Académique Documents
Professionnel Documents
Culture Documents
INSTRUMENTATION
Development
and
valida8on
of
data-‐
collec8on
tools
• Ques8onnaires
• Interview
schedules
• FGD
guides
• Lab
form
sheets
• Observa8on
checklists
QUESTIONNAIRE
DESIGNS!
DEFINITION
• A
ques%onnaire
is
a
data-‐collec6on
tool
used
to
gather
informa8on
about
an
individual,
group,
things,
phenomena
or
events.
• It
is
a
wri;en
document
which…
1) respondent
fills
in
by
himself
(self-‐
administered)
or
with
assistance
from
the
researcher
(interviewer-‐assisted)
2) collected
by
field
interviewers
(interviewer-‐
administered)
3)
may
be
sent
through
standard
mail
or
email.
Developing
a
ques6onnaire
is
not
easy.
It
is
a
tedious
and
6me-‐
consuming
process.
• List
-‐
The
respondent
is
given
a
list
of
choices,
any
of
which
may
be
selected
Quan6ty
ques6on
• When
did
you
start
seeing
a
dermatologist?
________
To
what
extent
does
your
skin
disease
affect
your
level
of
self-‐esteem?
/___________x_________________________/
Not
affected
extremely
at
all
affected
QUANTIFYING
A
QUALITATIVE
VARIABLE
(Steps
in
Scale
Construc8on)
1. Determine
the
nature
of
the
scale
to
be
constructed.
2. Opera8onally
define
the
construct.
Determine
domains
to
be
included
in
the
scale.
3. Devise
items
for
each
domain.
4. Establish
content
validity
of
the
scale.
5. Develop
response
categories.
6. Translate
the
items,
if
necessary
7. Pretest
the
scale.
8. Do
item
analysis.
9. Pretest
scale
again
if
necessary.
Re-‐do
item
analysis.
10. Establish
validity
and
reliability
of
the
scale.
11. Finalize
the
scale.
1.
Determine
the
nature
of
the
scale
to
be
developed.
“Knowledge”
IQ
test
“Prac8ce”
Achievement
test
Mastery
Test
Performance
appraisal
Screening
test
Behavioral
checklist
Compliance
scale
“A_tude”
A_tude
scales
Monitoring
checklist
Ap8tude
test
Observa8on
guides
Personality
test
Quality
of
Life
Scale
Interest
scale
II.
Opera8onally
define
your
construct.
CONSTRUCT-‐
an
a]ribute
of
a
person,
thing
or
event
which
is
not
readily
observable
Examples:
a_tude,
intelligence,
quality
of
life,
compliance,
s8gma
DOMAINS-‐
dimensions
or
aspects
of
the
construct
Example:
Construct-‐A_tude
Domains:
Beliefs
Affect
Behavioral
inten8ons
VARIABLES-‐
concrete
measures
or
indicators
of
the
domains
Example
Construct:
Quality
of
life
–
(subjec8ve
well-‐being)
Domains:
Physical
well-‐being
Emo8onal
well-‐being
Social
well-‐being
Func8onality
Cogni8ve
well-‐being
Variables:
I.
For
emo8onal
well-‐being
In
the
past
2
weeks,
1.
did
you
feel
sad
because
of
your
disease?
2.
did
you
feel
anxious
because
of
your
disease?
3.
did
you
feel
irritable
because
of
your
disease?
II.
For
physical
well-‐being
In
the
past
2
weeks,
1.
how
much
pain
did
you
experience?
2.
how
easy
were
you
to
get
8red?
3.
how
oken
did
you
vomit?
III.
Devise
items
for
each
domain
1. Construct
an
item
pool
of
about
7-‐10
items
per
domain.
In
the
final
scale,
there
should
be
at
least
5
items
lek
(aker
item
analysis)
2. Items
can
be
in
the
form
of
statement
or
ques8on.
It
can
be
answerable
by
dichotomous
responses
or
mul8ple
responses.
• Make
sure
that
the
scale
has
covered
adequately
the
domains
of
the
construct
• Make
sure
that
there
are
enough
items
for
each
domain
(about
7-‐10
items
per
domain)
• Make
sure
that
the
items
are
relevant
to
the
culture
and
condi8on
of
the
target
respondent
(emic
content)
• Seek
expert
opinion
on
the
content
validity
of
the
scale
V.
Develop
response
categories
• The
best
response
scale
suits
the
level
of
literacy
of
the
target
respondents.
Rule
of
thumb
–
level
of
language
should
be
at
grade
6
level
even
if
target
respondents
are
professionals
• Psychometrically,
con8nuous
scales
are
more
preferred
than
discrete
/dichotomous
scales
TYPE
OF
RESPONSE
SCALES
How
much
pain
have
you
felt
in
the
past
2
weeks?
__________x_______________________________
No
pain
at
all
extreme
pain
B.
Adjec8val
scales
(Likert)
My
social
ac8vi8es
are
limited
by
my
disease.
____/______________/__________/_____________/_____________/___
Strongly
agree
agree
undecided
disagree
strongly
disagree
C. Faces scale
Do
you
feel
anxious
about
your
disease?
10
9
8
7
6
5
4
3
2
1
0
D. Ladder scale
Hedonic
Scale
VII
Pretest
the
scale
• 1.
to
self
• 2.
to
colleagues
• 3.
to
sample
of
target
respondents
Sample
size:
number
of
items
x
3
Take
note
of
comments
about
item
wording,
transla8ons,
format,
styles,
size
of
font,
color
of
paper
The
following
ques6ons
can
be
addressed
through
pilo6ng:
• Do
the
respondents
understand
what
the
ques8onnaire
is
about?
• Test-‐retest
method
No
significant
difference
in
scores
when
same
test
is
administered
in
2
separate
occasions
to
same
subjects
using
the
same
measurement
scale
Validity
Content
validity
1.
Does
each
item
pertain
to
one
aspect
of
the
construct
(content
relevance)?
3. Does
the
scale
not
correlate
with
other
indices
that
measure
different
a]ributes
(divergent
validity)?
4. If
item
is
“bad”
in
several
or
all
methods
but
is
considered
cri8cal
in
terms
of
clinical
or
social
significance,
then
consider
the
item
but
analysis
will
be
descrip8ve
for
that
par8cular
item
(i.e.,
do
not
include
the
item
in
the
computa8on
of
the
overall
numerical
scale
score).
Issues
related
to
scoring
Use
a
transi.on
statement,
i.e.,
an
introductory
phrase
to
explain
why
the
next
ques8ons
are
asked
so
that
respondents
retain
a
sense
of
the
logical
flow
of
the
ques8onnaire.
“The
past
ques8ons
asked
about
some
of
your
socio-‐demographic
characteris8cs,
now,
the
next
ques8ons
will
inquire
about
your
opinions
about
the
skin
care
system
in
the
country…”
Use
filter
ques.ons
if
some
ques8ons
are
not
applicable
to
all
respondents
1.
Are
you
currently
prac8cing
as
a
skin
specialist?
Yes/No
If
your
answer
is
“No”,
proceed
to
ques%on
3.
2.
How
long
have
you
been
prac8cing
as
a
skin
specialist?____
3.
What
factors
have
hindered
your
prac8ce?
Filter
ques8ons
should
not
require
extensive
page
flipping
by
the
respondents
or
memory
of
answers
to
earlier
ques8ons.
9.
Sequencing
of
ques6ons