-tive) and the Forgiveness Likelihood Scale consists of a single factor. Both scales have adequate internal consistency and test-retest reliability. Both subscales of the Forgiveness Scale were significantly correlated in the expected direction with mea-sures of forgiveness, religiousness, anger, hope, religious well-being, existential well-being, and social desirability. The Forgiveness Likelihood Scale was significantly correlated in the expected direction with measures of forgiveness, religiousness, trait anger, religious well-being, and social desirability. The authors thank Sarah Danko, Anne Hovancsek, Carla Kmett, Jennifer Martin, and Colleen Ryan for their assistance. We kindly request that researchers who use the forgiveness scales evaluated in this study provide us with a summary of their psychometric data for the scales. The idea for creating the Forgiveness Likelihood Scale was based upon the Willingness to Forgive Scale (Hebl & Enright, 1993).     

Evaluation of the psychometric properties of two forgiveness scales

This study examined the psychometric properties of two forgiveness scales using participants enrolled at a Midwestern Catholic university (N = 328). The Forgiveness Scale is a 15—item Likert-type scale designed to measure forgiveness toward an of-fender. The Forgiveness Likelihood Scale is a 10—item Likert-type scale designed to measure tendency to forgive across situations. Factor analyses revealed that the For-giveness Scale contains two subscales (i.e., Absence of Negative, Presence of Posi-tive) and the Forgiveness Likelihood Scale consists of a single factor. Both scales have adequate internal consistency and test-retest reliability. Both subscales of the Forgiveness Scale were significantly correlated in the expected direction with mea-sures of forgiveness, religiousness, anger, hope, religious well-being, existential well-being, and social desirability. The Forgiveness Likelihood Scale was significantly correlated in the expected direction with measures of forgiveness, religiousness, trait anger, religious well-being, and social desirability. The authors thank Sarah Danko, Anne Hovancsek, Carla Kmett, Jennifer Martin, and Colleen Ryan for their assistance. We kindly request that researchers who use the forgiveness scales evaluated in this study provide us with a summary of their psychometric data for the scales. The idea for creating the Forgiveness Likelihood Scale was based upon the Willingness to Forgive Scale (Hebl & Enright, 1993).     

Interrelationships Among Name Desirability, Name Uniqueness, Emotion Characteristics Connoted by Names, and Temperament

The subjects in Study 1 (the target group) provided their names and data on their own temperaments. The subjects in Study 2 rated the target group's names on uniqueness and desirability. The subjects in Study 3 rated the temperament qualities connoted by the names given by the target group. The name uniqueness and desirability scales had high reliabilities and correlated -0.44, showing that more unique names were less desirable. The qualities of pleasantness and dominance, which were connoted by a person's name, were judged very reliably. Names that connoted psychological health versus maladjustment (pleasant/ unpleasant temperament) were indeed associated with pleasant/unpleasant temperament attributes. The pleasantness and dominance connoted by names enhanced the desirability of those names, whereas greater unpleasantness and submissiveness were connoted by more unique names.     

Interrelationships Among Name Desirability, Name Uniqueness, Emotion Characteristics Connoted by Names, and Temperament

The subjects in Study 1 (the target group) provided their names and data on their own temperaments. The subjects in Study 2 rated the target group's names on uniqueness and desirability. The subjects in Study 3 rated the temperament qualities connoted by the names given by the target group. The name uniqueness and desirability scales had high reliabilities and correlated -0.44, showing that more unique names were less desirable. The qualities of pleasantness and dominance, which were connoted by a person's name, were judged very reliably. Names that connoted psychological health versus maladjustment (pleasant/ unpleasant temperament) were indeed associated with pleasant/unpleasant temperament attributes. The pleasantness and dominance connoted by names enhanced the desirability of those names, whereas greater unpleasantness and submissiveness were connoted by more unique names.     

INCREMENTAL VALIDITY OF EMPIRICALLY KEYED BIODATA SCALES OVER GMA AND THE FIVE FACTOR PERSONALITY CONSTRUCTS

The major purpose of this study was to determine whether empirically keyed, cross-validated biodata scales accounted for incremental variance over that accounted for by the five factor model (FFM) of personality and GMA predictors. A concurrent validation study was employed using 376 employees in a clerical job (222 in the developmental sample and 154 in the cross-validation sample). Results for the cross-validation sample provided support for the hypothesis that biodata predictors accounted for substantial incremental variance beyond that accounted for by the FFM predictors and GMA for 3 of the 4 criteria. Support was also found for the hypothesized zero-order correlations between GMA, FFM, and biodata predictors and the 4 criteria. Theoretical and practical implications are discussed.     

Reducing voluntary, avoidable turnover through selection

The authors investigated the efficacy of several variables used to predict voluntary, organizationally avoidable turnover even before the employee is hired. Analyses conducted on applicant data collected in 2 separate organizations (N = 445) confirmed that biodata, clear-purpose attitudes and intentions, and disguised-purpose dispositional retention scales predicted voluntary, avoidable turnover (rs ranged from -.16 to -.22, R = .37, adjusted R = .33). Results also revealed that biodata scales and disguised-purpose retention scales added incremental validity, whereas clear-purpose retention scales did not explain significant incremental variance in turnover beyond what was explained by biodata and disguised-purpose scales. Furthermore, disparate impact (subgroup differences on race, sex, and age) was consistently small (average d = 0.12 when the majority group scored higher than the minority group).     

履历数据测评的效度分析

履历数据是人员测评的重要手段, 但其组织情景性限制使得国外同类工具不能适用于国内人事选拔, 国内也缺乏实证研究验证其信度、效度。本研究结合某通信企业的招募选拔工作, 开发履历数据分析工具, 选取250名应聘者的履历数据、一般认知能力、大五个性和面试结果数据, 通过效标关联效度、增量效度思想, 分析履历数据有效性。结果表明, 以面试结果为效标, 履历数据具有良好的效标关联效度, 与其他测评工具组合使用时有良好的增量效度。     

Validating an Inventory for the Assessment of Egoistic Bias and Moralistic Bias as Two Separable Components of Social Desirability

The psychometric properties of the German adaptation of the Balanced Inventory of Desirable Responding were investigated. This 2-factor inventory taps the egoistic bias and the moralistic bias as the 2 distinguishable components of social desirability. In 2 validation studies, both subscales showed satisfactory internal consistency and temporal stability. To examine convergent and discriminant validity, the scales were correlated with those of a measure of unrealistic optimism, various measures of self-enhancement, 1-dimensional scales of social desirability, and the tendency to overclaim in a test of general knowledge.     

Situational influences on destructive acts

Social psychological approaches to destructive behavior have stressed the importance of environmental factors, but, the range of potentially important life events has not been identified. Using a personality-based model of destructiveness, a pool of relevant biodata items were generated. A sample of 285 college students completed the situational-based biodata measure, the CPI, and two standardized integrity screening measures. A factor analysis of the biodata items resulted in seven factors: alienation, nonsupportive family, negative adult models, life stressors, competitive pressures, negative peer group and financial need. In a series of blocked regressions the situational scales produced significant increases in prediction over the personality variables on the measures of integrity, socialization and delinquency. The implications of these findings for selection and research are discussed. This research was supported by a grant from the U.S. Office of Naval Research (Contract No. 00014-0339, Michael D. Mumford and Theodore L. Gessner, Principal Investigators.     

A CONCEPTUAL RATIONALE FOR THE DOMAIN AND ATTRIBUTES OF BIODATA ITEMS

There is considerable disagreement among researchers regarding the defining attributes of biodata items, especially in contrast to other self-report measures. This paper provides a conceptual rationale for the use of biodata in order to evaluate various proposed attributes of biodata, and to clarify similarities and differences between biodata and temperament items. It is suggested that the core attribute of biodata items is that the items pertain to historical events that may have shaped the person's behavior and identity. Other attributes advanced as antidotes to socially desirable responding and faking are that biodata items reflect external events, that they be limited to objective and first-hand recollections, that they be potentially verifiable, and that they measure discrete, unique events when appropriate. Some researchers also require that the life events being sampled be controllable and equally accessible to all respondents. Items may also need to be visibly job relevant and nonintrusive to avoid legal censure. Disagreements among researchers regarding the importance of these attributes are discussed, and implications for practice and research are proposed.     

Social desirability, approval and public good contribution

Behaviour in public good experiments is usually attributed partly to rational self-interest and partly to social norms and preferences. This paper examines if sensitivity to social desirability affects public good contribution and in what way. A pre-experimental measure of social desirability (SDS17) was used to match partners in a two-person public good game. Half the participants received experimenter approval based upon their investment. Contrary to predictions, the highest public good investment was by low social desirability participants in the approval condition. Social desirability was not positively related to pro-social behaviour. We consider its relation to experimental and social conformity.     

Look at me, I'm happy and creative: The effect of impression management on behavior in social presence

The present research tested competing approaches to individual differences in impression management (as measured with social desirability scales) and their implication for behavior in social contexts. A defensiveness approach argues that impression management is a source of defensive self-presentation, which causes performance impairment in public social settings. The competing adjustment approach argues that impression management measures friendliness and self-control, which should bring about performance facilitation in public social settings. To decide between these approaches, two experiments utilized a social facilitation paradigm, whereby task performance was compared between an alone and a public condition. The results supported the predictions of the adjustment approach. Across different tasks, a high impression management score was associated with performance facilitation in social presence, expressed in greater creativity, positive implicit affect, and high self-control. The results reveal previously unnoticed constructive effects of impression management, supporting the reframing of the trait as reflecting interpersonally oriented self-control.     

BEYOND THE MEAN BIAS: THE EFFECT OF WARNING AGAINST FAKING ON BIODATA ITEM VARIANCES

We studied the effects of faking biodata test items by randomly warning 214 of 429 applicants for a nurse's assistant position against faking. While the warning mitigated the propensity to fake, the specific warning effects depended on item transparency. For transparent items, warning reduced the extremeness of item means and increased item variances. For nontransparent items, warning did not have an effect on item means and reduced item variances. These faking effects were best predicted when transparency was operationalized in terms of item-specific job desirability in addition to the item-general social desirability. We also demonstrated a psychometric principle: The effect of warning on means at the item level is preserved in scales constructed from those items, but the effect on variances at the item level is masked at the scale level. These results raise new questions regarding the attenuating effects of faking on validity, and regarding the benefit of warning applicants against faking.     

ADDITIONAL EVIDENCE FOR THE ACCURACY OF BIOGRAPHICAL DATA: LONG-TERM RETEST AND OBSERVER RATINGS

Although the predictive validity of biographical information has been extensively documented in a variety of environments, the extent to which responses to a biodata questionnaire are accurate has not been adequately investigated. In particular, the accuracy of unverifi-able biodata items has not been determined. Evidence of accuracy was obtained in the present study using a test-retest design and informed external observers for verification. Responses were obtained from 237 study participants and 200 observers. Factor and item re-test correlations were obtained for subject and observer data, and responses were compared using product-moment correlations and t tests. Results provide evidence that many responses to a biodata questionnaire are accurate. Sources of inaccuracy were assessed, including the impact of social desirability, and applications of the findings to industrial uses of biodata were noted.