This Section contains papers on Performance Management and Performance Appraisal.

Performance Management

Written 27th Feb 2003

Authored by Kelly Hooper

Performance is a concept that means different things to different people. As such, it tends to be measured in a number of different ways. The underlying issue for organisations is to turn performance assessment into a management and communication tool, which in turn becomes a performance management system. Banks and May (1999) suggest that such a system must have at least three elements to be effective. These include: (i) a well-articulated set of performance dimensions indicating an employee's valu e to the organisation; (ii) a behaviourally anchored continuum of effectiveness for each dimension to verify degrees of contribution and (iii) a structured process for communicating the feedback on performance. If performance management, including appraisal, is included as part of a managers' everyday duties and responsibilities, it is more likely that they will engage in it on a continuous basis, thereby providing feedback and coaching to employees as needed.

Over the years the focus has been on improving the accuracy of assessing performance as well as designing effective performance development tools. Some of the ways in which performance management effectiveness can be enhanced include designing the 'right' format and framing this within the context of setting goals (Banks & May, 1999). In other words, the format should be user-friendly, ensure it captures the appropriate job content and minimizes both judgmental and perceptual errors. It shoul d also allow performance to be a matter of setting an employee's annual goals then measuring the degree of attainment. In addition, it would be ideal for assessment to be received from all levels of the organisation (superiors, subordinates and peers) as well as from its customers.

Performance appraisals can have a direct bearing on an individual's motivation and can lead to both increased and decreased performance. Therefore it is important to design an effective system that aligns with organisation and employee needs. Some of the issues to consider when designing the appraisal process include:
- what the results will be used for
- how will the information for the appraisal be gathered
- rater training
- rater motivation
- self and peer appraisal
- feedback of information to employees
- assessing the system

JOB ANALYSIS
A thorough job analysis is an essential first step in developing a legally valid performance management system as it derives the performance dimensions, defines tasks and activities in behavioural terms and defines important and critical aspects of the job.
This systematic process involves a better understanding of the context, content and requirements of an existing job and is used when a new job is being created, changed or when there is uncertainty that the results of an old job analysis are still accurate. A good job analysis helps ensure that the focus of the performance appraisal is actually upon the work that is being performed, rather than something irrelevant like an employees' personality or political beliefs.

How to conduct a job analysis
Algera and Greuter (1998) list four approaches that can be adopted in conducting a job analysis. All methods of job analysis can be classified as belonging to one of the four approaches described by Algera and Grueter. These approaches are:

1. Behaviour description, where the focus is upon the behaviours displayed by task performers.
2. Behaviour requirements. This approach is aimed at the behaviour the task performer should display to perform a task successfully. The emphasis is particularly placed on the intervening processes between stimulus and response.
3. Ability requirements where tasks are analysed in terms of human abilities and personality traits required for successful task performance.
4. Task characteristics, where the focus is upon the objective characteristics of the task. This material is considered separately from the behaviour which should be displayed (behaviour description) or the required human abilities (ability requirements). The emphasis in this approach is placed on the actual objective characteristics of the tasks. This approach is preferred when job or work redesign is being considered with a view to increasing job satisfaction and motivation.

It is possible and, in fact, advisable to utilise more than one of these approaches when conducting job analyses to ensure a comprehensive coverage of the job content domain. Information sources can include present job incumbents, training manuals, role descriptions and supervisory accounts. Saville and Holdsworth (1998) advise structuring the survey sample for jobholders from different ethnic backgrounds, age classifications and gender mix to ensure an adequate representation of jobholder charac teristics.

DUE PROCESS
An important legal consideration in performance management is the concept of due process. That is, those criteria must cover only those aspects of performance, which are within the employee's control. The primary objectives for due process is the need for an adequate notice period for when one's performance is to be managed, fair hearing and judgement based on evidence (Folger, Konovsky & Cropanzano, 1998). Therefore, employees should have input into the formulation of objectives and standards or at least opportunities to question the content of standards, the process whereby these were established, and the manner in which these standards will be implemented. In other words, it is necessary to conduct a job analysis using subject matter experts, and to conduct formal consultation with employees about the ways in which goals are going to be set. Performance criteria must reflect relevant specific clearly defined and observable dimensions of the job performance.

The objectives of due process are met by establishing employee input, two-way communication, the right to challenge, rater familiarity with the employee's work and an application of consistent standards.

Most organisations use the performance appraisal system to assess an employees strengths and weaknesses, on the basis of which decisions are made about rewards, promotion and development. The practical importance of these decisions places a premium importance on the quality of appraisal information. However, rater distortions of employee appraisals and employee attempts to exaggerate apparent performance are facts of organisational life that can never be completely eliminated. These distortions m ay be intentional or unintentional (Kozlowski, Chao & Morrison,1998). Unintentional biases are generally regarded as a function of the rater's privately held evaluations influencing ones perception and judgment of the ratee. Intentional bias is a deliberate rating distortion and tends to be more prevalent. Contextual factors that may promote conscious manipulation of appraisals by raters include administrative policies, purpose and degree of accountability. Additional influences also include characteris tics of the organisation such as its culture, climate, socialisation processes and reward structure (Kozlowski et al, 1998)

Most managers and employees are aware of the types of problems that arise in performance appraisal. 'Easy' and 'tough' managers and managers who play favourites create situations which result in unfair ad inaccurate ratings. Common examples of rater errors / perceptual biases include:

Halo effect
The most common rating effect is halo and occurs when a rater assigns the same rating to all aspects of an employee's performance, regardless of the actual level, because of an overall favourable or unfavourable impression of the employee. Halo can occur at any point along the scale, although the most frequent tendency is for people to give uniformly positive ratings. In many cases halo represents an error in that the rater incorrectly generalizes from performance on one dimension to other dimensi ons rather than carefully considering each separate aspect of performance.

Central tendency
This effect is caused by a manager giving everyone an average or acceptable rating, regardless of actual performance. These errors arise from 'playing safe', having few opportunities to observe the employee's performance and being unwilling to justify high or low ratings to the employee or the organisation

Leniency / strictness bias
Occurs when managers rate their employees either consistently high or low. Severity (strictness) occurs when a rater gives mainly negative evaluations. Leniency occurs when raters assign all ratees favourably. The most serious problem with these types of effects is that they do not allow a differentiation among employees

Order effect
The order in which information appears can influence appraisals. Primacy effect - the first information about a rate influences the final evaluation more than information appearing later. Recency effect - when a manager overemphasizes the employee's most recent behaviour. Research suggests that primacy effects are more likely when one overall evaluation is made after observing information on the rate, whereas recency effects are more likely when the rater evaluates the ratee on each item of inform ation as it occurs

Impression management
In addition to the rater's motivation to distort performance ratings, ratees have the potential to indirectly exert influence as well. This is done through impression management, or selectively behaving so as to enhance their image in the raters eyes. This is likely to affect the extent to which the supervisor liked the subordinate and perceived similarities with the subordinate which in turn directly influences their ratings of their subordinate's performance.

Characteristics of the performance evaluation system that minimise political and impression management influences tap all phases of evaluations. Kozlowski, Chao & Morrison (1998) suggest the following recommendations to manage politics and impression management in performance appraisals.

In the fast-paced competitive world in which most global businesses run today, ongoing performance management is a valuable tool for staying on course or changing direction, as the marketplace demands. It should be remembered that if performance appraisal is to remain an effective management tool, the system must be consistently monitored to ensure that its results are fair, accurate and related to performance. If performance management and appraisal is interwoven into the fabric of managers' eve ryday duties and responsibilities, it is more likely that managers will engage in it on a continuous basis, thereby providing feedback and coaching to employees as needed.

Disclaimer:
The information provided on this page does not constitute treatment and in no way replaces direct advice from qualified professionals providing tailored solutions to particular workplaces and individuals.

References

Banks, C.G. & May, K.E. (1999) Performance Management: The real glue in organisations. In Kraut, A.I. & Korman, A.K. (Eds). Evolving Practices in Human Resource Management. San Francisco: Jossey-Bass.

Kozlowski, S.W.J., Chao, G.T. & Morrison, R.F. (1998). Games raters play: Politics, strategies and impression management in performance appraisal. In Smither, W. (Ed). Performance Appraisal: State of the Art in Practice. Jossey-Bass.

Algera and Greuter (1998) - RSA Performance Management seminar
Saville and Holdsworth (1998) - RSA Performance Management seminar
Folger, Konovsky & Cropanzano (1998) - RSA Performance Management seminar

Performance Appraisal: A Literature Review

This paper was prepared for the Australian Centre for Industrial Relations Research and Teaching, Sydney University September 1999

Authored by Martha Knox-Haly

Introduction

Performance appraisal is an increasingly significant aspect of human resources practices in Australian Workplaces. It is important that the development of performance appraisal systems occur in a systematic, methodologically rigorous and informed manner. The following paper provides a detailed outline of the steps associated with the development of appropriate performance appraisal system. The literature review is not exhaustive of all the topics referred to in this paper. However there has been an effort to be as comprehensive and practical as possible in the compilation of this material.

This paper describes the processes involved with conducting a job analysis and alludes to some of the more popular job analysis methodologies. Consideration is given to the reasons for conducting a comprehensive job analysis. These reasons include establishing the relevance of a performance appraisal system to the job at hand, as well as ensuring that the performance appraisal is representative of workforce and workplace characteristics. This section describes methodologies such as the Critical I ncidents Technique, The Repertory Grid Method, Strategic Job Analysis Interviews, The Position Analysis Questionnaire, the Work Profiling System and Assessment Centres.

The second section refers to the development of performance appraisal indicators. It touches upon the level of abstraction that can be associated with performance indicators. This area incorporates debate over whether personality or task behaviour is the preferred approach in the development of performance indicators. It reflects upon the legitimacy of including organisational citizenship behaviours as part of a performance appraisal process. The trend within performance appraisal literature has been to favour performance appraisal systems, which are based exclusively on task behaviours. However there has been a recent push for greater acceptance of personality measures as performance appraisal tools. Reference is made to the role of distinction between task behaviour and personality in the application of assessment centre methodologies.

The third section refers to the types of supports that are required in creating a fully-fledged performance appraisal system. This section includes diary keeping, training and the process of delivering feedback. A key consideration in this section is whether the system should include multiple rating sources, and who should conduct performance appraisal ratings. The available evidence supports the use of multiple raters, and in particular points to the utility of incorporating peer ratings to incr ease the validity of the performance appraisal process. The need for training that addresses rater bias and due process is also raised in this section.

The conclusion is a summary of the key attributes of a well-developed performance appraisal system and describes the implications of these attributes for Australian Workplaces. This section explores the ramifications of personality based performance appraisal measures and multi-rater systems for Australian workforces.

Section I: Job Analysis

Why have a job analysis?

A thorough job analysis is an essential first step in developing a valid performance appraisal system. This type of analysis ensures that the performance appraisal indicators reflect the realities of job content and how work is performed within that job. Saville and Holdsworth (1998) have emphasised the importance of job analysis being structured, analytical and documented so it will satisfy workplace legisaltive requirements. Algera and Greuter (1998) list four approaches that can be adopted in conducting a job analysis. All methods of job analysis can be classified as belonging to one of the four approaches described by Algera and Grueter. These approaches are:

1. Behaviour description, where the focus is upon the behaviours displayed by task performers.

2. Behaviour requirements. This approach is aimed at the behaviour the task performer should display to perform a task successfully. The emphasis is particularly placed on the intervening processes between stimulus and response.

3. Ability requirements where tasks are analysed in terms of human abilities and personality traits required for successful task performance.

4. Task characteristics, where the focus is upon the objective characteristics of the task. This material is considered separately from the behaviour which should be displayed (behaviour description) or the required human abilities (ability requirements). The emphasis in this approach is placed on the actual objective characteristics of the tasks. This approach is preferred when job or work redesign is being considered with a view to increasing job satisfaction and motivation.

It is possible to utilise more than one of the above approaches in conducting a job analysis to ensure comprehensive coverage of the job content domain. Saville and Holdsworth (1998) stress the importance of using more than one source of information in conducting a job analysis. Information sources can include present job incumbents, training manuals, role descriptions and supervisory accounts. Saville and Holdsworth also advise structuring the survey sample for representativeness of job holder c haracteristics. This means surveying job holders from different ethnic backgrounds, age classifications and having an appropriate gender mix in the survey.

In conducting a job analysis, it is necessary to consider whether the analyst should focus on performance results or work behaviour activities. Algera and Greuter (1998) note that in many cases choice is simplified by the fact that there are no directly measurable indicators of performance outcome, (i.e. an operators job may consist of service functions, control functions and panel watching). Practitioners conducting a job analysis can ask What is done on the job, (the objective question), what s hould be done on the job (the prescriptive question), and what will be done on the job (the predictive question), (Cronback, 1998)?

The process of job analysis may be conducted through interviews, structured questionnaires, identification of critical incidents or through application of the repertory grid method.

Critical Incidents Technique

According to SHL, C.J. Flanagan first developed this approach for analysing the role of an air-force pilot. Job incumbents are interviewed to identify examples of critical job functions which are either performed very well or very poorly. The job data is generated with the assistance of a trained interviewer. After the incidents have been collected (often 200 to 300 per job) and grouped into categories of 10 to 20 incidents. Job incumbents, supervisors and others are consulted each time to assess and correct intermediate results. The resulting categories called behavioural dimensions are used as the framework for identifying human characteristics and behavioural requirements associated with the job.

It is necessary to document the complete role set for the job in question by interviewing the various categories of people involved. This means gathering information from supervisors, job incumbents, support staff, suppliers, planners and senior management. This approach makes it possible to compose an extensive list with specific behaviour requirements which can be used in the selection of candidates.
The benefit of this approach is that it explores actual events, which have occurred, on the job in the past. It also provides a rich source of material for assessment centres and situational interviews.

Repertory Grid

The American Psychologist George Kelly developed the repertory grid technique in the 1950s. Repertory grid interviews were originally used to determine how people perceived and classified things that occurred in the environment around them.

This technique has been applied to job analysis. It is most commonly used with supervisors who are asked to compare individuals who work for them with a view to identifying the skills associated with successful job performance. The supervisor is asked to generate job constructs by picking three job holders, and identifying how two are different from the third in terms of the way in which they perform their jobs. The supervisor is asked to consider examples of successful and not so successful job performance which has been demonstrated by their subordinates. This process is repeated until a number of job constructs associated with job effectiveness have been built up. This technique has the advantage of being able to be statistically analysed when large numbers of grids have been generated.

Strategic Job Analysis Interviews

A job analysis can also be conducted through strategic job analysis interviews, (Bobrow and Leonards, 1997). This method can be used when the job does not yet actually exist or when the number of subject matter experts is small enough such that a survey would not be warranted. These types of interviews focus on not only what the new job knowledge, skills and abilities would be, but also how the new job would differ from the current job.

Structured Questionnaires

Computer scored questionnaires have been developed for job analysis procedures. McCormick's Position Analysis Questionnaire, the Job Analysis Questionnaire and Work Profiling System are examples of such questionnaires. Questionnaires are usually completed by job holders, but can be completed by line management or staff involved in job design.

The Position Analysis Questionnaire gathers descriptive and attribute based information about a job. The descriptive aspect of the Position Analysis Questionnaire consists of 189 items. Each item related to an independent work activity or job element. The questionnaire is divided into six sections, which are:

1) Information input: items relating to the kind of information one has to process or handle in a job. E.g. written material.

2) Mental processes: items on the required mental processes. Example taking decisions, solving problems, degree of planning.

3) Work output: items that have to do with the execution of work, and more particularly use of machines and tools. Example: Drawing tools, measuring instruments.

4) Relations with other persons: items about contacts that are necessary to do that work. Example: contacts with the public.

5) Job context: items on the work environment and work situation. Example: outside work, high temperatures.

6) Other job characteristics: category for miscellaneous items about shift work, clothing, responsibilities for materials.

Information generated by the descriptive aspect of the questionnaires is called job data. Attribute requirements (attribute data) has also been developed for each job element. Judges or subject matter experts are asked to consider the relevance of sixty-eight attributes. Judges use a six-point scale from 0 to 5 to assess the level of importance of each attribute to successful job performance. The profile of required human characteristics (attribute profile) was determined for each job element, by calculating the median for each of the 68 characteristics.

Saville and Holdsworth developed the Work Profiling System (WPS) in 1989. The Work Profiling System - this has three different questionnaires for the following job groups: management and staff, administrative and service personnel, skilled and technical workers. The full WPS contains more than 800 items. Human attribute profiles have been constructed for the activity described in each item.

In the design of the WPS, Human Attribute Profiles (HAPs) have been constructed for the activity described in each item. Human Attribute Profiles are a weighted linear composite of important human characteristics, which have been set up for each item. The computer output consists of a short report of the most important tasks, the job context and the profile of required characteristics.

The Job Analysis Questionnaire assesses 20 abilities and nineteen personal traits. Rating scales have been developed for each of these characteristics with three to six different examples of each task.

The benefit of using such questionnaires is that they add more scientific rigour to the job analysis process and allow combination of the views of different raters. A structured approach also enables comparability of different jobs. The disadvantage is that such questionnaires are expensive, time consuming and do not utilise the language or terminology of a given workplace.

Assessment Centres

The literature on assessment centres generally focuses more on their usage as a performance appraisal process. However Lowry (1997) has written about assessment centres as a job analysis process. According to Lowry (1997), the task specific assessment centre is concerned primarily with evaluating the subjects performance in tasks (work samples) related to the job. Therefore the desired outcome from such a job analysis should be the tasks performed on the job, rank ordered based on importance, fre quency, and need to perform on entry to the position. The final product is called a task analysis.

One approach to doing the task analysis is to develop a questionnaire, which lists the tasks derived from the job description and interviews with subject matter experts. Incumbent captains and their supervisors determine the relative rank of the task responding to three questions concerning each task a) the relative contribution of this task to effective operations b) time spent on this task relative to other listed tasks c) the importance of being able to do this task correctly on entry to the j ob. Selecting the most critical tasks to be included in the exercises concludes the process. This approach differs from a job analysis for a dimension specific centre in that there is no need to attempt to identify required knowledge, skills or abilities.

Section II: Performance Indicators

Formulating Performance Indicators

Once the raw material has been created from the job analysis, the next task is development of performance indicators. This touches upon the level of abstraction that is required in writing descriptors. Algera and Greuter (1998) comment that there are three levels of abstraction, and these levels are incidental, habitual or abstract. An example of an incidental descriptor might be that an employee "is always on time", a habitual behaviour might be "keeps agreements", whereas th e abstract descriptor would be "attention to duty".

Algera and Greuter (1998) take the view that the first two levels of description are to be preferred as they are closer to observable behaviour and do not lend themselves to personality theory. The choice between habitual or incidental behaviour descriptors rests on whether generalisability across job types is required, and whether there is sufficient specificity. Lack of definitional specificity results in insufficient agreement between raters, and between the appraiser and the appraised. This i n turn creates problems for the review component of the performance appraisal system, which can become bogged down in semantic arguments between appraiser and appraisee. Algera and Greuter (1998) favour the habitual behaviour descriptors as a good compromise between observable (specific) and abstract (generalisable). If generalisation is not required across job categories, then incidental behaviours are to be preferred.

Clarity of standards is alomost important in promoting inter-rater reliability. Schrader and Steiner (1996) found that providing an explicit comparison standard (internal, absolute, relative or multiple) produced significantly greater inter-rater agreement correlations than when no comparison was provided (the ambiguous condition). It was also found that ratings produced in the ambiguous condition (no standards) were lower than those using more specific comparisons were. These researchers also fo und that there were no differences between self and supervisory ratings of performance in this study.

The nature of standards is quite variable and depends on the requirements of the performance appraisal context. An absolute standard identifies a specific, objective benchmark for evaluation, such as typing 60 words per minute. A relative standard requests that the evaluation be based on a comparison with other members of the workgroup. An internal standard entails a rating relative to the ratee's past performance and skill level. Finally a multiple standard requests ratings on consideration of a bsolute, relative and internal standards.

Personality or Behaviour?

There is a division within the psychological literature on performance appraisal regarding the utility of personality measures. According to Hogan and Shelton (1998) there is now some consensus in industrial/organisational psychology that well constructed personality measures can predict occupational performance. Conversely Algera and Greuter (1998) comment that there is a distinct disadvantage to this approach in that personality theories provide few clues as to the concrete behaviour which will be necessary to achieve an improvement in performance. For this reason there has been a favouring of systems, which emphasise concrete information, which refers directly to the job behaviour in question. Algera and Greuter (1998) argue that only where it concerns an estimate of someone's potential for a series of jobs in the (distant) future which may alter as to their exact content, is appraisal for more general capacities and personality traits inescapable. These authors argue that for all other cases it is advisable to link appraisal to concrete behaviour requirements that stem from the job in question. There is also evidence that personality based measures of performance can be influenced by manipulative impression management exercised by subordinates, (Day et al, 1996).

Drenth (1998) notes that the content domain of most performance appraisal instruments developed in the last three decades falls within the category of behaviour description For managerial jobs, such an approach is often operationalised by means of an assessment centre. The centre is based on a series of simulations and exercises which reflect tasks that are considered essential to the job. Other performance appraisal systems rely on behavioural observation scales, (BOS). For example, Ridderbos (1 992) developed a series of BOS scales for operators in the petrochemical industry. Originally these scales were developed as a criterion measure to validate a workshop sample type of a predictor instrument, but they were also found to be excellent as an appraisal instrument for regular personnel evaluation.

The acceptance of personality measures as an appropriate means of appraising performance depends on how the job domain is conceptualised. According to Hattrup, O'Connell and Wingate (1998) there is a theoretical consensus that the job performance domain is inherently multi-dimensional, although there is currently no clear agreement about which dimensions should be used to characterise performance for most jobs. Borman and Motowidlo (1997) drew a distinction between task and contextual performance , where the former is defined by specific job requirements and the latter is defined in terms of actions that enhance team or organisational effectiveness. Contextual performance refers to organisational citizenship behaviour and prosocial behaviour. These behaviours include persisting, volunteering, helping and co-operating, following rules and procedures and supporting organisational objectives.

Motowidlo and Van Scotter (1994) supported the validity of distinguishing between task based and contextually based performance. Their research demonstrated that task performance involved executing, maintaining or supporting the technical processes of the organisation. Contextual performance did not support the technical core directly, rather contextual performance created and maintained the psychological, social and organisational environment in which task performance occurred. When employees en gaged in task performance they produced goods and services. When employees engaged in contextual performance, employees use their skills and energy to build the social and organisational structure that maintained the technical core. Campbell et al (1996) listed eight primary dimensions of performance, which were presumed to account for the latent structure of performance in all jobs. These dimensions were: job specific task proficiency, non-job specific task proficiency, written and oral communication, demo nstrating effort, maintaining personal discipline, facilitating team and peer performance, supervision/leadership and management administration.

Borman and Motowidlo (1997) hypothesised that the antecedents of contextual performance are most likely to involve personality variables. They cite some indirect evidence to support this relation; however only one direct test of this hypothesis is available. Motowidlo and Van Scotter (1994) showed that six personality scales correlated more highly with ratings of contextual performance than with task performance. Although they answered their research question, it is still not clear what personali ty dimensions are most relevant for contextual performance and why.

The theoretical utility of personality measures is dependent on the organisational culture. To the extent that contextual performance is discretionary, organisational practices and procedures may offer rewards encouraging incumbents to engage in discretionary behaviours. When work conditions reward contextual performance, they introduce the possibility that individual differences might affect contextual performance through the effects of the perceived value of rewards. Hogan and Hogan explain the relationship between personality and contextual performance by reference to the dual motives of cooperation - "getting along", and status - "getting ahead". Hogan and Hogan (1998) speculate that where there are no opportunities for advancement, employees will perform contextual acts because they are conscientious. Where there are opportunities for advancement, employees engage in contextual acts because they are ambitious.

The research in contextual performance is hampered by definitional difficulties. Less is known about contextual performance as researchers have failed to agree on terms, sometimes substituting examples for definitions. Overlapping terms in the literature include extra-role behaviour, organisational citizenship behaviour, prosocial behaviour, principled organisational dissent, employee reliability and contextual performance. There is also a tendency to couch contextual performance in behaviouristi c terminology. In Borman and Motowidlo's (1998) view, observable behaviour represents only one perspective on contextual performance - it is the other or observer's performance. Although important for appraising contextual performance, this behaviourist view has little to offer as an explanation for individual differences in performance.

How personality is measured

Personality measures of performance have drawn upon either the Five Factor Model or Trait theory such as that utilised by Hogan and Hogan (1998). The dimensions of the Five-Factor Model are extraversion, agreeableness, conscientiousness, emotional stability and openness to experience. The evidence for the Five-Factor Model points to it having mixed utility in predicting performance. Mount et al (1998) found that there was substantial variability in the validity coefficients across jobs - within p ersonality predictors and across predictors within job categories for these factors. Mount et al (1998) found that emotional stability was an important predictor of team performance.

Mount et al (1998) looked at the degree to which the five-factor model of personality was related to jobs in which interpersonal reactions are reported. Results supported the hypothesis that conscientiousness, agreeableness and emotional stability are positively related to performance in jobs involving interpersonal interactions. Emotional stability and agreeableness are strongly related to performance in jobs involving team work. Hogan and Shelton (1998) were also concerned with an appropriate t axonomy of personality variables that were needed to study job performance, and how to align personality features with performance criteria. Hogan and Hogan presented data that five general factors were too broad to capture all the facets of job evaluation. Tett, Jackson and Rothstein (1991) demonstrated using a meta-analysis that the validities for personality measures increase when measures are chosen on the basis of job analysis, thereby supporting Day and Silverman's (1996) research.

Borman and Motowidlo (1993) predicted those personality variables such as conscientiousness and extraversion correlated with performance amongst sales personnel. Borman and Motowidlo (1993) found that absenteeism was significantly predicted by conscientiousness, but conscientiousness did not have any relationship with general cognitive ability. Conscientiousness also correlated with citizenship behaviour. Those who perceive their job responsibilities more broadly are more apt to exhibit citizensh ip behaviour at work, suggesting that prosocial behaviours at work may be motivationally and perceptually determined.

The Hogan Personality Inventory is another instrument, which has featured strongly in the literature on personality measures of performance. Hogan and Shelton (1998) conceptualised individual differences in performance as an outcome of differences in an employee's interests in "getting along" as opposed to "getting ahead".

Borman and Motowidlo (1997) describe the Hogan Personality Inventory (HPI) as a 206-item assessment of normal personality. The HPI is based on socioanalaytic theory, and contains seven scales that include measures of the big five personality dimensions. The test reliability coefficients, both in terms of scale internal consistency (average coefficient alpha = .80) and test retest reliability coefficients (average coefficient = .71) are within acceptable range. The HPI is used primarily within per sonnel selection and its validity defined in terms of non-test correlates is well established.

They measured employees in three different industries to minimise the possibility of industry effects. They focused on marketing analysts, financial bankers and hospital administrators. They found that the characteristics of an organisation and a job could moderate the relationship between personality and contextual performance. Conscientiousness predicted contextual performance, but in the second study where promotions were more possible, ambition/urgency predicted contextual performance. Specif ic personality constructs that do appear to predict contextual performance appear to depend on the job and whether rewards are offered for contextual performance.

These studies also point to the role of understanding job context behaviour. Borman and Motowidlo (1998) found that ambitious individuals were unlikely to volunteer or otherwise exhibit contextual performance. It has been found that three factors are salient as to whether people demonstrate organisational citizenship: organisational commitment, prosocial values and impression management.

Hogan, Rybicki, Motowidlo and Borman (1998) had employees complete the Hogan Personality Inventory as applicants and be subsequently rated by their supervisors for contextual performance. Again results indicated that conscientiousness was significantly related to ratings of work dedication and interpersonal facilitation, which are measures of contextual performance.

Personality as a moderating variable in performance appraisal

The research on the impact of supervisor's perceptions of personality indicates that it does have an impact on the manner in which performance is appraised. There just seems to be a question mark over how whether perceptions of personality influence performance appraisal directly or indirectly. Wayne and Liden (1995) found that subordinate's supervisor focused impression management behavior was only weakly correlated with supervisor's ratings of employee performance. However these authors found t hat a subordinate's impression management behaviour influences the supervisor's liking of the subordinate and the supervisor's perceived similarity to the subordinate, and these in turn affect the supervisor's performance appraisal ratings.

Warech, Smither, Reilly, Millsap and Reilly (1998) examined the impact of various personality traits on performance appraisal ratings. These authors found that supervisors tended to be more critical of subordinates less often than they were of subordinates who were low in impression management. Impression management in this context refers to ingratiation, self-promotion or behaviours, which are designed to manipulate or control a supervisors impression of the subordinate. The tendency to engage i n manipulative impression management behaviours has been established as having a strong link with self monitoring.

The role of personality traits in influencing performance appraisal also depends upon the way in which performance is conceptualised. It depends on whether job performance is conceived in exclusively task-based terms, or whether performance is described in contextual terms. For example, Warech et al (1998) found that self-monitoring ability was positively related to assessors and supervisors ratings of interpersonal effectiveness, but was unrelated to ratings of technical business competence. Thi s pattern is clearly consistent with the interpretation that self-monitoring ability is positively associated with job related interpersonal effectiveness. These authors also found that target managers who were high in self-monitoring (or motivated to control the way that others view them) were rated more negatively with regard to business competence by their peers than target managers who were low in self motivation. This finding has particular meaning for performance appraisal systems that are exclusively reliant on supervisor ratings and lack sufficient structure. One possible implication being those performance appraisal systems with these characteristics are much more vulnerable to manipulation by subordinates who are high in impression management. It is conceivable that the ratings, which are based on personality rather than behaviour, are also more vulnerable to such manipulations. For example, Day et al (1996) reported that the correlation between self-monitoring and job performance was higher for sub jective measures of job performance.

Assessment Centres

A similar distinction between personality or task behaviour focused intruments can also be applied to Assessment Centre methodology. Assessment Centres are commonly used performance appraisal instruments. Assessment Centres generally involve the assessment process being conducted as a separate exercise from an employee's daily performance behaviour.

Lowry (1997) classifies assessment centres as either being task specific or dimension specific. Dimension specific assessment centres are essentially only forums that give subjects the opportunity to display requisite traits and constructs. The key word that best describes the exercises in a task specific assessment centre is work sample. In other words the exercise must be directly related to the job. The assessors in task specific centres are not attempting to determine how much leadership or j udgement etc a subject has. They are attempting to measure how well the subject handles a specific job-related situation.

The basic objective of the task specific assessment centre is to obtain a score from each assessor that reflects his or her best individual judgement of how well the subject performed a work sample. In a dimension-specific centre where the assessors make judgements about how much of a trait or given construct a subject has. There are no differences between the assessor criteria for dimension specific or task specific centres.

Role players are an important part of the task specific assessment centre process. They are often used to stimulate employees with personal or work related problems, with members of the work team, or as superior officers. The exercise design should include a carefully crafted scenario and instructions for the role player. The role players must be given ample time to rehearse the role and to make suggestions for improvements. The jobs of role playing and assessing are generally undertaken by separ ate indviduals. It is very easy for an assessor/role player to become emotionally involved in the exercise to the detriment of objectively evaluating subject performance. It follows that role players should not participate in the evaluation process.

In dimension specific centres, the assessors are expected to observe subject behaviours, to record the behaviours, to classify the behaviours into trait or construct categories, and to score them. The task specific centre uses the same observation record sequence, but does not require classification into complex trait or construct categories.

In both types of assessment centres, there is a requirement for behaviour observation aids such as checklists. In task specific centres the checklists can include a short list of items considered important for successfully accomplishing the work sample exercise.

The behaviour observation procedure should include a method for recording the subject's actions. At the conclusion of the exercise, the assessors should have an opportunity to share with each other what they saw and heard, without making any evaluation or divulging their score. Upon conclusion of the exercise, the assessors should prepare a written summary of the subject's strengths and weaknesses to be given to the candidate after the Centre has been completed. In essence the assessors do not at tempt to achieve any form of consensus on the scores; they pool the observed behaviours to ensure that all relevant behaviours are considered.

Scoring based on all observed behaviours, the assessor will enter a score for each of the listed actions. An overall score may be required. This overall score is not an arithmetic average of the checklist items, rather it is based on the assessors overall opinion of the effectiveness of the subjects in this exercise. When you take this approach it is possible to calculate coefficients for inter-rater reliability and the reliability of the exercise.

Section III: The Performance Appraisal Process

Who should do the ratings?

Conway and Huffcut (1997) have observed that multi-source-rating systens have become increasingly popular in recent years. These systems typically make use of supervisors, peers, and the workers themselves (self-ratings). Murphy and Cleland (1995) conjectured that subordinates and peers have more opportunity than supervisors to observe interpersonal behaviours. Conversely supervisors have been found to have higher reliabilities, and Conway and Huffcut speculate that this is possibly because it is part of a supervisor's job to evaluate and be observant of staff.

In their meta-analysis, Conway and Huffcutt (1997) found that no one source had a particularly high reliability for a single rater, and managerial jobs in particular produced low reliabilities. Fortunately multiple raters are often available (at least for peers and subordinates). There has been a demonstration of moderately positive correlations with objective criteria such as productivity, salary change and promotions.
Conway and Huffcut (1997) have found that composite ratings were found to maximise reliabilities. Conway and Huffcut concluded that each source demonstrated some evidence of validity. Viswesvaran et al (1996) meta-analytically summarised supervisor and peer reliabilities. For overall performance ratings (either a single rating or a composite) they found mean single rater reliabilities of .52 for supervisors and .42 for peers.

Saavedra and Kwun (1993) are worth citing in some detail on the role of peer ratings. Like Cleveland and Murphy (1995) these authors argue that peer ratings can make a unique contribution to performance appraisal. Particularly in contexts where self managing work groups are the main form of work organisation. These authors argue that peers are possibly the best-informed source of information on performance appraisal ratings because they are able to observe each others' task behaviours and interpe rsonal behaviours. Westerman (1996) cites research to the effect that peers provide information on different performance dimensions than do traditional supervisory ratings. According to Westerman, peers are more differentiate effort from performance and focus on task-relevant abilities and competencies.

Norton (1992) conducted a meta-analysis of studies of peer ratings. Norton (1992) made the following comment: "It appears that, given certain favourable contextual factors, peer assessments of ability and/or performance may show substantial validity. Much of the observed variance in the validity of peer assessments may actually be due to statistical artifacts (such as sampling error and imperfect reliability) and other less than ideal measurement conditions. …. Based on the above resu lts, it appears that peer assessments are most informative (and hence most valid) when (a) peers have had sufficient time to become familiar with each other's work-related qualifications, skills and abilities, (b) objectively measured criteria will be used for validation, and (c) peers believe that the data they provide will only be used for research purposes, as opposed to impacting upon practical decisions that affect the individual being rated. The author concluded that even when ratings were conducted f or administrative purposes, the size of validity correlation was still impressive.

There are limitations associated with peer ratings. In their study of peer assessments made by business graduate students, Saavedra and Kwun (1993) mention that generally peer ratings are not well accepted by raters or ratees except when they are used for developmental purposes.

Also group members will typically underrate peers who are more capable than average members and overrate those members who are less capable. Saavedra and Kwun (1993) also note that peer ratings tend to be less accurate than other peer assessment methods such as peer rankings and peer nominations. The question that Saavedra and Kwun explored in their 1993 article was whether adjustments in ratings were due to incomplete knowledge of the performance distribution (i.e. an inaccurate anchoring point) or whether it reflected self interest (i.e. adjustment processes that enhance own standing). Saavedra and Kwun (1993) found that a group members performance level influenced his or her rating of peers. Below average contributors and average contributors were less discriminating in their ratings of peers, and it was inferred that their anchoring and adjustment processes were biased. Outstanding contributors were more discriminating evaluators than average or below average contributors even when the evaluato r asked for inclusion of a self rating. For this reason, Saavedra and Kwun (1993) included the instruction for raters to allocate performance appraisal scores among group members including the raters themselves.

Saavedra and Kwun (1993) also mention that peer ratings provide a psychometric advantage when they are pooled in that they increase reliability and partially remove the idiosyncratic biases of any particular rater. Irrespective of the rater's perspective, Kulik and Ambrose (1992) suggest that both self and supervisory raters prefer the multiple comparison standard because it incorporates more points of view allowing for a more comprehensive evaluation.

Peer appraisals may offer other procedural justice benefits as well. Of the seven procedural elements used y individuals to form evaluations of outcome fairness, five may be positively influenced by use of peer appraisals: setting ground rules, change mechanisms, gathering information, safeguards and selection of agents. Peer appraisals that allow groups to stipulate the dimensions of performance that are relevant criteria for appraisal, grant employees voice in the setting of ground rules, and t his may have the added effect of increasing the validity of the appraisal, (Westerman, 1996).

The next question is how many raters there should be and under what conditions should ratings take place. Antonioni (1996) argued that contributors to the rating process should remain anonymous. Results from Antonioni's survey showed that appraisers who are accountable give higher ratings than appraisers who are anonymous. Conversely managers who were being appraised by the 360-degree process wanted their raters to be accountable. Antonioni (1996) argued that to ensure anonymity, there should be five or more direct contributing raters of any single individual. Antonioni (1996) noted however that the organisations in the research program lowered their requirements to three raters, (including peers). Maurer, Raju and Collins (1998) compared peer and subordinate ratings and they found that the most notable difference between subordinate and peer ratings occurs on the threshold between the highest and second highest categories. Peer raters seemed to have slightly higher standards; it took more underlyi ng skill of a manager's part to receive a rating in the highest category.

Diary Maintenance

Diary keeping is regarded as a useful, but infrequently used device in performance appraisal systems, (Maroney and Buckley, 1992). DeNisi and Peters (1996) commented that unstructured information makes it inherently difficult for raters to accurately rate subordinates even if they wished to do so. Raters who maintained diaries were less prone to error than those who did not maintain diaries. It was found that diary keeping led to better memory organisation, which in turn promoted better recall an d ratings. These raters were also more confident in their ratings, than raters who did not maintain diaries.

There has been some discussion over the manner in which diaries should be structured. When diary information is structured by task, then raters would store information according to tasks, (DeNisi, Robbins and Cafferty, 1989). Conversely, when diaries are organised by "person" more information is recalled overall. Raters who organised their material on a "person" basis were found to recall a lower percentage of positive information and a greater percentage of descriptive inform ation, (DeNisi and Peters, 1996).

It seems that organising diaries in terms of people improved raters' ability to recall and differentiate examples of effective and ineffective performance exhibited by a given ratee. DeNisi (1989) also found that there was support for diary maintenance as a means of keeping structured information in memory that allows for more accurate recall and improved ratings. These findings occurred in the presence of an inbasket task that placed competing demands on raters' attention, highlighting the benef its of diary keeping even during the short duration of the experiment.

Maroney and Buckley (1992) also recommended recording exemplary incidents that cover the range of behaviours in each dimension. These authors also comment that although employees may not exhibit equal numbers of positive and negative behaviours, a set number of recordings in each dimension are still recommended, in order that any one dimension may not be over-represented. They suggest this on the basis that it will be easier to detect rater bias, as it is also unlikely that all an employees' beha viours will be either positively or negatively skewed for all dimensions. Maroney and Buckley (1992) recommendations presuppose that the performance appraisal system will be based on critical incident methodology. The difficulty of this approach is that it can lead to an exemption of lesser but still none the less important mundane job performance behaviours that are necessary for every day performance of the job. For this reason it is preferable to structure diaries in the context of behaviourally anchored rating scales, where a predesignated number of incidents exist for each dimension within a set period of time. Also that incidents are recorded according to a distribution that has been witnessed for each employee. DeNisi (1989) also found that given choice raters' preferred to organise their diaries in terms of people when they were given a choice.

It is also important that diary keeping be undertaken in conjunction with rater training. Bernadin and Beatty (1984) found that diary keeping promoted greater inter-rater agreement, and that diary keeping was less subject to leniency and halo effect when raters had been trained.

Training and Rater Error

Armstrong (1998) argues that both managers and employees should be trained in the skills required for successful performance management. This author discusses the need for training which incorporates the skills of objective setting, providing feedback, coaching and counselling. Both parties in the performance management process would require guidance and training in the use of competences, the preparation of performance agreements and plans, the preparation for and conduct of performance reviews, ratings and the completion of review forms.

Training should cover the following areas:

1. The aims and objectives of performance management;
2. The sequence of activities that will take place
3. How to carry out or participate in the following processes
4. Agreeing key tasks
5. Setting objectives
6. Agreeing skill, knowledge and competence requirements
7. Reviewing performance on a continuing basis
8. Providing feedback
9. Counselling and coaching
10. Preparing work and development plans

Armstrong recommends developing a series of training modules, as it would be difficult for trainees to absorb more than one module at a time. It is also suggested that the modules be presented in two programs, the first program would incorporate modules 1 to 4, whilst the second program would incorporate modules 5 to 8.

The modules could consist of:

1. Introduction to performance management
2. Defining key tasks and objective setting
3. Working with attributes and competences - analysis and measurement
4. Preparing work and development plans
5. Conducting review meetings
6. Rating
7. Providing feedback
8. Coaching and counselling

On the specifics of coaching training, Antonioni (1996) has argued that training for coaches (whether they be consultants or supervisors) should include active listening, focused interviewing, selecting targeted improvement areas, dealing with feelings and reactions to unexpected negative feedback, setting specific goals and action plans, follow-up, and ways of shaping the development of new behavioural habits.

The research on performance appraisal training has mainly focused on the efficacy of different types of performance appraisal rater training in mitigating rater error. Woehr and Huffcut's (1994) review of performance appraisal training indicated that a combination of different training types was required. These authors found that frame of reference training (FOR) followed by behavioural observation training was most effective in terms of reducing rater error. The effectiveness of FOR training was also demonstrated in a study by Stamoulis and Hauenstein (1993). The primary goal of FOR training is for raters to share and use common conceptualisations of performance when making evaluations. It has been hypothesised by Woehr (1994) that FOR training will be used by raters as the basis for information processing in the evaluation process and will be demonstrated through improved ratings using different indices of accuracy, (i.e. differential elevation, differential accuracy). The comments of Arvey and M urphy (1998) are quite germaine. These authors note that little effort has been made in the studies, which evaluate the reasons for the efficacy of FOR training to directly measure the information processing that hypothetically goes on as a result of the training.

Feedback

Fletcher (1991) has commented that feedback is an essential element if effective learning is to take place.
It is well established within the literature that employees respond more positively to participatory styles of feedback. DeGregorio and Fisher (1988) compared several participative formats, and found that these approaches tended to be more likely to result in more positive subordinate perceptions than unilateral methods of feedback. No one particular participative method was considered consistently superior, and appraisal methods did not have a differential impact on post feedback performance. Thi s is in contrast to earlier research which established that critical, unilateral methods will actually increase defensiveness and reduce subsequent subordinate performance, (Kay, Meyer and French, 1965). There is also a disadvantage to supervisors being the sole source of feedback, as the information provided by supervisors can be seen as irrelevant or error prone. Greller (1980) discovered that subordinates placed more credence in information provided to them by sources that were central to the employees' performance. These sources included the task, co-workers and self-assessments.

Antonioni (1996) argues that most appraisees need someone they can openly discuss their appraisal results with. The appraisee's reaction to feedback will determine whether he or she makes constructive use of the feedback. Thus it is important for appraisees and their coaches to be aware of the range of possible individual reactions. Antonioni specifies two dimensions which can influence the range of reaction1) the appraiser's desire for changes in the appraisee's work performance 2) the extent to which feedback from the performance appraisal system has come as a surprise to the appraisee.

Reactions to type one feedback are positive because self-appraisal is congruent with positive feedback. Reactions to type two feedback are also positive because the appraisee received positive feedback that they themselves were not expecting. Reactions to type three feedback are usually neutral because the appraisee expected and agreed with the low 360-degree feedback. This area is definitely labeled as an area for improvement. Reactions to unexpected negative feedback (type 4) can vary from conf usion to defensiveness.

Unexpected negative feedback is most difficult for appraisee's to accept and they will require more time to process such feedback. The manner in which feedback is delivered is also critical. Negative feedback should only be delivered after the appraisee has been given a full opportunity to present their position and views. If this opportunity is not presented, the appraisee's defence mechanisms will cause them to discount any negative feedback on the grounds that the rater has not accepted all th e information pertaining to their performance.

The role of feedback is also important in predicting the possible direction of ratings. Shore, Adams and Tashchian (1998) found that the identified purpose of the rating process had a significant impact on the direction of the rations. The identified purpose also had an interaction with the system of accountabilities attached to the performance appraisal system. There were two conditions of identified purpose in Shore et al 's study. The first condition was developmental for the subordinate's per formance, the second condition was administrative (i.e. would the subordinate be recommended for a research position on campus).

These authors found that raters with knowledge of high self assessment by a subordinate rated the subordinate significantly more positively than did those rating subordinates with a reported low self assessment. The subjects inflated ratings relative to actual performance levels when presented with a high self assessment rating, but assigned performance ratings closer to the actual performance when given access to low (accurate) self assessed ratings. In other words, managers can soften the ratin g to be more congruent with the subordinate's self-perception, so as to avoid a negative confrontation.

However, when ratings are to be fed back to a higher organisational agent (such as senior managers) then ratings are less likely to be inflated. This was particularly the case when raters were led to believe that their decisive would result in administrative outcomes, (i.e. that employment of the subordinate in a research role).

Section IV: Conclusions

It is clear that the practice of performance appraisal has much to gain from being informed by the psychological literature in this area. Irrespective of whether a performance appraisal system relies on personality measures or task behaviour measures, the performance appraisal system must be based upon a job analysis. All job analysis methodologies rely on a comprehensive documentation of the tasks performed by job incumbents. The inclusion of skills, attributes, and requisite personality traits depends on whether the developers of the performance appraisal system wish to conceive of the job domain in terms of both job context and job content. A sound job analysis is necessary to ensure that performance indicators are relevant and representative of the job domain. This initial step is also crucial from the point of ensuring that the performance appraisal system meets legal and legislative requirements, and is seen as being a fair and unbiased measure of an employee's performance.

In the past, proponents of performance appraisal decried the application of personality measures in predicting and managing employee performance. It was argued that personality measures did not provide information that was sufficiently specific in creating behavioural change. It is noteworthy that available literature on personality measures of performance have not yet addressed the role of feedback and coaching for employees. It is also noteworthy that the literature on personality measures of p erformance does not address the possibility of mutliple values systems and subcultures existing within a single workplace or organisation. This has particular implications for Australian workplaces which have a strong history of industrial militancy, where there is a pragmatic requirement to accommodate additional cultural values systems beside those values which are espoused by management. Indeed the most of the research on personality measures of performance has been conducted on white collar sales workfo rces or in a military context. Such workplaces do not have an institutional history of competing power bases outside of the managerial chain of command. It is therefore relatively simple to establish a uniform set of cultural values which express what is considered to be contextual "good performance".

Definition of organisational citizenship behaviours or pro-social behaviours becomes more complex when there is a contradiction between awards, custom and practice and management ideology. This contradiction in turn affects the relevance and utility of personality measures which may rely exclusively on a managerial definition of pro-organisational citizenship behaviours. There was also a suggestion that personality measures may be more vulnerable to manipulation from subordinates who practice &qu ot;impression management behaviours" with their supervisors and superiors.

The need for multiple raters in the performance appraisal process was also promoted in the present paper. It is quite evident from available research that no single ratings source was particularly valid. Validity of performance appraisal ratings was greatly enhanced when a system of combined supervisory, peer and self ratings was used. Each ratings source was seen to contribute a different type of information, and mutiple ratings were a useful device to counter rating error stemming from inadequa te opportunity to observe the appraisee. This may also be a useful device to counter possible bias which could occur in situations where there has been prolonged and bitter industrial disputation. Multi-rater systems are a structural counter to accusations of management or peer bias.

Diary keeping and the structure of diaries was considered by this paper. This paper highlighted the value of maintaining contemporaneous notes from a procedural justice perspective and with a view to reducing rater bias. Diaries can also be structured to promoted balanced recall of events. This is significant as unstructured diaries can have the effect of encouraging raters to overlook more mundane, but no less essential job behaviours. It was noted that diary keeping became particularly effectiv e in reducing rater bais when diaries in maintained in conjunction with rater bias training. Training was presented as a key support feature in the performance appraisal process.

The material outlined in this paper indicated that specific training modules need to be developed for each step of the performance appraisal process. It is not sufficient to provide training that simply explains a performance appraisal system to all the participants. It is equally crucial for ratings accuracy, that raters be trained to adopt a specific frame of reference and be cognisant of the different types of bias.

Reference List

Algera, J.A. and Greuter, M.A.M. Job Analysis, Chapter 7, Personnel Psychology: Handbook of Work and Organisational Psychology, 1998. Eds. Pieter Drenth, Henk Thierry, Charles de Wolff. 2nd Edition.

Armstrong, M. (1998). Performance Management. Kogan Page, London.

Antioni, D. (1996). Designing an effective 360-degree appraisal feedback process. Organisational Dynamics. 25(2), 24-38.

Arvey, R.D. and Murphy, K.R. (1998) Performance Evaluation in Work Settings. Annual Review of Psychology, 49, 141 - 168

Bernadin, H. J. and Buckley, M.R. (1981) Strategies in Rater Training, Academy of Management Review, 6, 205-221.

Bernadin, H. J. and Beatty, R. W. (1984) Performance appraisal: Assessing human behaviour at work. Boston: PWS-Kent.

Bobrow, W. and Leonards, J.S. (1997) Development and Validation of an assessment centre during organisational change. Journal of Social Behaviour and Personality, 12 (5), 217 - 236

Conway, J. Understanding method variance in multi-trait multi-rater performance appraisal matrices: examples using general impressions and interpersonal affect as measured method factors.

Conway, J.M. and Huffcut, A.I. (1997) Psychometric Properties of Multisource Performance Ratings: A Meta-Analysis of Subordinate, Supervisor, Peer and Self-Ratings. Human Performance, Vol 10(4), 331-360

Cronbach, S.F. (1998). Job Analysis: Changing Nature of Work. Canadian Psychology, 39 (1-2), 5-13.

Day, D.V., Schleicher, D.J. and Unckless, A.L. (1996). Self monitoring and work-related outcomes: A meta-analysis. Paper presented at the 11th Annual Conference of the Society of Industrial and Organisational Psychology, San Diego, CA.

DeGregorio, M.B. and Fisher, C.D. (1988) Providing performance feedback: Reactions to alternate methods. Journal of Management, 14 (4), 605 - 616

DeNisi, A. S. and Peters, L.H. (1996) Organisation of Information in Memory and the Performance Appraisal Process: Evidence from the field. Journal of Applied Psychology, Vol 81 (6), 717 - 737

DeNisi, A.S., Robbins, T. and Cafferty, T.P. (1989). The organisation of information used for performance appraisals: The role of diary keeping. Journal of Applied Psychology, 74, 124-129.

Drenth, P. J. D. (1998) Personnel Appraisal, Chapter 4, Personnel Psychology: Handbook of Work and Organisational Psychology, Eds. Pieter Drenth, Henk Thierry, Charles de Wolff. 2nd Edition

Fletcher, C. (1991). Assessment, feedback counselling and individual differences: an exploratory study. Counselling Psychology Quarterly, 4 (2/3), 103 - 107

Greller, M. M. (1980). Evaluation of feedback sources as a function of role and organisational level. Journal of Applied Psychology, 65, 24-27

Hattrup, K., O'Connell, M.S. and Wingate, P.H. (1998) Prediction of multi-dimensional criteria: distinguishing task and contextual performance. Human Performance, 11 (4), 305 - 319.

Hogan, J., Rybicki, S.L., Motowidlo, S.J. and Borman, W.C. (1998). Relations between contextual performance, personality and occupational advancement. Human Performance, 11 (2/3), 189 - 207.

Hogan, R. and Shelton, D. (1998) A Socioanalytic Perspective on Job Performance. Human Performance, 11 (2/3), 129 - 144.

Kay, E., Meyer, H.H. and French, JR. (1965) Effects of threat in a performance appraisal interview. Journal of applied psychology, 49, 311 - 317

Lowry, P.E. (1997) The Assessment Centre Process: New Directions. Journal of Social Behaviour and Personality. Vol 12, No. 5, 53 - 63

Maddux, R.B. (1993) Effective Performance Appraisals: A Practical Guide for more productive and positive performance appraisals. Crisp Publications, Inc. Menlo Park, California.

Maroney, B.P. and Buckley, R. (1992), Does Research in Performance Appraisal Influence the Practice of Performance Appraisal? Regretfully not! Public Personnel Management, Vol 21 (2), 185 - 196

Maurer , T.J. and Tarulli, B.A. (1997) Managerial Work, Job Analysis and Holland's RIASEC Vocational Environmental Dimensions. Journal of Vocational Behaviour, 50, 365-381

Maurer, T.J., Raju, N.S. and Collins, W.C. (1998). Peer and subordinate performance appraisal measurement equivalence. Journal of Applied Psychology, 83 (5), 693 - 702.

Mount, M.K., Barrick, M.R. and Stewart, G.L. (1998) Five factor model of personality and performance in jobs involving interpersonal interactions. Human Performance, 11 (2/3), 145 - 165.

Murphy, K.R. and Cleveland, J.N. (1995). Understanding Performance Appraisal: Social, Organisational and Goal Based Perspectives. Thousand Oaks, CA, Sage.

Nowack, K.M., (1997) Congruence between self -other ratings and assessment centre performance. From Riggio, R.E. and Mayes, B.T. (Eds.) Assessment Centres: Research and applications. Journal of Social Behaviour and Personality, 12 , (5), 145-166.

Norton, S.M. (1992). Peer assessments of performance and ability: An exploratory meta-analysis of statistical artifacts and contextual moderators. Journal of Business and Psychology 6 (3), 387 - 399.

Saavedra, R. and Kwun, S.K. (1993). Peer evaluation in self managing work groups. Journal of Applied Psychology, 78 (3); 450 462.

Schrader , B.W. and Steiner, D.D. (1996). Common comparison standards: an approach to improving agreement between self and supervisory performance ratings. Journal of Applied Psychology, 81 (6), 813 - 820.

Shore, T.H., Adams, J.S. and Tashchian, (1998) Effects of self appraisal information, appraisal purpose, and feedback target on performance appraisal ratings, 12 (3), 283 - 298

Stamoulis, D.T. and Hauenstein, N. M. A. (1993). Rater training and rater accuracy: Training for dimensional accuracy versus training for ratee differentiation. Journal of Applied Psychology, Vol 78 (6), pp 994 - 1003.

Sulsky, L.M. and Keown, J.L. (1998). Performance appraisal in the changing world of work: Implications for the meaning and measurement of work performance. Canadian Psychology, 39 (1-2), 52 - 59.

Tornow, W.W. (1993), Perceptions or reality? Is multi-perspective measurement a means or an end? Human Resource Management. 32: 2211-2229.

Varma, A., Denisi, A.S. and Peters, L.H. Interpersonal Affect and Performance Appraisal: A field study. Personnel Psychology, 49, 341 - 360

Viswesvaran, C. Ones, D.S. and Schmidt, F.L. (1996). Comparative analysis of the reliability of job performance ratings. Journal of Applied Psychology, 81(5); 557-574.

Warech, M.A., Smither, J.W., Reilly, R.R., Millsap, R.E., Reilly, S.P.. Self-Monitoring and 360-Degree Ratings. Leadership Quarterly, 9(4), 449 -473.

Wayne, S.J. and Liden R.C. (1995). Effects of impression management on performance ratings: A longitudinal study. Academy of Management Journal, 38, 232-260.

Westerman, J. W. (1996). Rethinking the role of performance appraisal in total quality management: An argument for the use of peers as raters. Employee Responsibilities and Rights Journal, 9 (4), 272 - 284

Williams, R.S. (1998). Performance Management: Perspectives on employee performance. International Thomson Business Press. London.

Woehr, D.J. and Huffcut, A.I. (1994). Rater training for performance appraisal: A quantitative review. Journal of Occupational and Organisational Psychology, 67, 189 - 205