Topic 5

Selection
HRM561

Hello! This week, we will learn about selection. In the previous module, we covered
recruitment. So selection is about making a choice of whom to hire and whom not to
hire among the recruited candidates. Like any other human decisions, hiring decision
is also subject to multiple errors and biases. To minimize these errors and biases,
companies should use valid assessments. This is critical not only for defending the
organization legally but also for identifying best qualified candidates. So this week’s
lecture contains concepts such as validity and reliability of different selection tools,
which provides solid scientific evidence for the choice of selection tools.

1
 Selection : Hiring the Best

u Selection has been argued to be the single most important factor

affecting organizations
u Legal concerns can play an important role particularly in
selection because selection is affected by a number of legal
constraints
u Firms should ensure selection techniques are job-related
u Selection tools as predictors of job performance
u Assessment to identify good hires and screen out poor hires

Selection begins with the candidates identified through recruitment. Selection

procedures based only on the “gut” impressions of hiring managers can be a legal
liability for employers. And research shows that this approach generally results in
poor hiring decisions. However, selection procedures based on systematic data
collection and analysis not only can result in more legally defensible hiring procedures
but also lead to better hiring decisions. So selection has been known to be the single
most important factor affecting organizations. The effectiveness of the selection
process heavily depends on the ability to gather necessary and accurate information
and to make a best prediction about job candidates’ future performance.

2
 Quality of Selection Tools

Ø Reliability : Consistency or similarity in the sets of scores as

a result of the measurement of an attribute
Ø Validity: The extent to which the technique measures the
intended knowledge, skill, or ability
Ø A selection technique that is not valid is useless and may
even present legal problems
Ø When discrimination in hiring practices is charged, the
critical evidence will be the job relatedness (validity) of
the selection technique

Some selection procedures are of higher quality than others. Two concepts come into
play in determining this. Validity and reliability are the essential ingredients for an
effective and legal selection tool.

3
 Validity
Accuracy of measurement

Selection Attributes
Test (e.g., Verbal &
(e.g., Cognitive ability test) Math ability)

e.g., Irrelevant material on e.g., Important skills but

a test, Interviewees are not assessed in the test
asked different questions

Validity is the concept relating to how accurately a selection method predicts actual
job performance. For example, we can use test scores of applicants and calculate
correlations between the scores and performance of the applicants who were hired.
When selection tests include irrelevant content but do not include core skills, the
results from the tests will not be useful in predicting the applicant’s future job
performance. Validity is the most important factor in considering whether or not to
use an assessment method, because an assessment that does not accurately identify
who will perform effectively on a job has no value to the organization. In this ban
diagram, The larger the overlap here, the more accurate. In other words, more valid
the performance measure is.

4
 Reliability : Consistency across different
raters or different times
e.g., if interviewers’
ratings are not
consistent à
low reliability of the
interview

e.g., if personality
test yields similar
results at times à
high reliability

Reliability refers to consistency over time and across different raters. For instance, a
reliable personality test should generate /consistent results if the same person takes
the test several times. MBTI personality assessment has been found to be lacking
reliability. Some of you might have done this assessment and received different
results every time you took it. Because this test has low reliability, it is not
appropriate to use MBTI as a selection tool. It also lacks validity because evidence
shows that MBTI is not significantly related to job performance. So HR professionals
and managers should look at information about validity and reliability before deciding
on selection methods. Depending on the skills and knowledge required to perform a
job, valid and reliable selection tools should be used and this information should
come from job analysis.

5
 If this is a question only asked to women, it
would be an example of sex discrimination.
How old are your
Given its irrelevance to the hiring process and
children?
potential discriminatory effects, this question is
best avoided.

You have an This question could reveal the national origin of

Improper/illegal
interesting accent.
Where did you grow
the applicant, which is a protected category.
Interview
up?

Have you ever been Statistically, Hispanic and Black men are more
Questions
arrested? likely to have been arrested, and arrest record
is not a reliable indicator of crime. The use of
this criterion could have a disparate impact.

Do you have a Even with good intent, it is best to avoid this

disability that requires question. If the employee has a disability, the
an accommodation request should come from the
accommodation? employee. It is acceptable to ask whether the
individual is able to perform the major functions
of the job.
Are you married? If only asked to women, it may be an example
of sex discrimination under Title VII. The
organization may be trying to screen out
female applicants who are likely to become
pregnant, violating pregnancy discrimination.

One of the most common personnel selection tools is interviews. Interviews are
where biases and stereotyping most easily show up. How old are your children?
Although this question sounds casual and personal, if it was only asked to women, it
would be an example of sex discrimination. Have you ever been arrested?
Statistically, Hispanic and Black men are more likely to have been arrested, and arrest
record is not a reliable indicator of crime.

6
 Types of Structured Interviews
— Behavioral interviews: using information about what the applicant
has done in the past to predict future behaviors
— Situational interviews: asking people how they might react to
hypothetical situations e.g., “suppose,” “what if”
— Case interviews: give candidates a situation, problem, or challenge
and ask him
— All three outperform unstructured interviews and result in scores
that can be used to compare candidates

Continuing on interviews, we’ve learned some interview questions can lead to biased
decisions. Then, how can we better use interviews? Is there a better type of
interviews? Yes, research suggests that structured interviews are better, in other
words, they are more valid than unstructured interviews, what that means is that
when you use structured interviews, you can do a better job predicting new hires’
later job performance. I'm going to jump in and start telling you about different types
of interviews. Structured interviews mean asking the same questions of all
interviewees. On the other hand, unstructured interviews are more like random
individual conversations that do not necessarily cover all the same questions with
every candidate. Therefore, the ability of unstructured interviews to predict job
performance vary because of inconsistency across applicants and interviewer biases.
This is why unstructured interviews have low reliability and validity. However, all
three structured interviews you see on the slide have better validity than
unstructured interviews. Among them, you are probably familiar with behavioral
interviews. Behavioral interviews use information about what the applicant has done
in the past to predict future behaviors. Behavioral interviews are a standard part of
Google and Ford’s selection procedures.

7
 Example:
Structured
Interview

So what does a structure interview question look like? The example question on the slide is asking
about interpersonal relations or conflict management style. Now, there is a rating scale for the
question, low, moderate and high. The rating scale is ideally predetermined before the interview
and these ratings and description for each level of the rating should be defined for each question.
So the question and the rating scales reflect the type of skills and behaviors the interviewers are
looking for. So they are trying to identify using the questions and the rating scales. So where can
you find the rating scales? You can develop rating scales using some information from a job
analysis. So you can identify high performers’ behavior, their skillsets and credentials, they can be
profiled as highly productive employees’ competencies and traits. Rating scales also provide
interviewers with a consistent framework for rating job candidates in the interview scene. This is
why structured interviews have high validity. In addition to asking a same set of questions, having
a well-defined rating scale can illustrate poor, average, and high levels of skills and behavior for
each requirement. This can help to increase the predictive power of interviews. To make
interviews most effective, interviewers also need to be trained on how to conduct an interview
using this type of questionnaires and rating scales. And we need to make sure all interviewers use
the same procedures. Behavioral interviews use information about what the applicant has done
in the past to predict future behaviors. Now, the issue here is that does people’s past
performance always predict 100% their future performance? Probably not. But, the point here is
that structured interview does a better job predicting employee future performance and behavior
relatively, compared to unstructured interviews.

8
 Assessment
Center

Next selection tool I like to tell you about is assessment center. Please note that
although assessment centers sound like a facility that physically exists but they
are just one of the selection techniques comprising multiple tasks such as role-
play, tests, group discussions, case studies and so on. Assessment centers are
often held over multiple days, generally for managerial positions or management
trainees. I wanted to bring your attention to this as this is just a type of selection
method that uses a set of simulated tasks or exercises that candidates are asked
to perform.

9
 Comparison of Selection Methods
Assessment Method Average Applicant Relative Costs Adverse Usability
Validity Reactions (Development/Administratio Impact
n)
Assessment centers .37 Good High/High Low Difficult
Cognitive ability tests .51 Good Low/Low High Easy
Integrity tests .41 Low Low/Low Low Easy * Validity values range
from –1 to 1, with
Job knowledge tests .48 Good High/Low Low Easy
numbers closer to –1 or
Reference checks .26 Good Low/Low Low Easy
+1 reflecting better
Situational judgment .34 Good High/Low Low Moderat prediction of job
tests e
performance
Structured interviews .51–.63 Good High/High Mixed Moderat
e
Unstructured interviews .20–.38 OK Low/High Mixed Easy
Personality testing –.13–.33 Good High/Low Low Easy
Biodata .35 Good High/Low Low Easy
Graphology .02 Low High/High Low Low
Weighted application .50 Good High/Low Low Easy
forms
Simulations .54 Good High/High Low Difficult
Work samples .54 Good High/High Low Difficult

Then, more specifically, how do we know which selection tool to choose? And how do we
evaluate the effectiveness of a Selection Method’? Research suggests that some selection
procedures are of higher quality, better received by job applicants, cheaper than than others.
In this slide, there are five factors determining the effectiveness of assessment tools. First,
validity refers to the extent to which the assessment method predicts relevant components
of job performance. For example, structured interviews have much higher validity than
unstructured interview. That’s why you have to use structure interviews Personality
assessments are among the least effective in predicting job performance. However, when
they are combined with others measures, such as cognitive ability or integrity tests,
personality tests have a higher validity. Applicant reactions: perceived job relatedness and
fairness of the assessment method used in the selection process. Work samples are highly
correlated with job performance and preferred by applicants. Due to high administrative
costs, work samples may be administered as a final hurdle to a smaller group of finalists
rather than to all applicants in the early stages of the hiring process.

Next, assessment centers are expensive and managers will need to be involved to evaluate
the candidates during the assessment. So the usability might be low. But they have high
validity and are viewed positively by candidates because assessment centers, like work
samples, mirror actual activities and knowledge that are required to perform the job.

Last, adverse impact refers to the extent to which an assessment discriminates against
members of a protected class. Research has shown mechanical ability tests have shown
adverse impact on women and general cognitive ability tests have shown disparate impact on

10
Black and Hispanic groups. So when you are choosing to use cognitive ability tests
based on its high validity, you will have to also examine if this cognitive ability tests
may have adverse impact. So the decision comes down to what levels of validity and
adverse impact are desirable in your organization. Job simulations can take many
different forms, such as in-person assignments, online exams, take-home
assignments, role-playing, presentations or even virtual simulations.
----------------
— Return on investment: whether the assessment method generates a financial return
that exceeds the cost associated with using it
— Applicant reactions: perceived job relatedness and fairness of the assessment method
influences perceptions of the fairness of the selection process
— Usability: people in the organization must be willing and able to use the method
consistently and correctly
— Adverse impact: an assessment method is more effective if it predicts job performance
and other important hiring outcomes without discriminating against members of a
protected class

10
 Choosing Assessment Methods for
Selection (1)

Then how do we choose which selection tool to use for a position? Consider what
behaviors are performed on the job and what Knowledge, Skills, Abilities and
Other Characteristics (KSAO) are needed to perform them. These behaviors and
skills can be derived from a job analysis. The assessment in the selection process
should relate directly to one or more important aspects of the job. For this
reason, job analysis is important for choosing selection procedures because
different selection procedures capture certain KSAOs more effectively than
others. If you determined that a product-design job required engineering skills
and teamwork, you would want to use assessment tools that reflect these
requirements. You might consider a personality test that reflects a person’s
abilities to work well as part of a team, and you might include relevant
engineering and teamwork questions in the job interview. So the bottom line is
some sort of job analysis is legally required for selection procedures and it is also
a good practice to have in place.

11
 Choosing Assessment Methods for
Selection (2)

Evaluative Methods
• Cognitive ability tests
Screening Methods
• Values assessments
• Resumes and cover • Personality assessments
letters • Integrity tests
• Job applications • Job knowledge tests
• Behavioral, situational, Contingent
• Biographical and case interviews Methods
information • Situational judgment tests • Medical test
• Telephone screens • Job simulations • Drug test
• Work samples • Background checks
• Reference checks

Assessment methods for selection can be grouped into three categories depending
on their primary purposes. The screening methods which include resumes, job
applications, biographical information and telephone screens, narrow a pool of job
applicants down to a smaller group of job candidates. The job candidates are then
assessed in more depth using evaluative methods that evaluate the pool of job
candidates to determine whom to hire. Examples include cognitive ability tests,
personality tests, interviews, job knowledge tests, simulations and work samples.
Finally, job offers may be made contingent on passing contingent methods such as
medical test, drug test, and background checks. If the finalist passes the contingent
assessment, their contingent job offer becomes a formal job offer. The use of criminal
background checks is beginning to decline for many jobs and in certain U.S. states due
to “ban the box” laws, companies such as American Airlines, Starbucks, and Xerox
were among many organizations pledging to voluntarily remove the box from their
applications.

12
 Combining Selection Tools
u Validity of the candidate’s assessment is more valid when using
multiple predictors

u Two ways of combining assessment scores

u Multiple-hurdle approach: candidates must receive a passing score
on an assessment to continue in the selection process
u Compensatory approach: high scores on some assessments can
compensate for low scores on other assessments
u Combination of the two above

So far, we have discussed that organizations often use assessment methods to collect
info about applicants to accurately predict the applicants’ future performance.
We also learned how to increase the validity of the assessment tools, for
example, using multiple selection tools is better than relying on one selection
tool. No assessment method is best used alone.
Then how do we go about combining the results from the assessments we conducted?
There are three ways to combine these pieces of info to make an effective
selection decision. Multiple-hurdle approach is a sequential approach where
candidates must pass each hurdle to move to the next one. For example, they
have to pass a cultural fit test to move to the next, job knowledge test. The
compensatory approach is when high scores on some assessments can
compensate for low scores on other assessments. So if you have certain types of
certificates in the field, even though your GPA is not great, you can still move to
the next hurdle in the hiring process. Finally these two approaches can be
combined in a selection process. So you could use one approach in one area and
then use the other approach in the other KSAOs or capacities. Depending on the
type of people and talent the company is looking for, and the labor market in the
given time companies make a choice of how to select the final job candidates
based on one or two of these combining selection tool methods. Alight, this is it
for this week. We covered selection.

13