7. Passing Scores Applicant Sex Score Omar M 98 Eric M 80 Mia F 70 (passing score) Morris M 69 Tammy F 58 Drew M 40
8.
9.
10.
11.
12.
13.
14.
15.
16.
17.
18.
19.
20.
21.
22.
23.
24.
25.
26.
27.
28.
29.
30.
31.
32.
33.
34.
Editor's Notes
Show Slide 5-2 We will now consider the three approaches to making selection decisions: Unadjusted top-down selection, passing scores and banding. As we discuss each, think about the one that is most consistent with your organizations selection system. We will first discuss the top-down approach.
Show Slide 5-3 Selection decisions using top-down selection are very straightforward. Top-down selection assumes that those with higher test scores will be better workers. Test scores are literally indicative of performance. You simply choose the applicants with the highest test scores. The basic idea here is that if a test is criterion valid, it will have a linear relationship with performance. This means that applicants with higher scores would be expected to perform better on the job, than applicants with lower scores. The way this works is that applicants are ranked from highest to lowest based on their test scores and you select the employee with the highest scores and move down until all the openings are filled.
Show Slide 5-4 Advantages: The major advantage with this approach is that it results in higher utility. In other words, one would expect a better quality group of selected applicants (Aamodt, 1999). Disadvantages: Top-down selection has some serious disadvantages. For example, this approach allows for little flexibility in decision making, when it may be desirable to use to use nontest factors such as references and organizational fit. When the top-down approach is used, the most valid selection method (cognitive ability) typically results in the highest levels of adverse impact against protected groups. Tests that result in subgroup differences may be illegal if not defensible. Less workforce diversity. Increased likelihood that selection procedure would be viewed as unfair. Strict rank-ordering ignores measurement error. It assumes that your test score alone accounts for all the variance in performance (Zedeck, Cascio, Goldstein & Outtz, 1996). We will now talk about the use of passing scores in making selection decisions.
Show Slide 5-5 One common way of making selection decisions is to classify applicants into two groups: a high scoring group and a low scoring group (Livingston & Zieky, 1982). A passing score will accomplish this task . A passing score is a point in a distribution of scores that distinguishes acceptable from unacceptable performance (Kane, 1994). Uniform Guidelines (1978) Section 5H: Passing scores should “normally be set so as to be reasonable and consistent with normal expectations of acceptable proficiency within the workforce.”
Show Slide 5-6 The organization would have to determine the lowest score on the test that is associated with acceptable performance or minimum competency. Once you set your passing score, you can fill the openings with any applicant who scores at or above that level. So what do you think? Are minimum standards acceptable? What could go wrong?
Show Slide 5-18 Advantages: Passing scores have been used to reduce adverse impact and increase workforce diversity (Biddle & Sill, 1999). The use of passing scores also allows for flexibility in reaching affirmative action goals or any other goals established by the organization. Passing scores allow for the greatest flexibility of any selection method. Disadvantages: The major disadvantage with this approach is that quality of selected applicants may be lower, therefore reducing the utility of the selection device (Biddle & Sill, 1999). Two decision errors: An unqualified person may get a score above the passing score and a qualified person may get a score below the passing score (Livingston & Zieky, 1982). By establishing a passing score, you are ignoring a lot of the variability around the middle of the score distribution. To reduce the chances of making these types of errors, you can establish an “uncertain” category for people that fall in the middle category. For these people, a pass/fail decision would only be made after additional information is obtained. The “uncertain” category is useful only when additional information can be obtained.
Show Slide 5-20 Top-Down (a band of one) Essentially, top-down bands are a band of one. Rules of “three” of “five” Rules of “three or five” allow a supervisor to choose from the top 3 or 5 of the highest scores. Traditional bands Traditional banding has been used for decades and is based on expert judgment. Expectancy bands These bands are based on predicted performance in reference to some normative group. SEM bands SEM banding is based on statistical significance testing, in which you are testing whether scores are significantly different from each other. With this approach you are placing a zone around a group of scores, and treating all scores that fall within that zone as statistically equivalent. Pass/fail bands Pass/fail bands are based on scores and not the number of openings. The amount of flexibility you can use in making decisions progressively increases as you go from top-down bands to pass/fail bands.
Show Slide 5-27 There are several choices of SEM bands. Each has its strengths and weakness. Choice of band type will depend on organizational goals and culture (Zedeck, Cascio, Goldstein & Outtz, 1996). Fixed bands: Here, the initial band is set in reference to the highest raw score. For example, if the highest raw score is 100 and the band is 10 points wide, the band will be set from 100 to 90. Applicants will be selected from within the band until all positions are filled (Murphy, Osten & Myors, 1995). Sliding bands: Width of the band remains the same, but the referent score is now the top highest that remains after the first hiring decision is made. In other words, a new band is established every time the highest scorer is hired (eliminated) from the applicant pool. For example, let’s say the bandwidth is 10 points and the applicant with highest score (100) in a band is hired. The next highest score is 98. The bandwidth (10) would then be subtracted from the next highest score (98). The band will now slide from 98 to 88. As you can see, the band slides through the distribution every time a decision is made, allowing for the inclusion of lower scoring applicants. This would not be possible with fixed bands (Zedeck, Cascio, Goldstein and Outtz, 1996). Diversity-based bands: Females and minorities are given preference when selecting from within a band. Preferential selection only applies to those who fall within the band. Which do you prefer?
Show Slide 5-31 Research suggests that banding: Can help reduce adverse impact, increase workforce diversity, and increase perceptions of fairness, with little sacrifice in utility. Banding also gives you more flexibility and autonomy in making selection decisions. It allows you to consider other factors that are relevant to the job (Zedeck, Cascio, Goldstein & Outtz, 1996). Secondary criteria Nontest factors are called secondary criteria, and they include things such as job experience, education, training, attendance records, seniority, residency requirements, special skills and organizational fit. For example, it is sometimes more desirable to hire a lower scoring applicant with computer skills or who is bilingual over a higher scoring applicant without such skills. With banding one can select equally qualified people that otherwise would have been rejected. Using secondary criteria, some argue that the job performance of those selected using the top-down method is not greater than that of those selected using banding (Campion et al, 2001). As you can see, banding forces manager’s to consider other factors in the decision making process. However, it is suggested that secondary criteria be evaluated using an objective scoring system in which each factor is weighed according to its relevance to its criterion (Campion et al, 2001).
Show Slide 5-32 There may be scores that fall below the band that actually belong in the band. Job performance within a band is not truly equal (e.g. between the top and low scores within a band). By grouping scores, you may lose valuable information. For example, wide bands may in fact result in utility loss. Banding creates the impression that reducing adverse impact is more important than reliability and validity. Sliding bands are cumbersome to administer, especially in the private sector where employment volume is continuous and banding requires that you adjust selection rule based who has already been selected. The score indifference may compel organizations to hire lower quality people. Finally, banding without minority preference may not reduce adverse impact. Further, its usefulness in achieving affirmative action goals is affected by the percentage of minority applicants (Campion et al., 2001).
Show Slide 5-33 Narrow bands are preferred: Narrow bands may be more desirable in jobs that require a select number of qualified applicants. With narrower bands, you can be more confident that there is less difference in performance. Consequences or errors in selection: The greater the consequences of hiring a poor performer, the narrower the band needs to be (e.g. police officer). In jobs where most applicants are qualified and there is little difference between them, a larger band may be used (e.g. garbage collector). Criterion space covered by selection device: Most tests tap no more than 10% of the criterion space (Campion et al, 2001). This is because the predictor is not fully representative of the criterion. Banding allows you to consider other factors not measured by the test, but indicated on job analyses. It allows managers to take a multidimensional view of performance. Reliability of selection device Selection tests are not perfect. The lower the reliability of the test, the more confident you can be that differences between applicants are also unreliable. Validity evidence If a test has low reliability, it is less likely to be valid. This places the inferences or decisions made based on test scores in doubt. Diversity issues Banding can be thought of as a way to incorporate social/legal responsibilities into the decision making process.
Show Slide 5-35 What the organization should do to protect itself: The company should have established rules and procedures for making choices within a band. These should be applied consistently and be outlined in company policy or in employee manuals. Applicants should be informed about the use and logic behind banding in addition to organizational values and objectives. This information can be presented at some point during the selection process. Now that we have discussed how to make selection decisions. It is time to take action. In the following section we will discuss some considerations in conveying rejections and making offers.
