Title: Mgmt 441 Staffing
1Welcome!
- Mgmt 441 Staffing
- Prof. Howard Miller
2Staffing Function
- Among several human resource functions
- Benefits
- Compensation
- Safety
- Labor Relations
- Training
3Staffing . . .
- Involves getting people in, or out, of a company
- Well concentrate on the getting people in part
4Staffing has 2 main purposes
- Get the best talent for the firm at the price
we can offer - Conform to national, state and local regulations
concerning staffing practice, or in short keep
it legal -
5How to Staff?
- Derive demand for labor
- Update job requirement information
- Identify knowledge, skills and abilities required
for success on job - Develop measures of job-related KSAs
- Recruit from relevant labor markets
- Screen using valid tests
- Make offer, provide orientation to accepts
6Staffing Project Steps
- Choose job to study
- Find real setting to perform job analysis
- Perform job analysis
- Develop job description, job specification
- Find/develop tests to measure job specifications
- Identify Relevant labor markets
- Define recruiting methods
- Spell out hiring process to client in full detail
7What do you mean valid test?
- Note importance of Supreme Court case
- Watson v. Ft Worth Bank and Trust (1988)
- Supreme court states (paraphrasing)
- A test is any hurdle you have to clear to get
a job - especially important for interviewing,
which was at core of case - A valid test is a hurdle that allows better
talent to get over, while lesser talent is
screened out
8How do we establish if a test is valid
- Note importance of The Uniform Guidelines for
Employee Selection Procedures (1978) Uniform
Guidelines link - Three methods recognized by courts
- Content validation
- Criterion-related validation
- Construct validation
9The logic of hiring validity illustrated
- Consider classic payoff matrix, which well label
selection decision matrix in our setting. - Good decisions result when applicants who will
succeed are hired (true positive), and
applicants who will not succeed are rejected
(true negative) - Bad decisions involve rejecting people who will
work out (false negative), and accepting those
who dont work out (false positive)
10Selection Decision Matrix
11Content validation
- A logical analysis by subject matter experts
(SMEs) of the overlap between the content of
screening tests and the content of job
requirements - Note the deceptive simplicity of the requirements
for a typist at the university
12Criterion-related validation
- The main way its done
- 2 types Predictive Validation and Concurrent
Validation - Predictive uses test data from applicants, and
job performance data from those hired - Concurrent uses both test and performance data
from current employees
13Construct Validation . . .
- More complex than the other two
- One has to show that measures of applicant traits
and job performance really measure those things
(Classic construct validation) - AND then show trait measures correlate with job
performance measures
14the vast majority of the time
- well be referring to concurrent,
criterion-related validity evidence because of
its overwhelming use
15Valid tests mean lower error rates in hiring
decisions
- More true positives and negatives
- Fewer false positives and negatives
- Lower exposure to successful litigation
- Higher utility of the staffing function
16Research on validity tells us what works!
- Find tests that are shown to be valid
- How is this done?
- Within the concurrent, criterion-related
validation approach, it means showing a
significant correlation between test scores and
job performance scores the essence of a
concurrent validity study.
17What would such validation data look like?
- See Ma and Pa Consumer Electronics Store data
set illustration - Ma and Pa Consumer Electronics Data Set 1.xls
18Valid Hiring Tests
- A hiring test anything you must get through to
get the job - A Valid hiring test one where applicants who
score better on the test do a better job if hired - In criterion-related empirical test validation,
it is one where there is a significant
correlation between hiring test scores and job
performance scores
19Empirical validation
- Obtain a representative sample of people
- Have them take the hiring test(s), and measure
their job performance - Compute the statistical correlation of hiring
test scores and job performance scores - Compute the statistical significance of the
sample correlation - If significant, cross-validate in new sample
- If correlation remains significant, put test(s)
into use
20Some key statistical concepts
- Mean the average score for a group of people
- Standard Deviation the average variability
around the average score for a group of people - Correlation a number that reveals the degree
of linear association between hiring test scores
and job performance scores
21Correlation properties
- Correlation is notated with lower case r
- It can range in value from -1.00 to 1.00
- r0 means zero correlation, no linear
association between the test (x) and job
performance (y) thats not happy - r1.00 (or r -1.00) means there is a perfect
association of hiring test and job performance -
doesnt happen in reality, tho wed love it if
it did! - Correlations from real samples RARELY exceed
values of r.50
22More on correlation
- Correlation can be used to summarize the
pattern in a 2-variable scatterplot, like the
hiring test (x axis) versus job performance (y
axis) scatterplot - In this application, correlation is a special
case of linear regression using a straight line
to summarize whats happening in a data set - Plot interview score against job performance in
the Ma and Pa data set
23Statistical analysis of interviews in relation to
monthly sales
- Compute the correlation of hiring test scores
with job performance scores - Compute the regression of job performance scores
(Y) on hiring test scores (X)
24Things to do
- Review material covered, so you ace the quiz next
week - Read the legal chapter in GFB
- Read the Uniform Guidelines
- Find a company/organization where you can perform
job analysis on real job