Improving Selection Outcomes With the Use of Situational Interviews: Empirical Evidence from a Study of Correctional Officers for New Generation Jails
| Author | Nannette Kistler,Mary K. Stohr-Gillmore,Michael W. Stohr-Gillmore |
| Published date | 01 April 1990 |
| Date | 01 April 1990 |
| DOI | http://doi.org/10.1177/0734371X9001000201 |
| Subject Matter | Articles |
1
Improving
Selection
Outcomes
With
the
Use
of
Situational
Interviews:
Empirical
Evidence
from
a
Study
of
Correctional
Officers
for
New
Generation
Jails
Mary
K.
Stohr-Gillmore
New
Mexico
State
University
Michael
W.
Stohr-Gillmore
Nannette
Kistler
Washington
State
University
The
post-Watson
case
legal
environment
calls
into
question
the
too
frequent
reliance
upon
interviews
for
attaining
good
’bottom
line’
representativeness
in
the
work
force—both
in
the
initial
selection
and
in
promotional
selection.
The
’situational
interview’—entailing
the
use
of
questions
about
appropriate
behavior
in
actual
job
situations—has
been
touted
as
a
way
to
improve
the
typically
unreliable
interview
to
an
extent
that
validatable
scores
can
be
derived
from
group-rater
interview
sessions.
A
comparison
of
selection
outcomes
for
correctional
offi-
cers
in
New
Generation
jails
employing
first
no
situational
content
and
then
including
situa-
tional
content
reveals
that
the
situational
interview-based
scores
reflect
a
higher
level
of
inter-
rater
agreement
(reliability)
and
a
higher
correlation
with
performance
on
the
job
as
reflected
in
performance
appraisal
scores
(validity)
than
occurs
in
interview
sessions
lacking
situation-
al
content.
The
policy
implications
of
these
findings
are
discussed
briefly,
and
a
plea
is
made
for further
replication
of
such
research
in
additional
public
sector
settings.
Spring
1990
(Vol. JOINO
2). 1-18
he
Supreme
Court
ruled
in
Griggs
v.
Duke
Power
Company
( 1971 )
~
that &dquo;...an
employer
could
not
use
a
selection
technique
having
an
T
adverse
impact
on
minorities
unless
that
technique
has
been
shown
to
measure
job-related
skills&dquo;
(Thomas
and
Heisel,
1983).~
Because
of
the
Griggs
ruling,
public
managers
are
faced
with
a
selection
process
dilemma:
(1)
continue with
use
of
standardized
but
unvalidated
tests
and
risk
lawsuits
if
it
can
be
shown
that
those
tests
have
an
adverse
impact
on
minority
groups;
or,
(2)
resort
to
unstructured
interviews
that
have
no
adverse
impact,
but
which
are
also
likely
to
represent
poor
selection
practice
for the
identification
of
quality
personnel.2
Students
of
selection
processes
make
note
that
public
managers
more
often
than
not
have
chosen
the
latter
course
of
action
in
the
interest
of
meeting
affirmative
action
goals.
Instead
of
developing
selection
tests
that
can
be
validated,
many
employers
have
elected
to
take
the
&dquo;course
of
least
resistance&dquo;
and
have
returned
to
the
interview
format
for
employee
selection
to
the
detriment
of
employee
quality.
As
Daniel
( 1986:1 )
has
aptly
noted:
The
courts’
and
the
Equal
Employment
Opportunity
Commission’s
(EEOC)
stringent
requirements
have
reduced
employment
testing
efforts
in
the
United
States
without
neces-
sarily
improvmg
the
position
of
women
and
mmonties
in
the
labor
force.
The
argument
that
EEO
standards
have
upgraded
selection
by
stimulating
mcreased
validation
may
be
2
true
for
a
few
organizations,
but
for
others
the
strict
requirements
have
had
a
chilling
effect
leading
to
the
abandonment
of
old
tests
and
an
unwillingness
to
develop
new
ones.
As
a
consequence
of
the
widespread
return
to
the
old
unstructured
inter-
views,
some
authors
(e.g.,
Dyer,
1981)
argue
that
the
quality
of
the
work
force
will
deteriorate
primarily
because
interviews
are
not
very
good
in dis-
criminating
between
those
applicants
who
prove
to
be
good
and
those
who
prove
to
be
poor
employees.
Specifically,
scholars
are
skeptical
of
the
utility
of
interviews
for
predicting
applicant
success
on
most
kinds
of
jobs
(Wagner,
1949;
Mayfield,
1964;
Ghiselli,
1966;
Latham
and
Saari,
1980;
Silverman
and
Wexley,
1987).
They
question
the
utility
of
face-to-face
interviews
for
gathering
true
impressions
of
personality
characteristics
(Ulrich
and
Trumbo,
1965),
and
they
decry
the
generally
&dquo;nebulous
and
intangible
character
of
most
interview
processes
(Ghiselli,
1966:
389;
Dyer,
1981).
As
should
be
clear
from
the
foregoing,
the
Griggs
ruling
clearly
placed
public
personnel
managers
in
a
fix.
If
they
continue
to
use
unvalidated
tests
they
risk
discrimination
charges,
but
if
they
return
to
subjective
interviewing
techniques
they
risk
hiring
persons
who
would
be
incapable
of
performing
well
on
the
job.
It
is
in
this
context
of
an
unenviable
choice
between
a
rock
and
a
hard
place
that
the
proponents
of
the
&dquo;situational
interview&dquo;
process
laud
their
technique
as
the
solution
to
this
personnel
dilemma.
According
to
Latham
and
Wexley
(1982),
a
proper
selection
procedure
is
important
both
because
it
alleviates
concerns
about
adverse
impact
(it
can
be
validated)
and
because
it
assists
organizations
in
screening
for
quality
employees.
Before
a
proper
situational
interview
can
be
constructed,
however,
the
effective
behaviors
for
a
particular
job
must
first
be
identified.
A
valid
selection
test
cannot
be
developed
until
the
organization
agrees
upon
an
acceptable
definition
(i.e.,
measure)
of
employee
behavior.
This
is
because
the
validity
of
a
test
is
determined
by
measuring
the
performance
of
people
on
the
test
and
measuring
the
performance
of
the
same
people
on
important
aspects
of
the
job.
If
there
is
a
signifi-
cant
correlation
between
these
two
measures
the
selection
procedure
is
valid
(Latham
and
Wexley,
1982:3).
The
&dquo;critical
incident
technique&dquo;
for
job
analysis
was
developed
originally
by
Flanagan
(1954)
and
has
been
further
developed
by
numerous
scholars
in
a
number
of
settings.
This
technique
is
used
by
personnel
specialists
in
the
job
analysis
process
to
identify
the
effective
and
ineffective
BEHAVIORS
associated
with
a
given
job
(Flanagan,
1954;
Latham
and
Wexley,
1982).
In
accord
with
this
technique
those
persons
(job
incumbents
and
supervisors)
who
currently
perform
or
have
recently
observed
the
performance
of
the
job
in
question
are
asked
to
provide
several
examples
of
effective
and
ineffective
behaviors
related
to
that
job.
Upon
collection
of
these
critical
job
behaviors,
researchers
distribute
the
behaviors
under
local
dimensions-such
as
&dquo;Managing
the
living
unit
to
assure
a
safe
and
humane
environment&dquo;3
in
New
Generation
Jails
(Latham
and
Wexley,
1982).
To
minimize
sampling
bias,
a
To continue reading
Request your trial