对于胜任力感到迷惑？一份胜任力模型的进化和应用评估 毕业论文外文翻译

对于胜任力感到迷惑？一份胜任力模型的进化和应用评估 毕业论文外文翻译 外文文献: Confounded by Competencies? An Evaluation of the Evolution and Use of Competency Models Leanne H. Markus Performance Group International Ltd., Auckland Helena D. Cooper-Thomas Department of Psychology, University of Auckland Keith N. Allpress Centranum Ltd., Auckland Over the last ten years there has been a world-wide expansion in the use of competency models as a major underpinning of Human Resources (HR) strategy. The use of the competency approach is promoted by consultants and software vendors on the basis that this will improve both individual job performance and organizational effectiveness. Yet there is a substantial, and largely unquestioned, gap between the many claims and the actual benefits measurably delivered by competency initiatives. Industrial and Organizational (I/O) Psychologists are often involved in developing and implementing competency models, yet there is little research validating the approach. As scientist-practitioners, we should be concerned about this. This article will review the theoretical perspectives that have informed the competency movement, review our experience of the use of competency models in New Zealand, and critically examine the assumptions that underpin their use. The research that exists is reviewed with particular reference to the outcome measures used to substantiate the value of competency models. Finally we identify various research areas and questions that should clearly be investigated by I/O psychologists if they are to be involved in the promotion, development and implementation of competency models in an organizational setting. What is a competency? Three main approaches The numerous published definitions can be grouped into three distinct approaches: educational standards, behavioral repertoires, and organizational competencies. 1. The Educational Approach (The development of skills, achievement of standards, award of credentials) The modem competency movement originated from the educational discipline. In the US 1 'competencies' were based on functional role analysis and described either role outcomes, or knowledge, skills and attitudes, or both, required for role performance, and assessed by a criterion, usually a behavioural standard. In the UK, industry bodies especially those requiring trades and technical skills, developed standards of occupational competence based on expected work outcomes (Fletcher, 1992). A 'competence' was defined narrowly as an action, behaviour or outcome to be demonstrated, or a minimum standard, with different levels of mastery defined by different statements (Bourke et al., 1975, Elam, 1971). 2. The Psychological Approach- (Behavioural repertoires) In 1973 David McClelland, working in the educational field in the US, wrote a paper suggesting that personal competencies, which he defined as motives and personality traits, are a better means of predicting occupational success than traditional psychometrics such as IQ and aptitude tests. McClelland's work was to be enormously influential. Of particular interest was the idea that the factors or inputs associated with individual success could be identified, and then taught to others. McClelland and Boyatzis (1980) developed a methodology for identifying competencies, based on the skilled behavioural repertoires of recognised star performers within particular organisations. They defined competencies as "a generic body of knowledge, motives, traits, self images and social roles and skills that are causally related to superior or effective performance in the job." (p.369, italics added). 3. The Business Approach (Organisational competencies for competitive advantage) The concept of competencies was taken up by business strategists in the late 1980s. Hamel and Prahalad (1989) advanced the idea of "Core Competencies" and "Capabilities". Their definition of core competencies as the "collective learning" of the organization has been much cited, and contributes to the current interest in "competencies" (Shipmann et al., 2000). Thus Sparrow (1995) suggests that practitioners should aim at defining "higher level" future oriented organisational competencies. What are the potential benefits of the Competency approach? Performance benefits are promised by the various definitions which include the causal or instrumental relationship of competencies and job performance (Boyatzis 1982) and competencies and organisational performance. (Organ, 1988; Hamel & Prahalad, 1989). In addition, Sparrow (1995) has observed that the competency literature includes a huge 2 range of claimed benefits specific to HR processes in organisations. In summary, these are: • improved recruitment and selection practices through a focus on required competencies; • improved individual, organisational and career development programmes; • improved performance management processes due to improved assessment; and lastly • improved communication on strategic and HR issues through a common language. What is a competency model? Organizations adopting a competency approach must create or utilize a competency model, at minimum a simple list or catalogue, specifying desirable competencies. The structure of this model must support the use of competencies across the selected HR functions. Models designed for selection and educational purposes usually describe technical competencies in terms of their antecedent skills and knowledge, at a detailed level. Those designed to promulgate behavioural repertoires and citizenship behaviours or organizational competencies typically describe competencies at a much higher level. Regardless of approach, a competency model should provide an operational definition for each competency and subcompetency, together with measurable or observable performance indicators or standards against which to evaluate individuals. How do competencies link to other constructs used in I/O Psychology? As pointed out by Shippmann et al. (2000) competency modelling is a huge trend in HR. While job analysis focuses mainly at the individual level, examining the specific knowledge, skills, abilities and other attributes required for individual job performance, much competency modelling represents an attempt to identify dimensions of performance applicable to many different roles and situations. Relevant to this is the extensive literature in I/O psychology representing many decades of research into factors associated with both job performance and organizational effectiveness. O’ Reilly and Chatman (1986) suggested that two distinct variables relate to job performance; firstly the in-role behaviours required in the job, and secondly prosocial behaviours which are not specifically prescribed in a particular role. Brief and Motowidio (1986) identified 13 aspects of prosocial organisational behaviour (POB) defined as behaviours aimed at promoting the welfare of other individuals or groups within the organisation. Prosocial behaviour is hypothesised to improve communications, job and customer satisfaction, and therefore 3 organizational effectiveness. Aspects of both in-role and prosocial behaviours may be included within competency models. Organisational Citizenship Behaviour (OCB) is a similar construct which has spawned a considerable literature. Organ (1988) defined OCBs as individual behaviours, beyond that required in the role or job description, which, in the aggregate, contribute to organizational effectiveness. Similar to O’ Reilly and Chatman's model (1986), Motowidio et al. (1997) have identified two elements of overall job performance; task performance, and contextual performance. The latter is essentially the socialisation, application and effort required to facilitate task performance, and is equivalent to OCB (Organ, 1997). Motowidio et al. suggest that the activities involved in task performance are most likely to vary between roles, while those involved in contextual performance are often similar. Further, they propose that the antecedents or predictors of task performance are more likely to involve cognitive ability, while personality is more likely to affect contextual performance. Task performance includes the application of technical and task knowledge, and task habits, defined as characteristic responses to task situations (Borman et al 2001). Contextual performance includes behaviours and traits such as persistence and effort, volunteering, helping and cooperation, loyalty, policy and procedural compliance, endorsement and promotion of organisational objectives, initiative and self development (Borman & Motowidio 1997). The many generic competency models and catalogues emphasise aspects of contextual rather than task performance. Personality has also been related to job performance. The Five Factor Model of Personality includes a multidimensional factor of Conscientiousness, which describes aspects of effort and application (Anastasi 1997). This factor has been found to correlate with contextual performance, particularly in the aspect of Job-Task Conscientiousness, with overall performance across a wide range of jobs, (Tett and Burnett 2003), and with career advancement (Viswesvaran & Ones 2000). This raises the question of whether investment in extensive competency models, addressing primarily contextual performance, provides any incremental utility. The literature on organizational commitment distinguishes between three types of commitment, attitudinal commitment; belief in the organisation, instrumental commitment, given on the basis of perceived costs and benefits, and normative commitment, the result of socialisation procedures. (Mathieu & Zajac, 1990). This is relevant to the use of competency models to promote and reward behaviours which exemplify desired organisational values and core competencies. 4 Perceived organizational support (POS) is the extent to which employees believe that they are valued by the organisation. It is related to organisational commitment, and job performance. Rhoades and Eisenberger (2002) found fairness to be the most important factor in POS, followed by supervisor support. Colquitt et al (2001) summarise the elements of organisational justice as consistency of treatment between individuals and over time, the absence of bias, the accuracy of information, conformance to current ethical standards, a voice for affected individuals and groups, and a mechanism to review and correct flawed decisions. It is therefore important that competency models used for assessment and performance appraisal purposes are perceived by employees as fair. How should competency models be implemented? In an attempt to quantify the quality of competency models, Shippman et al. (2000) proposed a 10 point level of rigor scale in establishing competency taxonomies from job or competency analysis. This covered effective data collection methods, competency descriptor development procedures and quality requirements, links to business strategy, validation procedures, and documentation. Attempts at model definition often canvas ideas from the wider organisation in order to create buy in. Since there are potentially many ways of defining and phrasing competencies, this can lead to a long drawn out costly process, with results subject to the Abilene effect - you get what no-one disagreed with, not necessarily the best definitions. The alternative, buying an off-the-shelf system, is likely to be cheaper up front, but may require ongoing effort from users to adapt it to fit their situation. Either way, once implemented, the competency assessment process carries a significant administrative burden, and organisations need to be assured that such investments are worthwhile. Yet there are major validity issues with the use of competency models, and as yet little evidence to support their claimed benefits. Issues with the Competency approach 1. Construct validity- What is a competency- can a competency be operationalised so that it can be observed and measured? The aim of construct validity is to assess whether a measure of an individual trait or characteristic actually measures what it is meant to. As with many psychological constructs, 5 there is no real world aspect of competencies (Cronbach & Meehl, 1955). Establishing construct validity therefore requires finding a suitable proxy criterion of the construct (Ghiselli, Campbell, & Zedeck, 1981). This is clearly a problem for competencies, with a number of studies documenting disagreement between managers, staff and even experts asked to categorise operational definitions of competencies, usually example behaviours (American Management Association, 2003; Horton et al., 2002). This also illustrates the difficulty in using competency models to communicate, promote and reward organisational norms. 1a. Content and Face validity - are competencies credible in organisations? A major problem with the use of competency models is the lack of agreement on what is meant by the term 'competency'. Sparrow (1995) has suggested that the interchangeable use of the various competency approaches encourages organisations to "build and integrate HRM systems on a bed of shifting sand" (p. 168). So what should competency models comprise - that is, how can we be sure of their content and face validity? Content validity means that the descriptors of competencies are a representative sample of the universe of interest. Face validity means that the competencies themselves feel accurate and appropriate, as judged by their users. For any particular competency model, content and face validity are essentially subjective judgements. All those to whose roles they are applied are in a position to judge whether the competencies match their role. Thus content and face validity issues may arise despite the use of subject matter experts and regardless of how systematic information gathering methods may be. Hayes et al. (2000) argue that it may be impossible to break down a competency into an exhaustive list of elements. This suggests that competency models will always be incomplete. They cite examples of studies where managers have not been able to describe all the competencies required for a role. Certainly the few behavioural statements in most generic competency models could not be regarded as exhaustive. Associated with this is the difficulty in arriving at a suitable structure for a competency model. In defining competencies, Stuart (1983) highlights the trade off between universality and fidelity problem, and between complexity and simplicity (see Table specificity, the bandwidth – 1). Universal or generic competencies are those which are applicable across roles and organisations, whereas specific competencies are those particular to roles and organisations. Universal or generic competencies run the risk of being so broadly defined that they are not perceived by individuals as relevant. Competencies can be defined simply, as a headline plus a 6 few sample behaviours, or they can be designed to cater for multiple levels of detail and mastery. While complex models permit more accurate communication of requirements, and evaluation, they can become an administrative burden. In practice, the universal or generic approach is the most commonly adopted. As an example Tett et al. (2000) attempted to identify and validate a "hyper dimensional taxonomy of managerial competence". The many proprietary generic competency catalogues, also corresponding to Stuart's (1983) universal competencies, typically include management and OCB factors. Many of these competencies are so broadly defined that they subsume a mix of personality factors, motivation and cognitive abilities (Bartram, 2004). In a recent international survey, of the 28% of larger firms which had a competency catalogue, almost half were using proprietary software with generic competency libraries (Metagroup, 2004). Yet the one size fits all approach of generic competencies is unlikely to be appropriate for organisations working with different settings, different products and different customers (Chiabaru, 2000). As Stuart (1983) suggested, the more simple and the more universal the competency model, the less the perceived relevance at the individual level. 1b. Criterion validity – Can competencies be accurately measured? An associated issue is that many competencies, especially those related to contextual performance, are defined in very broad terms, and with few performance indicators. In these cases it is unlikely that accurate evaluation is possible. This has implications for perceptions of organizational justice (Colquitt et al 2001), and perceived organisational support (Rhoades & Eisemberger 2002) which impact employee commitment to the organisation. Apart from the inadequacy of measurement criteria, competencies are usually evaluated using self and supervisor ratings, and sometimes by peers as well. Thus, the assessment of competencies is likely to suffer from all the same reliability problems, such as rater bias, that the extensive literature records for performance appraisal in general (Fletcher, 2001). Yet accurate measurement of competencies is a key issue, especially when evaluations are used in pay for performance schemes. A major pre-occupation of organisations is to accurately discriminate between different levels of success in order to ensure that "top talent" feels valued and is rewarded appropriately. For example, Hunter et al. (1990) found that in complex roles such as professional services, individual output can vary by a factor as much as 12 to 1 between best and worst performers. Table 1. An illustration of competency modelling options using Stuart's (1983) framework 7 Definitions Simple Complex Quadrant 1 Quadrant 2 , High level , High level , User-friendly , Complex structure, , May not be perceived as potentially difficult to use relevant , More relevant to role Universal , Allows comparison , Allows comparison across across roles roles , Evaluation less accurate , Evaluation more accurate , Off-the-shelf purchase , Customised solution Competencies possible Quadrant 3 Quadrant 4 , Role specific/technical , Role specific/technical , User-friendly , Multilevel, complex , No comparison across structure, difficult to use Specific roles/organisation , No comparison across , Evaluation accurate roles/organisation , Customised solution , Evaluation accurate , Customised solution 2. Validation of the competency model A second major issue is the way that organisations have implemented competency models; that is adoption without validation (Shippman et al., 2000). Validation is important because competencies describe normative behaviours, behaviours the organization wishes to promote and develop to enhance organizational effectiveness. 3. Predictive validity – do improved competencies predict improved individual Job performance and/or improved organisational performance? The third and major issue is the lack of evidence for benefits that result from adopting a competency approach. The underlying assumption of all competency initiatives is that individual skill development, exemplified by particular behaviours, will lead to improved job performance 8 and, in turn, organisational performance. Barrett and Depinet's (1991) review of the research into competency measurement provided little empirical support for McClelland's (1980) claim that competencies are better predictors of job performance than traditional psychometric tests of mental ability. Later Laber and O'Connor (2000) highlighted the lack of empirical research into the effectiveness of competency models. Our search of the literature reveals only a handful of studies investigating the link between competencies and objective job performance, leaving the situation largely unchanged four years later. One of the many criticisms of the competency movement is the implicit confusion of competency and job performance. We suggest that this confusion has arisen through the language associated with the competency movement; the confusion of behaviours, knowledge and personality traits which are inputs to the job, with results or outcomes, objective job performance. Thus competencies are assessed by "performance" of behaviours deemed to be criteria of competence/competency. Illustrating this confusion, Campbell (cited in Bartram, 2004, p.5) states "Performance is behavior. It is something that people do and is reflected in the actions that people take... Performance is not the consequence(s) or result(s) of action; it is the action itself (see also Hackett, 2002). This confusion of terminology has created an inherent circularity in the use of competency models. Competencies are identified using a variety of information gathering methods, behavioural criteria are defined, and then in the absence of objective measures of job outputs or performance, the subjective evaluation of the occurrence of these behaviours is assumed to equate to (job) performance and validate the competency construct itself. As an illustration, Mayer (2003) reports on a study examining whether health workforce competencies are predictive of essential service performance. He measured the relationship of self-assessed core competency levels and self-assessed service performance, defined as frequency of performance of public health Job tasks, at a US metropolitan health department. He found that competency level had only a very modest association with what are essentially service performance inputs. Hunter and Schmidt (1996) point out that there is little correlation between OCBs and objective output-based measures of individual job performance. However when supervisors estimate job performance, there is a high correlation between ratings of OCBs and their subjective ratings of overall job performance. They suggest this is because supervisors tolerate poor task performance in people with high levels of OCBs. Thus any association of generic 9 competencies with job performance may be due to the weighting given by supervisors to OCBs (Johnson, 2001). 10 外文文献译文: 对于胜任力感到迷惑,一份胜任力模型的进化和应用评估 琳恩 H.马库斯 海伦娜 D.库珀-托马斯 凯斯 N.奥普瑞斯 在最近的十年间，作为人力资源策略的主要基础的胜任力模型的应用正在世界范围上扩张。咨询公司和软件供应商在提高个人工作绩效和组织有效性的基础上促进了胜任力方法的应用。然而在大量的要求和由胜任力主动带来的可预测的实际利益之间有着本质的、很大程度上毋庸置疑的空白。工业和组织心理学家经常参与开发和执行胜任力模型，但是很少有研究能验证这些方法。作为科学实践者，我们应该关注这方面。 本文将通过回顾理论观点使大家了解的胜任力运动，回顾我们在新西兰使用胜任力模型的经验，并仔细检查支持它们应用的假设。现存的研究是与曾用于证实胜任力模型价值的结果测量的特殊参考一起回顾的。最后我们确定了各种各样的研究领域和那些I/O心理学家如果被卷入了胜任力模型在组织环境中的提升、发展和实现时，明显会研究的问题。 什么是胜任力,三种主要途径 许多已发行的定义可以被归纳为三种有区别的途径:教育标准，行为技能，组织能力。 1. 教育途径(技能发展，成就的标准，授予证书) 现代的胜任力运动起源于教育学科。在美国，‘胜任特征’基于功能角色分析，描述角色结果或者知识、技能和态度两者之一，或两者皆有，要求角色绩效，并且被按照标准评估，通常是一种行为标准。在英国，行业主体，尤其是那些需要交易和专业技术能力的行业，在预期工作结果的基础上开发了职业的胜任力标准(弗莱彻，1992)。‘胜任力’被狭隘的定义为一种要证明的行动、行为或结果，或者是一种最低标准，不同程度的掌握由不同的陈述定义(伯克等，1975，埃兰，1971)。 2. 基于心理学的研究-(行为技能) 1973年，在美国教育领域工作的麦克里兰写了一份报告，指出他定义为动机和个人特质的个人胜任特征与传统的心理测验(比如IQ和能力倾向测验)相比，是一种更好的预测职业成就的手段。 麦克里兰的工作产生了极大的影响。特别有趣的是一个想法,即因素或与个人成功有关的输入是可以被鉴定出来，然后传授给别人。麦克里兰和博亚兹(1980)开发了一种研究方法来识别胜任特征，基于特定组织中公认的明星演奏家熟练的行为技能。他们把胜任特 11 征定义为“一个与工作中优秀的或有效的绩效有因果联系的知识、动机、特质、自身形象、社会角色和技能的通用体系”。(P.369，斜体字) 3. 商业途径(组织胜任力的竞争优势) 在20世纪80年代末，商业策略家开始重视胜任特征的概念。哈默和普哈拉德(1989)增加了“核心竞争力”和“能力”的概念。他们把核心竞争力看做一个组织的“集体的学习”的定义被广泛的引用，并有助于“胜任特征”的现行利息(谢普曼等，2000)。因此，斯拜罗(1995)建议实践者应该着眼于界定“更高等级”未来导向的组织胜任特征。 胜任力研究有哪些潜在利益, 绩效利益是有指望的通过多样的定义，包括胜任特征和工作绩效的因果关系或工具性关系(博亚兹，1982)以及胜任特征和组织绩效。(敖根，1988;哈默&普哈拉德，1989) 另外，斯拜罗(1995)观察过胜任力文献，包括针对组织中人力资源处理过程的一个巨大范围的声称的利益。总之，有: • 通过必须的胜任特征的焦点来改进招聘与选拔实践; • 改进个人、组织和职业发展计划; • 为了改进评估而改进绩效管理程序; • 通过一种共同语言来改进策略和人力资源问题上的交流。 什么是胜任力模型 组织采用胜任力的方法，必须创造或利用一种胜任力模型，至少是一个简单的列表或目录，具体说明合适的胜任特征。这个模型的构成必须支持胜任特征在选定的人力资源功能中的应用。 为挑选和教育的目的设计的模型通常在一个详细的水平描述技术性的胜任特征，依据他们的前期的技能和知识。那些旨在公布行为技能和公民行为或组织胜任特征的模型，以更高的水平代表性的描述胜任特征。不管什么方法，一个胜任力模型应该为每个胜任特征和潜在胜任特征提供一个操作性定义，连同可测量的或可观察的绩效指标或标准来准对个人评估。 如何把胜任特征联系到其他应用于I/O心理学的概念, 谢普曼等(2000)指出，胜任力建模将会在人力资源中呈巨大趋势。当工作分析的焦点主要在个人水平，检验专业知识、技能、能力和其他个人工作绩效所必须的特性时，胜任力建模表现了一种鉴别适用于不同任务和情境的绩效的尝试。与此相关的是I/O心理学 12 中广泛的文献代表了数十年对与工作绩效和组织绩效有关的因素的研究。 O’ Reilly和Chatman(1986)建议有两种明显不同的变量和工作绩效相关;首先是工作中必须的角色内行为，其次是在特殊角色中没有明确规定的亲社会性行为。Brief和Motowidio(1986)发现了亲社会组织行为(POB)的13个方面，亲社会组织行为定义为目的在于提升组织中其他个人或团体的福利的行为。亲社会行为是假设改善沟通、工作和客户满意度，继而工作绩效。角色内行为和亲社会行为的各方面也许都包含在胜任力模型中。组织公民行为(OCB)是一个类似的概念，并能引起值得思考的文献。Organ(1988)把OCBs定义为个人的行为，除了那些在角色或工作描述中必须的，也就是，作为总体，有助于组织绩效的。 与O’ Reilly和Chatman的模型(1986)相似，Motowidio等(1997)发现了整体工作绩效的两个元素;任务绩效和周边绩效(关系绩效)。后者本质上是促进任务绩效所必须的社会化，应用和努力，相当于OCB(Organ，1997)。Motowidio等建议任务绩效包含的活动最可能在角色之间不等，然而周边绩效包含的活动通常都是相似的。更进一步，他们建议任务绩效的前因或预测更可能涉及认知能力，而个性更可能影响周边绩效。任务绩效包括技术性的和任务知识的应用，以及任务习惯，定义为对任务情境反应的特征(Borman等2001)。周边绩效包括行为和特质，比如毅力和努力，志愿活动，帮助和合作，忠诚，政策和程序上的服从，对组织目标的支持和促进，主动权和自我发展(Borman & Motowidio 1997)。多数的通用的胜任力模型和目录强调周边绩效的方面相对于任务绩效。 个性也和工作绩效有关。个性的五因素模型包括一个责任感的多维的因素，即描述努力和应用的方面(Anastasi 1997)。这个因素被发现与周边绩效有关，特别是在工作任务责任心的方面，以及整体绩效穿过大范围的工作，(Tett和Burnett 2003)，和职业提升(Viswesvaran & Ones 2000)。这增加了关于是否投资大量的胜任力模型的问题，访问主要的周边绩效，提供任何增加的效用。 在组织认同感方面的文献区分了三种类型的承诺，态度性承诺;对组织的信仰，工作性的承诺，基于感知的损失和收益，和规范性承诺，社会化程序的结果。(Mathieu & Zajac，1990)。这和胜任力模型应用于提升和奖励例证期望的组织价值和核心竞争力的行为有关。 知觉组织支持(POS)是职员相信他们被组织评估的程度。这和组织承诺和工作绩效有关。Rhoades和Eisenberger(2002)发现公平是POS中最重要的因素，随后是监督人的支持。Colquitt等(2001)概括了组织公正的元素作为处理的一致性介于个人和超时，没有偏见，信息的准确性，符合当前的道德标准，一个受影响的个人和组织的声音，和一种机制来查阅及改正错误的决定。重要的是胜任力模型用于评估和绩效评价的目的是被职员感到公平的。 13 胜任力模型该如何实施, 试图量化胜任力模型的质量，Shippman等(2000)提出一个建立从工作到胜任特征分析的胜任特征分类标准的10分水平的精确衡量。这包括有效的数据收集方法，胜任特征描述符号开发程序和质量要求，与商业策略联系，确认方法，和文件材料。 模型定义方面的尝试经常从广阔的组织中得到想法，以便创造买进。由于有许多潜在的方法来定义和表达胜任特征，这会导致一个拉长的昂贵的处理方法，以及结果受制于效果比，你将得到没有人不同意的，不一定是最好的定义。 供替代的选择，买一个现成的系统，有可能一开始更便宜，但是可能需要应用者不间断的努力来改进它，使之适用于他们的情况。无论哪种方法，一旦实施，能力评估过程携带了一个重要的行政负担，以及组织最需要确定的是这些投资是值得的。 然而，有主要的有效性问题与胜任力模型的应用有关，以及到目前为止有很少的证据来支持他们宣称的利益。 胜任力途径的议题 1. 结构效度—什么是胜任力—胜任力可以被实施以便它可以被观察和测量, 结构效度的目的在于评价一个个人特质的测量或特征事实上测量它意味着什么。正如许多心理结构，不可能有真正的世界方面的胜任特征(Cronbach & Meehl, 1955)。建立结构效度因而需要找到建构的适当的代理人标准(Ghiselli, Campbell, & Zedeck, 1981)。这很明显是胜任特征的问题，包括很多研究文件不同意管理者、职员甚至专家要求分类胜任特征的操作性定义，通常以行为为例(美国管理协会，2003;Horton等，2002)。这也说明了应用胜任力模型来传递、提升和奖励组织规范的困难。 1a. 内容效度和表面效度—在组织中胜任特征是不是可信的, 胜任力模型的应用有一个主要问题，就是缺少对于术语‘胜任力’的意思的一致性。Sparrow(1995)提出各种胜任力途径的可交换的应用鼓励组织“在流动的沙床上建立和整合HRM系统”(p. 168)。 那么，胜任力模型应该包含什么呢—即，我们如何确定它们的内容效度和表面效度,内容效度的意思是胜任特征的描述符号是兴趣的世界的一个代表性样本。表面效度的意思是胜任特征自身被使用者判断并认为是精确和适当的。至于一些特别的胜任力模型，内容效度和表面效度是本质上的主观判断。那些对于谁的任务他们在申请是在一种情况下来判断胜任特征是否合适他们的任务。因此内容效度和表面效度的议题可能出现不管主题有关专家的应用和出现什么样的系统的信息收集方法。 14 Hayes等(2000)主张把胜任力分解为一个详尽的元素列表也许是不可能的。这就意味着胜任力模型会一直都是不完全的。他们引用管理者不能描绘一个角色必须的所有的胜任特征的研究作为例子。当然，在最通用的胜任力模型中的很少的行为报告不可以被认为是详尽的。 与此相联系的是为一个胜任力模型达成一个适当的结构的困难。在定义胜任特征，Stuart(1983)强调普遍性和特殊性之间，带宽-诚信问题，以及复杂性和简单性之间的交替换位(见表1)。普遍的或通用的胜任特征就是那些通过角色和组织合适的，反之，特殊胜任特征就是那些对于角色和组织来说特别的。普通的或通用的胜任特征具有定义过于广泛的危险以至于它们不被人们认为是有关的。胜任特征可以被简单的定义，作为一个内容提要加上一点行为例子，或者它们可以被设计用于满足多层次的细节和掌握。同时，复杂的模型允许更精确的要求和评估的交流，它们可以变成一个管理的责任。 事实上，普遍的或通用的途径是最一般的被采用的。作为一个例子，Tett等(2000)企图鉴定和验证一个“管理能力的高度紧张的空间分类法”。很多专有的通用胜任力目录，也与Stuart(1983)的普遍胜任特征相一致，代表性的包括管理和OCB因素。这些胜任特征中的很多都被广泛的定义以至于它们包括了一个个性因素，动机和认知能力的混合(Bartram，2004)。在一个近期的国际性调查中，在拥有胜任特征目录的28%的大型公司中，有将近一半的公司在用有通用胜任特征函数库的专有的软件(Metagroup，2004)。但是一个规格适合所有的通用胜任特征途径是不太可能适合致力于不同设置、不同产品和不同顾客的组织(Chiabaru，2000)。正如Stuart(1983)提到的，胜任力模型越简单越普遍，个人水平的感知相关性越低。 1b. 效标效度—胜任特征可以被精确测量吗, 一个有关问题是很多胜任特征，尤其是那些与周边绩效有关的，被广泛的定义，并且几乎没有绩效指标。在这种情况下不见得精确的评估是合理的。这对组织的公平是有含义的(Colquitt等2001)，以及感觉组织的支持(Rhoades & Eisemberger 2002)影响员工对组织的承诺。 除了测量标准的不足，胜任特征通常用自己或管理人的评价来评估，有时候也会是同事之间的。因此，胜任特征的评定有可能遭受相同的可靠性问题，比如评定人的偏见，广泛的文献记录为了一般的绩效评估(Fletcher，2001)。 然而胜任特征的精确测量是一个关键问题，尤其是当评估被用于付钱的绩效计划。组织的一个主要的前职业是精确的区别不同等级的成功，以便确保“顶尖人才”感到值得并且得到适当的奖励。例如，Hunter等(1990)发现在复杂的任务中，比如职业的服务，个人的产出可以被一个因素改变多达12比1在最好和最差的执行者之间。 15 表1. 一个使用了Stuart(1983)的结构的胜任力模化选择图解 定义 简单的 复杂的 象限 1 象限 2 , 高水平 , 高水平 , 容易使用 , 复合结构，可能难使用 , 可能不被看作是相关的 , 与任务更相关 通用的 , 允许任务间的比较 , 允许任务间的比较 , 评估较少准确 , 评估更精确 , 可以购买现货 , 可定制解决 胜任特征 象限 3 象限 4 , 任务特殊/技术化 , 任务特殊/技术化 , 容易使用 , 多层次，复合结构，难使 特殊的 , 没有任务/组织间的比较 用 , 评估精确 , 没有任务/组织间的比较 , 可定制解决方案 , 评估精确 , 可定制解决方案 2. 胜任力模型的确认 第二个主要问题是组织应用胜任力模型的方法;即没有确认就采用(Shippman等，2000)。确认是很重要的，因为胜任特征描述了标准的行为，即组织希望提升和发展来提高组织绩效的行为。 3. 预测效度—改进的胜任特征能预测改进的个人工作绩效或改进的组织绩效吗, 第三个主要问题是缺乏采用胜任力途径所得结果的利益的证据。所有胜任力主动性的潜在假设是个人技能发展，以特殊的行为作为例证，会导致工作绩效以及组织绩效的好转。Barrett和Depinet(1991)的回顾，关于研究胜任力测量为McClelland(1980)的断言，即胜任特征相对于传统的心智能力的心理计量测验是更好的工作绩效预测工具，提供了很少的经验上的支持。 后来，Laber和O’Connor(2000)注重胜任力模型的效能的实证研究的缺乏。我们对于文献的检索揭示只有少量的研究调查胜任特征和客观的工作绩效之间的联系，四年以后留下了基本不变的情况。 16 关于胜任力运动的众多批评之一是胜任力和工作绩效隐含的混乱。我们认为这个混换已经出现了，通过和胜任力运动有联系的语言;行为、知识和人格特质，即工作的输入以及结果或输出，和客观的工作绩效的混乱。因此，胜任特征被行为“绩效”评估并认为是标准的胜任力。阐述这个混乱，Campbell(引用Bartram，2004，p.5)陈述“绩效是行为。是一些人们做并反映在人们行动中的东西„绩效不是行动的结果;它是行动本身(见Hackett，2002)。 这种对术语的混淆造成了胜任力模型应用中的一个固有的循环。胜任特征被鉴别通过使用各种各样的信息收集方法，行为标准被定义，然后缺乏对工作输出或工作绩效的客观的测量，对这些行为事件的主观评估被假设与工作绩效相当，并且确认胜任力概念自身。作为说明，Mayer(2003)报告了一个研究，检查健康的劳动力胜任特征是否可以预测基本的服务绩效。他测量自我评估核心胜任力等级和自我评估服务绩效之间的关系，即在美国的一个大城市的卫生部，定义为公共健康工作任务的绩效的频率。他发现，胜任水平和本质上的服务绩效输入只有一个适度的联系。 Hunter和Schmidt(1996)指出OCBs和对个人工作绩效以客观输出为基础的测量之间的相关是很少的。然而，当监察员估计工作绩效时，在OCBs的评定等级和他们对整体的工作绩效的主观评定等级之间有很高的相关。他们指出，这是因为监察人宽容差的任务绩效出现在有高水平的OCBs的人身上。因此，任何协会的通用胜任特征和工作绩效都可能会由于监察员给出的对OCBs的衡量而改变(Johnson，2001)。 17