Based on German law (§ 630 BGB or § 109 of the Commercial Code) employers are legally obligated to provide work references for former employees, regardless of the reasons they left the company. So an estimated 2.58 million references are written in Germany (Huesmann, 2008). Polls have shown that references are still considered an important, commonly used tool in the staff selection processes.
Studies show that more than 200 German companies attribute a weighting of 34.8% to references in deciding whether or not to invite applicants for further interviews. Whereas the cover letter only accounts for 16.7% and the CV for 25.3% of the weighting. Similarly, in 2012, an estimated 34% of German recruiters stated that the reference is of great importance in candidate selection (Online Survey, “deutschen Bildung”, 2012).However scientific and academic studies on the importance of references say very little, as very few actually exist. Those that do exist tend to show a devastatingly small or negative significance of references on candidate suitability. Where does this discrepancy between practice and theory arise from, how references can cause problems in the selection process, and how they can nevertheless be used, is explained here.
Sources of Error in a Reference
Any German employee has a right to a job reference. Thus, a work certificate typically includes information about the position, employee related tasks, an evaluation of performance, and social behavior. The author of the reference is committed to being honest and writing in goodwill, which prohibits most negative characterizations of the employees. Why this is one of the biggest problems in terms of aptitude diagnostics, we will discuss later. As a selection diagnostic tool, references are used particularly in the process of pre-selection. It seems that the benefits often stated by HR managers are more often than not outweighed by the drawbacks. References are not reliable or valid instruments and have high error rates and mistakes in regards to candidate selection (see Table 1). This vulnerability to produce errors undermines the validity of job references in many ways.
Table 1. Possible Errors in Reference Letters (Weuster, 2012)
Scientifically testing Recruiter Concealment Techniques
In Table 1 we outlined various sources of error. Of these, the most serious problem is the “secret code” used by HR managers. Because employers are not allowed to clearly describe negative aspects of an employee’s performance, a secret code has been established. Interpreting this code can, at times, be compared to deciphering hieroglyphics, rather than assessing a diagnostically reliable aptitude testing tool. Employers are forced to use methods of encryption though, relying on the receiver to read between the lines. Some of these techniques are:
- The Vacancy Technique: Anything negative is omitted – the reader should therefore attach particular importance to what is NOT stated
- The Discrepancy Technique: Negatives are expressed through hidden contradictions, “After a long initial familiarization phase, Mr. X worked with a quick grasp on new information.”
- The Positive Scale Technique: Positive formulations are graded, ranging from the meaning “very good” to “very bad”, but formulated only in positive terms.
Although there isn’t necessarily a lot of psychological research on references, the “The Positive Scale” has already been looked at a bit closer. According to a German research team led by Huesmann (2008): The written expression “to our full satisfaction” was meant in code as “satisfactory” in a reference. When the reference was judged by laypeople and professional HR managers, it was read as meaning anything from “Very Good” to “Unsatisfactory”. The reader may be anxiously wondering to what extent is writing a work reference even still a good thing? To gather more specific scientific information regarding references, the study used a contrasting view. Written references in the U.S.A. are “free”, meaning they are not coded. The study looked at the significance of free references and used the following combinations:
- One person judges and writes references for two employees (1-to-2 Combination)
- Two people judge and write references for one employee (2-to-1 Combination)
If reference letters were reliable and valid, one would expect a high degree of discernment in combination 1 and a high level of agreement in combination 2. The results of the study however show that instead: free references are both indistinctive and inconsistent and offer no accurate and reliable information on the candidates’ suitability.
One of the largest meta-analytic studies of the validity of aptitude diagnostics methods was carried out in 1998 by Schmidt and Hunter. In this study written references were certified as having a significant predictive validity of .23 to .26, but under severe restrictions. According to Schmidt and Hunter work references may well be significant. However, the following points must be observed. Psychometric quality can be obtained if both author and recipient understand and are capable of fully using the “secret code.” To get around the other sources of error, a good solution may be to simply call the former employer to discuss the reference and find out: Did the author personally see or know about the work and performance of the employee? How familiar was the author with the “secret code.” Another way to interpret references, lies in the comparison of several references. It should be noted, however, that only limited conclusions can be drawn and only when it comes to similar activities and industries.
To conclude, considering the validity and usefulness of references it could be said that: work references are associated with a myriad of problems. Job references can provide information in the context of aptitude diagnostics selection processes. To do this however, a number of important conditions first have to be met. The most important thing is probably that author and recipient both have knowledge of the “secret code” used for references. The fact that other errors can also occur, such as observer error, should be mitigated by applying various precautions such as structured assessment systems and using objective criteria in the evaluation.
