390-Surveys

advertisement
BHV 390
Surveys
Operationalization of
Concepts
Operationalize means to make
measurable
You operationalize concepts by turning them into
measurable variables.
Variables are, by definition, measurable.
Surveys are made up of items collecting data on
concepts the researcher is studying.
Each item was chosen to operationalize a concept
into a particular type of variable.
Operationalization
in Surveys
1. Validity
2. Level of Data (Nominal, Ordinal, Ratio)
3. The amount of detail
(See the section on Operationalization for 1-3
1. Exhaustiveness of items
2. Clarity
3. Lack of bias in wording
4. No double barreled questions.
5. Question, item format
6. Participant competence
7. Order of items
8. Contingency options
Kinds of Variables
General
Type
Specific
Type
Definition
Examples
Discrete
Nominal
Numbered for labeling
in a category based on
equality to or difference
from other members
of the category
African American
•Asian American
•Latino
•Euro-American
Continuous
Ordinal
Items are ranked in
categories with no
standard interval
implied
by their ranking
1)First Place
2)Second Place
3)Third Place
(the distance between first and
second place, and second and
third place may be very
different from one another)
Continuous
Interval
Items are ranked with a
standard interval, but
without a standardized
zero point.
32 degrees Fahrenheit
64 degrees Fahrenheit
(32 does not imply twice as
cold as 64, and the Fahrenheit
and Centigrade have different,
non-standard zeros.
Continuous
Ratio
Items are ranked with a
standard interval based
on a standardized zero
point, and a ranking of
4implies exactly twice
that of 2.
Age in years
Exploratory, Descriptive or
Explanatory Goals
• Descriptive Goals
Demographic variables as descriptive
tools
Open ended questions, items to elicit the
point of view of participants
Explanatory Goals
Demographic variables as independent
variables
Closed- ended questions to collect
identical, comparable information from
all participants
General Kinds of
Measures
• Closed Ended
• Multiple Choice
• Matrix
• Open Ended
• Fill in the Blank
• Short Answer
• Contingency items
• Single Item
• Composite Measures
Exhaustiveness of Items
• WRONG
1. What is your
ethnicity?
_____ African
American
_____Asian American
_____EuroAmerican
• RIGHT
1. What is your
ethnicity?
_____ African American
_____Asian American
_____EuroAmerican
_____Native American
_____Latino
_____Other, please specify:
_______________________
Clarity of Questions = Validity
If you want to know whether anyone in a
participant’s family has ever been in prison:
• WRONG
1. Has anyone in
your family ever
gone to jail?
OR
1. Has anyone in
your family ever
been arrested?
• RIGHT
1. Has anyone in
your family ever
served a prison
sentence?
_____Yes
_____No
Bias in Questions
Example:
1. Don’t you think that women
should be stay-at-home moms?
___ Yes
___ No
Be careful that the questions
in your items do not contain
the answer you hope to get.
Double Barreled
Questions
Example:
2. Did you pay your state and
federal taxes last year?
___ Yes
___ No
There are two questions
contained in this item. It is
possible that a participant paid
federal but not state taxes or vice
versa. Be sure that you separate
out your questions so that there is
only one per item.
Contingency Questions
WRONG:
1. Are your parents currently married
to one another? _____ yes
_____ no
2. How many years have they been
married? _____
3. Do you think they have a happy
marriage?
_____ yes
_____ no
4. When did your parent’s marriage
end?
_____
Contingency Questions
WRONG :
1. Are your parents currently married
to one another? _____ yes
_____ no
IF NO, GO TO # 4
2. How many years have they been
married? _____
3. Do you think they have a happy
marriage?
_____ yes
_____ no
GO TO # 5
4. When did your parent’s marriage
end?
_____
Contingency Questions
RIGHT:
1.
Have you parents ever been married?
_____Yes
_____no
IF NO, GO TO # 6
2. Are your parents currently married to one
another?
_____ yes
_____ no
IF NO, GO
TO # 5
3. How many years have they been married?
_____
4. Do you think they have a happy marriage?
_____ yes
_____ no
IF YES, GO
TO # 7
5. How many years were your parents married before their
marriage ended?
_____
6. Have you parents ever lived together?
_____yes
_____no
7.
How old was your mother when you were born? _____
Matrix Items
A matrix format sometimes makes it
easier for participants to respond to a
set of items that all have the same
options. The section of the State-Trait
Anger inventory below is arranged in
a matrix format.
Not at all Somewhat Moderately so Very much so
5. I am furious.
0
1
2
3
6. I feel irritated.
0
1
2
3
7. I feel angry.
0
1
2
3
8. I feel like banging on the table. 0
1
2
3
9. I feel like hitting someone.
0
1
2
3
10. I feel like breaking things.
0
1
2
3
11. I feel like yelling at someone. 0
1
2
3
12. I am mad.
0
1
2
3
13. I am burned up.
0
1
2
3
14. I feel like swearing.
0
1
2
3
Steps in Constructing A Survey
• List all variables that you will measure
• Determine which of the formats above you will use for
each variable.
• Determine what level of data you will collect for each
variable.
• Organize the items for presentation on a survey
according to:
format
matrix or other clustering of items
contingency series
influence on each other
visual flow
• Pretest the survey on several individuals who will not be
participating in the survey for
comprehension
logical consistency
time
• Revise according to suggestion that enhance the
survey’s effectiveness
Study Guide
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
Survey
Operationalization
Variable
Validity
Level of Data
Research goal
Amount of detail
Exhaustiveness of items
Clarity
Question/item format
Participant competence
Nominal data
Ordinal data
Ratio data
Discrete data
Continuous data
Open-ended
Closed-ended
Fill in the blank
Multiple choice
Contingency items
Matrix format
Short answer
Composite/multiple item measures
Index
Scale
Biased questions/items
Double-barreled questions/items
Pretesting the survey
Download