Pay-for-Performance

Rating Scale Disadvantages

We've covered several variations of rating scales and discussed how each of the variations differ from one another. Now let's examine some common disadvantages of rating scales:

• commonality
• performance factors
• total score
• weighting
• errors
• number of scales

Commonality

Rating scales require managers and employees to agree about the meaning of factors and their degrees. If there is no such agreement, the 'standard' is an illusion and becomes a disadvantage.

Performance factors

The advantage of commonality can also be seen as a disadvantage if in fact there is a great deal of difference among the performance factors required in different jobs. Different jobs may really require that different factors be used. This is a basic argument for the use of BARS.

Total score

The summation of a number of factors always assumes that a deficiency in one can be made up for by strength in others. Where this is not true, the summated scores are not useful.

Weighting

Weighting is a problem in rating scales. If factors overlap in the behavioral domain that they measure, then some dimensions of performance may be inadvertently over weighted.

Since the actual weighting of the scales may be quite different from the weights specified when the system was developed, weights should be applied after the ratings have been completed and checked statistically.

Errors

The most common criticisms of rating scales, particularly graphic rating scales, are the set of constant errors that occur in rating. These errors include:

• degree
• central tendency
• halo effect
• proximity

Degree. Degree errors occur when everyone is rated too leniently or too severely.

Central tendency. In central tendency errors, the rater overuses the middle of the scale making it hard to distinguish among employees.

Halo effect. The halo effect bias refers to a tendency for people to form a generalized positive impression of an employee, and rate the employee highly on all rating criteria even if the employee really doesn't deserve a high rating for every item. In other words, the manager thinks highly of an employee and rather than rate independently for each item, the manager inflates the ratings.

Proximity. Proximity can occur when raters assume that because 2 factors are next to each other on the ratings form, they are likely to correlate. Logical errors occur when the rater assumes that two factors are similar and should therefore be rated the same.

Number of scales

A criticism from BARS advocates is that using a single rating scale over a large number of jobs is not useful, since the behavioral dimensions of jobs differ greatly. This is a dilemma. While the criticism is valid, the solution – that of having many different scales – presents other problems. It is expensive and time-consuming to develop a series of rating scales and employing many scales can also make it more difficult to directly compare employees in different jobs.

End Result

Despite all of these criticisms of rating scales, the Human Resource Department must be able to use rating scales in a pay-for-performance program, since this is the most common performance appraisal program used by organizations. Where there are common factors and a summated score, the task is easy, if not always accurate.

All employees can be placed upon a single ranking, making the performance axis complete. If constant errors are prevalent in the rating process, the responses may need to be statistically spread out on the scale. The best results can be obtained if:

• unambiguous descriptions of factors and degrees are developed
• evaluations are NOT shared with the employee being rated
• raters have been trained