APPRAISALS BASED ON MERE SAMPLES OF WORK
One of the more notorious games of an unfair supervisor is manipulating the work selected to review for a performance appraisal. Those supervisors can set an employee up for great success or heartbreaking failure merely by how they choose the sample of the employee’s work. Consequently, unions need to focus on how work is chosen for review.
MSPB has given federal employees a little help in the fight against manipulated samples, particularly where work is numerically measured, e.g., by percentages, number of errors, or peer group counts.
While an agency is not required to provide an accounting of every item of work in order to prove unacceptable performance with respect to a critical element with a percent requirement, the agency must, at a minimum, establish some methodology for selecting the examples of alleged unacceptable performance so that a reasonable person might conclude that the appellant’s performance fell below the critical element’s percent standard. Ryerson, 35 MSPR 123.
However, the Board has required the agency, at a minimum, to establish some objective, systematic method for selecting examples of alleged unacceptable performance so that a reasonable person might conclude that the employee’s performance fell below the percentage standard. Bowling,47 MSPR 379
“Objective and systematic” are criteria that should be in every labor-management contract article dealing with performance appraisals based on samples of work. While the Board does not mandate that the samples be “random” in order to meet the “objective and systematic” criteria, it has repeatedly pointed to that as a way to satisfy those standards.
In reaching this conclusion, we note that, under some circumstances, an agency may establish that an appellant failed to satisfy a percent requirement for a critical element by examining a random sample of the appellant’s work, rather than all of the work performed during the period in question. . . . In the present case, however, the agency did not establish that the documents it examined constituted a random sample of appellant’s work. Player, 32 MSPR 448.
The Board has also mentioned the need for “sufficient” samples. Although we will not go into the details here, suffice it to say that not only must a sample be randomly chosen to meet the commonly accepted rules of systematic numerical analysis, but it also must be large enough. (But, if you feel up to it, we recommend you take a look at one site that offers a free calculator for determining what the size of a sample of work should be. If that is confusing, start with this one. )
We find that it is reasonable to require a similar method in Chapter 75 cases that implicate performance standards allowing a percentage of errors. . . . Because the agency has not shown that it established a systematic method for evaluating the appellant’s performance based on a sufficient sample of the appellant’s work, we find that the agency has failed to prove its charge against the appellant. Thus, we conclude that the agency has failed to support its removal of the appellant by a preponderance of the evidence. Bowling, 47 MSPR 379.
So, if a member walks into the union office with a complaint about her appraisal and that appraisal measures her work numerically, one of the first things you want to ask the supervisor is how he drew the sample of the work he reviewed to develop the count. When he answers that, ask about how large his sample was compared to how much work the employee did.
If you are going to the bargaining table, try real hard to get the concepts of objective, systematic, random, and sufficient written into the agreement. You can still demand management meet them if they are not mentioned in the contract, but it never hurts to put the standards right out there where managers can see them.