A popular trend for about the past twenty years has been for organizations to adopt competency models as a basis for their performance management and other human resources systems. One advantage of competency models is that they communicate what is important to the organization’s leadership and drive performance in desired areas. For example, if an organizational goal is to improve customer service, then including a Customer Service competency not only communicates that this is a critical success factor but also focuses attention and rewards on this aspect of performance. Competency models usually include a comprehensive array of factors associated with success technical, leadership, interpersonal, and personal.
While competency models have been enormously popular and adopted by many organizations, there has been debate and lack of clear defnition about what a competency is :
• Ability factor ?
• Skill factor ?
• Knowledge factor ?
• Personality factor ?
• Performance factor ?
• All of the above ?
Here, competencies are defined as the knowledge, skills, abilities, and other personal characteristics that are most instrumental for achieving important job outcomes and contributing to organizational success.
Many organizations have used fairly informal processes to develop their competencies, such as interviewing organizational leaders or convening focus groups to discuss important success factors, rather than using rigorous job analytic approaches such as those. If competency measures are used primarily for development, using informal processes to develop them is unlikely to yield significant negative consequences. However, if competency measures are used as the basis for human resources decisions, more formal and rigorous procedures for demonstrating their job relevance are prudent to employ. A suggested approach for doing this is outlined following discussion of two important questions that first need to be addressed :
•
How many competencies should be developed ?
•
Should highly job specific or more generic performance standards be developed for different jobs ?
How Many Competencies ?
One advantage that competency models offer is that they provide a foundation for developing integrated human resources systems, such as staffing, training, promotion, succession planning, and performance management systems. However, the number and specifi city of the competencies needed depend on their intended use, for example :
•
To make entry level staffing decisions, broad ability and personality competencies are typically developed and assessed, such as critical thinking, conscientiousness, and interpersonal skills
•
If the purpose is staffing for a job that requires highly specialized skills, more specifi c technical competencies need to be identifi ed and measured
•
For training and career development, competencies need to include more general capabilities (communication skills, planning skills) as well as specialized technical knowledge and skills
•
For performance management purposes, higher-level competencies are typically used that reflect the major performance requirements of a job Organizations usually identify between five and higher level competencies that are linked to their strategic objectives and critical success factors.
While novice developers of performance management systems are sometimes tempted to include a large number of very specific competencies in their systems, rating and providing feedback on a large number of competencies take a considerable amount of time. Systems that contain too many competencies are not viewed positively or as practical by managers with many direct reports to evaluate. In addition, the number of competencies selected for performance management purposes is much less important than ensuring that those selected are well defined and reflect the most critical aspects of the job.
The challenge in developing competency models to serve multiple purposes concerns the need for job information that is both specific and general. A way to address this challenge is to create hierarchical competency models, where more detailed competencies are categorized under more general competencies. Lower-level competencies that relate to the “Communicating with Others” competency appear in the example. The advantage of this approach is that more specific competencies are portrayed as interrelated instances of broader competencies within the context of a job. This provides an organized, understandable foundation of detailed information that is needed for staffing and training of specialized competencies, while also providing higher level competencies that are more practical for performance management purposes.
For performance management purposes, experienced practitioners agree that competencies should be defined in terms of performance standards that help managers differentiate between employees who are performing more or less effectively than others. However, decisions need to be made about how many different sets of performance standards should be developed and how customized these should be. There is no one best approach, as there are advantages and disadvantages to different options. There are three key decisions that need to be made :
•
How much customization for different jobs ? Will performance standards be developed that can be applied across all jobs or will they be customized to reflect the specific content of individual jobs ?
•
How many effectiveness levels ? Within each competency, how many levels of performance effectiveness will be defined and what will these be (e.g., below expectations, meets expectations, exceeds expectations) ?
•
How many experience levels ?
Will one set of performance standards be developed that applies across all experience levels or should separate standards be developed for different levels of work (e.g., entry-level, full performance, expert) ?
How Much Customization for Different Jobs ? Sometimes fairly general performance standards are developed for each competency that can be used across all of the jobs within an organization. In this scenario, technical aspects of the job are often handled by having a “Technical Proficiency” competency, also defined in a general manner that is applicable across jobs.
An alternative to using common standards that apply across jobs is to customize the competencies to some degree to reflect the specific requirements of individual jobs. Several examples of how this has been done appear below.
If common competencies and more general standards are used that apply across all jobs, a potential downside is that some organizational members may have difficulty seeing their specific jobs in the standards, and they may react less favorably to these than more customized standards. As a result, managers may need to translate more generally written standards into more specific expectations for a given employee’s job. On the other hand, a practical advantage of using common standards across jobs is that the development time and resources are significantly less than what is needed to develop customized competencies and performance standards for different jobs. Using common standards across jobs also results in communicating more consistent expectations to organizational members, which can be important if one goal of the competency model is to drive performance in certain strategic areas. The decision about how much to customize the competencies and performance standards depends on three things :
•
The goals the organization is trying to achieve with its performance management system
•
The resources available for developing and maintaining the system
•
What will be viewed as acceptable and effective to organizational members
How Many Effectiveness Levels ?
To anchor managers’ judgments of performance, it is important to define standards for different levels of effectiveness within each competency.
Obviously, performance standards could be defined for more or fewer effectiveness levels. Typically, at least two sets of performance standards, describing what it takes to “Meet” and “Exceed” expectations are defined. Sometimes, standards are also written for minimal or below expectations performance. It is rare, however, to see performance standards written for more than three effectiveness levels. The reason is because any standards written need to clearly and unambiguously describe the performance expectations at each level of effectiveness. When one attempts to write standards for too many different effectiveness levels, the descriptions start melding together and it is difficult to see clear and unambiguous distinctions between them. When there are not clear distinctions between the standards for different effectiveness levels, managers have difficulty systematically applying them as well as explaining to employees the rationale for one rating versus another. For this reason, it is recommended that performance standards be defined for no more than three effectiveness levels.
If performance information will be used for decision making, a numerical rating scale should be used in conjunction with the performance standards. A five or seven point scale is used most often because it provides a sufficient number of rating points to differentiate between employees in terms of their performance. A rating scale of less than five points does not usually provide adequate differentiation among employees, while a rating scale of more than seven points adds complexity without significantly more differentiation. With a fi vepoint scale, three ratings are commonly used, from 3 (meets expectations) to 5 (exceeds expectations). With a seven point scale, four ratings are commonly used, from 4 (meets expectations) to 7 (exceeds expectations). The choice of whether to use a five or seven point scale is personal preference.
Those who prefer a five point scale say . . .
•
More than five rating points is too many, and it’s not possible to make such fine grained performance distinctions.
•
They want to keep the performance management system as simple as possible, and a seven point scale is overly complex.
Those who prefer a seven point scale say . . .
•
The availability of seven rating points is important to spread out the ratings as much as possible, especially given managers’ tendencies to use the higher rating points.
Irrespective of the number of rating scale points, the relationship between these and the performance standards needs to be clearly articulated so that raters can apply them in a uniform and fair manner. Ratings based on numerical scales can easily be averaged or summed across competencies to derive a summary score for decision making.
How Many Job Levels ?
While organizations can choose to define more or less customized standards for different types of jobs, it is not recommended that the same standards be used for different job levels. Because employees at different job levels are paid differently, based on their knowledge, responsibility, and contributions, performance standards need to reflect the different levels of complexity, difficulty, and independence that characterize these.
Similar to the discussion above regarding definition of different effectiveness levels, standards for different job levels need to clearly and unambiguously describe the performance expectations at each level. For progression purposes, it is also important to clearly articulate how the performance expectations change as one moves through different job levels. Thus, the standards for different experience levels should focus on :
•
Differences in the tasks and activities performed at different levels
•
Differences in the complexity of the work at different levels
•
Differences in the amount of supervision at different levels
•
Any other concrete factor that describes how the work is changes at the different levels
The differences between job levels should not simply be about doing the same work more or less effectively. Finally, it is not sufficient to say that work is “complex” at one level and “highly complex” at the next level, because everyone will have different interpretations of what this means. Rather, it is important to describe what makes the work more complex from one level to the next.
One caveat in writing standards for different job levels is that there may not be different expectations for each and every competency.
Certainly, there should be differences for many, if not most, of the competencies showing the higher level responsibilities that accompany higher level salaries and expectations at different job levels. However, for things like integrity, initiative and motivation, and security consciousness, the same high standards are usually expected, irrespective of the job level.
It is important to think through how many job levels should be defined. If one attempts to write behavioral standards for too many levels, the descriptions start overlapping and it is difficult to see clear distinctions between them. Many organizations have a large number of job levels, sometimes as many as 15 or more, where there are overlapping requirements and unclear distinctions between them. Not only are systems with this many levels difficult to manage efficiently, but it is extremely difficult to articulate clear performance standards for this many levels, as well as multiple effectiveness levels within them. Prior to developing a new performance management system, it may be necessary to define fewer job levels so that the performance requirements for these levels are distinct and clear. Typically, anywhere from three job levels (entry level, full performance, and manager) to a maximum of six levels (entry level, full performance, expert, team leader, supervisor, and manager) are defined.
Certainly, there should be differences for many, if not most, of the competencies showing the higher level responsibilities that accompany higher level salaries and expectations at different job levels. However, for things like integrity, initiative and motivation, and security consciousness, the same high standards are usually expected, irrespective of the job level.
It is important to think through how many job levels should be defined. If one attempts to write behavioral standards for too many levels, the descriptions start overlapping and it is difficult to see clear distinctions between them. Many organizations have a large number of job levels, sometimes as many as 15 or more, where there are overlapping requirements and unclear distinctions between them. Not only are systems with this many levels difficult to manage efficiently, but it is extremely difficult to articulate clear performance standards for this many levels, as well as multiple effectiveness levels within them. Prior to developing a new performance management system, it may be necessary to define fewer job levels so that the performance requirements for these levels are distinct and clear. Typically, anywhere from three job levels (entry level, full performance, and manager) to a maximum of six levels (entry level, full performance, expert, team leader, supervisor, and manager) are defined.
As a practical matter, careful consideration needs to be given to the implications of the decisions regarding the number of performance standards to be developed. For example, if one chooses to define six job levels with three performance effectiveness levels for 12 competencies that are customized for 20 separate jobs, over 4,000 unique performance standards would need to be written. The resources and time it would take to accomplish this may be prohibitive, in which case it would be practical to use more general standards that can be applied across jobs. Performance standards that can be applied across jobs for three effectiveness levels and three experience levels are shown in the examples below.
An Efficient Approach to Developing Competencies and Performance Standards
Job analysis techniques, such as job observations, interviews, focus groups, and surveys are an effective means for identifying key competencies. To demonstrate that a performance measure is job relevant or content valid, it is necessary to show that it assesses performance on work behaviors that are important for performing the job. While there are several approaches for developing valid competency models, an effective and efficient process for defining content valid competencies and performance standards consists of three steps :
•
Step 1 :
Develop Competencies Defined by Important Work Behaviors
•
Step 2 :
Validate Competencies
• Step 3 :
Develop and Validate Performance Standards
Develop Competencies Defined by Important Work Behaviors
The first step is to become familiar with the target jobs by reviewing organizational and occupation related documentation and talking with senior management for its perspective about the key competencies that are required for success. In this first step, decisions will need to be made about the general number of competencies to develop and whether common competencies and performance standards will be developed to apply to all jobs or be customized for different jobs. Decisions will also need to be made about how many experience levels and effectiveness levels to define. For purposes of illustrating the approach, assume the decision has been made to develop :
•
Eight competencies and performance standards that will be applied across all jobs
•
Two levels of effectiveness (meets expectations and exceeds expectation) for each of three experience levels (entry level, full performance level, and first level manager).
Although these are the goals, the development and data collection processes discussed next will determine whether these goals are possible. For example, it may be that the exact same competencies and performance standards are not relevant (or valid) for all jobs.
Based on information collected from the organization’s leadership and the review of job materials, a preliminary set of competencies is identified and a general definition is written for each. Each competency is then further defined by approximately five to eight work behaviors. Articulating how the competencies are manifested in observable work behaviors is important for establishing the job relevance of the model. While the number of work behaviors used to define a competency can vary, five to eight is recommended for this approach. Work behaviors are written at a higher level of generality than task statements so that they will apply across jobs rather than be highly job specific. Also, performance standards will eventually be written reflecting the content of each work behavior. If there are too many very specific behaviors defining a competency, it becomes difficult to write performance standards that are straightforward and efficient for managers to use. Having performance standards that are clear, uncomplicated, and easy to use is essential for gaining buy-in from users of the system. As a practical matter, it is much more important to cover the essential job content efficiently and effectively than to have highly detailed, cumbersome, and complicated standards. Three sample work behaviors for the Planning Work competency are shown in the example.
•
Are the competencies comprehensive and correct for the jobs in question ?
•
Are the work behaviors defining each competency complete and correct for the jobs in question ?
•
What wording changes are revisions are needed to the competency definitions or work behaviors to make them maximally relevant for the jobs ?
Collectively, the job experts attending the focus group meeting should have sufficient knowledge of all of the occupations under consideration so that they can comment on the relevance of the competencies and work behaviors and revise them, if necessary. Participants should be selected to be representative in terms of occupation, location, race, gender, age (over 40), and any other factors that may yield differences in perceptions about the job requirements. Six to 10 participants are typically included in each focus group.
The number of focus groups needed depends on the size of the job incumbent pool and the number of factors that need to be represented (e.g., race, gender, different locations, shifts, etc.). For many jobs, two workshops with a total of 12–15 participants are sufficient to review draft competency and work behavior lists. Obviously, if competencies and work behaviors are being developed to apply across jobs, workshops need to be conducted with job experts who can collectively speak to all of the jobs. The end product of Step 1 is a set of competencies, each defined by approximately five to eight work behaviors that have been reviewed and edited by job experts.
Step 2 :
Validate Competency Model
To validate the competencies and work behaviors, a survey of job experts is recommended where they are asked to rate the importance of the work behaviors for performing their jobs. Usually, job incumbents are surveyed, although supervisors of the target jobs can also participate in addition to or instead of job incumbents. For validation purposes, it is sufficient to draw a representative sample of respondents to complete the survey. However, to promote buy in from the workforce, a better strategy is to allow any organizational member who will be covered by the system to complete a survey. This is especially important when a new system is threatening to employees or will involve something new, such as pay tied to performance when this has not been the case in the past. Drawing a sample of respondents is referred to as a sample based survey. Allowing all affected organizational members to participate is called a census survey.
The survey data provide a basis for demonstrating that the competencies and work behaviors are job-related, thus addressing both professional and legal standards. The competencies and work behaviors are usually rated on a five-point importance scale, ranging from extremely important to minor importance. A response option is also provided for indicating that a work behavior is not relevant to the job. Work behaviors that receive an average rating across respondents of “Important” or greater are generally considered to be sufficiently important to be retained. The idea is that it is not appropriate to hold employees accountable for or evaluate their performance on a work behavior that is not an important job requirement. Another criterion that is sometimes used in conjunction with the average importance rating is that not more than 25 % of respondents report that a work behavior is Not Relevant for their jobs. In the example below, only the work behaviors are rated for importance and not the competency overall. This is because the competency is defined by the work behaviors. However, if desired, an importance rating of the competency itself can also be collected in the job analysis survey.
In situations where the competencies and work behaviors will be applied across jobs, it is important to analyze the data separately for the respondents within each job. Likewise, if standards will be written for different experience levels, it is also important to analyze the data separately for respondents who have these different levels of experience. This ensures that the competencies and work behaviors on which the performance standards will be based reflect requirements that are important for each job and level of experience within that job. Sometimes, for example, a work behavior may only be important for employees with higher levels of experience but not at the entry level. Or, a work behavior may apply to only supervisors and managers but not non supervisory employees. Likewise, although the work behaviors were designed to apply across jobs, some may not be important for certain jobs. If the work behaviors and competencies cannot be shown to be relevant and sufficiently important for each job and each experience level within a job, then slightly different competency models and performance standards will need to be selectively used based on the survey results. Under these circumstances, it may be necessary to remove an entire competency or perhaps one or two work behaviors defining a competency for some jobs or experience levels within those jobs.
An example will help to clarify how the survey data are used. Shown in the example are partial job analysis results for full performance employees in six different jobs. For a work behavior to be retained, we will use two criteria :
•
The behavior must have an average importance rating of at least 3.0 (“Important”)
•
No more than 25% of respondents can report that the behavior is “Not Relevant” for their job
Using these criteria to interpret the data in the example below :
•
One behavior is not valid for Support Jobs, Plans, prioritizes, and balances projects and/or assignments to accomplish work in a timely manner. This behavior will need to be excluded from the competency model and performance standards for this job.
•
Because all three behaviors are not valid for Maintenance Jobs, the entire Planning Work competency will need to be excluded for these jobs.
Step 3 :
Develop and Confirm Appropriateness of Performance Standards
The development of standards to guide the performance ratings is an important component of a transparent, effective, and fair performance management system. Performance standards need to be written that describe how important work behaviors are performed at different levels of effectiveness. For each work behavior, a standard is written describing the type of performance that would be considered meeting expectations, while a higher level standard would be written describing the type of performance that would be considered exceeding expectations. Performance standards should be written to be decidedly results oriented, thus reinforcing the importance of achieving important work outcomes. To incorporate a results focus, the performance standards should not simply describe desirable behaviors, but they should also describe the expected outcomes from these behaviors.
Typically, initial performance standards are written by Human Resources professionals. To refine the standards, workshops are conducted with representative job experts, who review and comment on the standards and suggest additions, deletions, or revisions. Identification of participants for these workshops needs to follow the same guidelines for representation discussed previously for reviewing the draft competencies and work behaviors.
To validate the standards, another set of workshops is held with job experts who did not participate in the first set of workshops. Participants in these workshops are asked to make ratings regarding whether or not the standards comprehensively measure the important work behaviors. Although the standards were specifically written to reflect different levels of effectiveness in performing the work behaviors, it is possible that some aspects of the work behaviors may have been lost in developing the standards. Accordingly, these ratings help to ensure that the integrity of the work behaviors was retained, supporting the job relevance of the performance standards and their use as a basis for making human resources decisions.
Workshop participants are also asked to rate the extent to which they feel that the standards accurately describe “Meets” and “Exceeds” performance expectations for the job and experience levels (entry, full performance, and manager). Since this group of job experts is specifically asked to judge whether the standards appropriately reflect the job requirements for different experience levels, supervisors are usually used because they have knowledge about how the performance expectations change at different job levels.
Different people use different criteria for how much agreement is needed among respondents on the preceding exercise to feel comfortable with the performance standards, ranging from a minimum of 60 % to a maximum of 90 %. A reasonable level of agreement is 70 %. That is, at least 70 % of respondents should independently agree that each work behavior is covered by its intended performance standards. At least 70 % of respondents should also agree that the standards accurately describe the performance expectations at the different effectiveness and experience levels.
•
If less than 70 % agree that a behavior is covered by its target competency, this indicates that there is probably redundancy or overlap between the standards for different competencies that might be better differentiated.
•
If less than 70 % agree that the standards properly describe the expected performance for the different experience and effectiveness levels, this should be further investigated to understand which standards are problematic and revisions should be made accordingly.
Weighting Competencies
Managers and employees frequently jump to the conclusion that competencies should be weighted in deriving an overall rating score. The rationale is that they feel different competencies are more important to overall job success than others. For example, technically oriented competencies are viewed by some managers as more critical than softer skill competencies, like teamwork and communication. The job analysis data can be used to examine whether there are significant differences in the importance of different competencies and thus if there is any justification for weighting them differently. In most cases, the importance ratings of the competencies that are critical for a job tend to be similar and do not provide support for differential weighting.
Some managers want to go even further than weighting the competencies differently overall. They want to separately weight the competencies for each individual employee to reflect what is most important for that employee’s specific work. For development purposes, there is no problem if managers discuss competencies they feel are more or less important for a given employee. However, if performance ratings will be used for decision making, managers should not idiosyncratically weight competencies for each employee, as this would be akin to holding employees to different standards without justification for doing so.
Contrary to what many intuitively think, the most compelling reason not to differentially weight competencies is that ratings based on weighted and unweighted competencies usually result in the same rank order of employees. If the ratings were used as a basis for decision-making, what this means is that the same decisions would be made about whom to pay more or promote irrespective of whether or not the competencies were differentially weighted. However, so strong is the belief that weighting competencies will make a difference that it is often necessary to perform demonstration studies with real ratings to show that weighted and unweighted competency ratings yield the same rank order of employees.
No comments:
Post a Comment
Note: Only a member of this blog may post a comment.