Behaviorally Anchored Rating Scales
(BARS)

Behaviorally Anchored Rating Scales(BARS)
BARS refers to Behaviorally Anchored Rating Scales. It was developed by Smith and Kendall to provide a better method of rating employees. It differs from "standard" rating scales in one central respect, in that it focuses on behaviors that are determined to be important for completing a job task or doing the job properly, rather than looking at more general employee characteristics (e.g. personality, vague work habits).
Behaviorally Anchored Rating Scales (BARS) are designed to bring the benefits of both qualitative and quantitative data to the employee appraisal process. BARS compare an individual’s performance against specific examples of behavior that are anchored to numerical ratings. For example, a level four rating for a nurse may require them to show sympathy to patients while a level six rating may require them to show higher levels of empathy and ensure this comes across in all dealings with the patient.
There are different ways in effectively structuring a BARS system. They are
1. Identification of Critical Incidents: This step involves supervisors and other in authority identifying specific incidents of effective and unsatisfactory behavior. These incidents should be related to the job performance.
2. Selection of performance dimensions: These incidents are further dissected into performance dimensions. An average of 5 to 7 performance dimensions is ideal. This classification is based primarily on the Critical incidents selected.
3. Retranslation of the incidents: Another group of participants who are familiar with the job are asked to reclassify the critical incidents initially selected. This is done to safeguard against individual bias. There needs to be an agreement of 75% of the panel for the incident to be included in the rating procedure.
4. Assignment of values to the incidents: The individual incidents are now rated based on the performance dimensions. The rating scale usually ranges from 1 to 7. A rating of 1 indicates an ineffective and unsatisfactory performance, while a rating of 7 indicates an effective or satisfactory performance. The ratings thus arrived at are then subjected to mean and standard deviations. Incidents which have standard deviations of 1.5 or less find a place in the final scales.
5. Developing the Final rating document: The behavioral anchor for the ultimate performance dimensions is the subset of the incidents selected. The BARS in its final form is composed of a set of vertical scales and the corresponding incidents which endorse them.
While this system appears to be logical and with inherent safeguards against the effects of bias and preconception, it has its own share of advantages and disadvantages. The advantages are that since the ratings are by people who are familiar with the working environment they have a greater chance of being accurate. This also involves participation by the workers to a considerable degree. It is also specific to the particular job function and rates behavior that can be observed easily. However, critics point out that despite its intuitive appearance, the system has some inherent downsides. The designing of effective BARS is time-consuming in the first place. The behaviors used for rating are more activity oriented rather than result oriented. This leads to a situation where the supervisors will have to give a higher rating to employees who are just involved in an activity, but do not show results.
In utilizing Behaviorally Anchored Rating Scales the following process needs to be undertaken.
1. Generation of expert panels. Two panels of experts who due to their knowledge and experience in the area of study, are able to design an instrument to assess quality performance. 2. Designing a questionnaire based on a 5 point Likert scale of competencies required to measure performance.
3. Validating the instrument by sending the questionnaire to a wide sample.
4. Analysis of the results.

Advantages: * These types of rating scales are particularly effective for assessing competencies, skills and abilities. * BARS rating scales are highly valid and job-related because important job requirements are covered. * Objective benchmarks are provided against which observations can be rated, therefore, there is less rating error than when using other types of scales (e.g. numeric).
Cautions:
* BARS scales take some time and effort to create and usually cannot be used for job types other than those for which they were developed.
Developing a BARS Scale * Using subject matter experts, identify examples of job performance behaviors reflecting all different levels of effectiveness ranging from ineffective to superior for all the different parts of the job. These are key indicators only. It is not required to be an exhaustive list of every possible criterion. You may have already identified these behaviors when you conducted your job analysis to establish the qualifications and competencies. If so, use them here. * Examples are then clustered by content and categories of job performance and ranked according to importance. * Major, essential, or core criteria are distinguished from those that are minor or secondary within the group. * Tip: Focus on the extreme ends of each range (i.e. the 5 and the 0-1 points) and describe them fully first OR focus on describing the 3 or 4 Point passing answer first, then add or subtract to define other answers. * Decide how many points will be awarded and how irrelevant or incorrect responses will be scored. A wrong answer should result in a failing grade. * Assign marks that reflect the relative importance of the question and the competency being assessed. * The scale may be multiplied by a factor to increase the overall weighting. For example, a 5-point scale is multiplied by a factor of 4 to increase the weight of the assessment to 20 points in the overall competition.

Are Behaviorally Anchored Rating Scales Superior to More General Approaches?
As with most things related to performance management the success of a BARS (behaviorally anchored rating system) approach to employee reviews depends on how well the system is implemented. In theory, a BARS system, if properly implemented should result in fairer, and more accurate assessments of employee performance. In theory, they are indeed better than more vague rating systems where it's hard to get any two people to agree on what a particular rating item means.
However, BARS still involves RATINGS, and ratings still have inherent flaws, the most notable being that ratings themselves (let's say assigning some number to "reflect" performance) are not very helpful in helping employees improve performance because too much information is lost. Another problem is that there is a tendency for people to believe that BARS system ratings are objective, and that is definitely not the case. Ratings cannot, by definition, be objective, because they involved labelling and generalizations.
The issue with BARS is that it requires extensive upfront analysis necessary to identify the job behaviors. Without that, it's no improvement. And, if you are going to invest all that time into the upfront analysis, maybe it makes more sense to use that data without a rating system per se, or use it in a way similar to how goals and objectives are used in a Management by Objectives system (MBO).

-------------------------------------------------
References
Human Resource Management by Gary Dessler
Smith & Kendall (1963) Re translation of expectations: An approach to the construction of unambiguous anchors to rating scales. Journal of Applied Psychology