Part 3: Evaluating the Selection System

You have now created a selection system for the job of a teller.

The final step in a selection system is to make sure the system works properly.

As discussed in the overview, there are two ways to ensure that the system is working. One approach takes a legal perspective to ensure that organizations do not discriminate in hiring.

There are two types of discrimination: disparate treatment and disparate impact (also known as adverse impact). Disparate treatment discrimination refers to treating applicants differently based on a protected characteristic (for example, age, sex, national origin, religion). An example of disparate treatment discrimination is not considering women for leadership positions. This type of discrimination is considered intentional and therefore easy to identify and correct or prevent.

Disparate impact discrimination is considered unintentional. This form of discrimination indicates that all applicants were treated equally; however, this equal treatment had an unequal effect related to a protected characteristic. The most common approach to identify adverse impact is to apply the four-fifths rule. The four- fifths rule states that adverse impact exists if the selection ratio of the minority group is less than four-fifths (or 80 percent) of the selection ratio of the majority group. A selection ratio is the percentage of those hired based on the percentage of those who applied for the job. Selection ratios must be calculated for each protected group. The selection ratio of the minority group is compared with the selection ratio of the majority group (often "males" or "Caucasians"). The simplest way to calculate adverse impact is to divide the selection ratio of the minority group by the selection ratio of the majority group. If the result is less than 80%, then adverse impact exists.

For example, the bank collected the following data over the past five years:

Males applied=200 Males hired=40

Females applied=300 Females hired=45

Based on this information, the selection ratio for men is 20% (40/200), whereas the selection ratio for women is 15% (45/300). Dividing the minority group (the group with the lower selection ratio, women) by the majority group (in this case, men) results in an answer of 75% (15%/20%). Since the result is less than 80%, adverse impact exists. The organization needs to explore the selection process to identify what may be the cause of this disparity.

In this part of the exercise, you will conduct this analysis.

The bank compiled selection data on three racial groups during the past year: Caucasians, African-Americans and Latinos. The data is as follows:

	Number applied	Number hired
Caucasians	90	27
African-Americans	50	10
Latinos	40	10

1. 1. The selection ratios for the three groups are: Caucasians = African-Americans = Latinos =
2. 2. Does adverse impact exist when you compare the African-American applicant pool with the Caucasian applicant pool? Show your work.
3. 3. Does adverse impact exist when you compare the Latino applicant pool with the Caucasian applicant pool? Show your work.
4. Another way to assess the effectiveness of the selection system is to examine the decision-making accuracy. Employee selection is basically a system to predict which applicants will succeed on the job. Highly qualified applicants are expected to perform well on the job. Similarly, seemingly less qualified applicants are not expected to perform well on the job. Unfortunately, the hiring process is complex. Some applicants can put on a "good show" and still be mediocre employees, whereas other applicants may interview poorly but still be excellent employees. Organizations cannot assume the selection process is working well without collecting and analyzing data.
5. One option to calculate the decision-making accuracy of a selection system is to classify applicants into one of two categories: strong applicants (ones who appear to be highly qualified and are predicted to do the job wella good hire) and weak applicants (ones who do not appear to be highly qualified and are predicted to not do the job wella poor hire). Once hired and after sufficient time (typically six months to a year) has passed, employees can also be classified into two categories: good hire or poor hire. After all of this data is collected, the overall decision-making accuracy of the selection process (the total hit ratio) can be calculated as the percentage of correct predictions.
6. The table below shows data about the 200 tellers employed at the bank.

   1. 	Poor hire	Good hire	Totals
      Strong applicant	20	80	100
      Weak applicant	70	30	100

      1. Calculate the total hit ratio by adding the correct predictions and then dividing that number by the total number of decisions made. What is this percentage? Do you think this percentage is impressive?

   2. What percentage of weak applicants turned out to be good hires?

   3. What percentage of strong applicants turned out to be good hires? This percentage is known as the positive hit ratio.
   4. Compare your answers to questions 2 and 3. Do you think the bank's system is effective?

P.S. It is all one part...