Approaches to Scale Construction

Question 1

Reflective measure:

Answer 1

when the indicators of the construct are considered to be caused by the construct

Question 2

Give an example of a reflective measure:

Answer 2

pretty much all psychological constructs – extraversion, organizational commitment, work-family conflict

Question 3

Formative measure:

Answer 3

when the measured variables are considered to be the cause of the construct

Question 4

What are [pretty much] all psychological measures considered?

Answer 4

reflective

Question 5

Explain the deductive approach to scale construction:

Answer 5

this approach is all conceptual and does not involve data collection

Question 6

What are the 3 steps to the deductive scale construction approach?

Answer 6

1) define the construct based on theory
- break it down into dimensions (components)
- also defining what it is NOT
2) write items for each dimension
3) have experts “ensure” items are related to the dimensions

Question 7

What are the [two] advantages to using the deductive scale construction approach?

Answer 7

1) strong tie to theory
2) covers the full range of the construct

Question 8

What are the [two] disadvantages to using the deductive scale construction approach?

Answer 8

1) the quality of your measure is based off the strength of the theory you used
2) the items you create may not relate to one another (empirically) in the way you expected

Question 9

(T/F) The deductive approach is commonly used in it’s purest form.

Answer 9

False

Question 10

Explain the inductive approach to scale construction:

Answer 10

This approach does not require theory, but instead uses a sample of people to describe the construct.

Question 11

What are the 5 steps to the inductive scale construction approach?

Answer 11

1) define the construct space (generally, and not based on theory)
2) write a set of items that seem to fall within that space
3) administer the items to a lot of people
4) evaluate the results
- specifically, the quality of each specific item and uncover
the basic dimensions of the construct with a exploratory
factor analysis
5) interpret and name the dimensions of the construct

Question 12

What are the [three] advantages to using the inductive scale construction approach?

Answer 12

1) you can develop an understanding of the construct as you go
2) we can refine the understanding of a construct
3) you can get rid of bad items (data analysis)

Question 13

What are the [two] advantages to using the inductive scale construction approach?

Answer 13

1) subjective interpretation of the dimensions
2) understanding of the construct influenced by chance
- item correlations can change with as you test it on more samples

Question 14

(T/F) The inductive approach is commonly used in it’s purest form.

Answer 14

False

Question 15

When is it particularly helpful to use the inductive approach?

Answer 15

When developing new constructs or ideas

Question 16

Explain the deductive-inductive approach to scale construction:

Answer 16

Using the advantages of both the inductive and deductive approaches.

Question 17

What are the 5 steps to the deductive-inductive scale construction approach?

Answer 17

1) Specify the construct we well as you can
2) Write items to represent the construct
3) Collect data
4) Evaluate the items with respect to data and theory
5) Repeat as necessary

Question 18

Further explain step 4 of the deductive-inductive scale construction approach [evaluate the items with respect to data and theory]:

Answer 18

Just because there is a data-based problem, doesn’t mean you should get rid of the item automatically. You should evaluate the data in light of your own theory. When an item is not functioning well, but you think the item really should work, let the theory win.

Question 19

Discriminant validity:

Answer 19

the extent to which a test is not related to another test that measures a different construct

Question 20

Convergent validity:

Answer 20

how closely a test is related with other tests that measure the same or similar constructs

Question 21

What are the three assumptions of the inductive and deductive approaches?

Answer 21

1) people interpret the items the same way
2) people are aware of, and can report their own behavior
3) people are willing to tell you about their behavior

Question 22

Explain the Empirical Keying approach to scale construction:

Answer 22

Approach scale construction with the theory that item responses alone are interesting behavior. Therefore, an item is of value if it reliably differentiates a criterion group from a comparison group. Based on responses, you can make a prediction about your group membership.

Question 23

What is the empirical approach based on?

Answer 23

that the assumptions of the inductive and deductive approaches are untenable (not able to be supported)

Question 24

What are the 4 steps to the empirical keying scale construction approach?

Answer 24

1) Generate items with response format of true/false
2) Have two different samples respond to the items
3) Compare each group’s responses to each item
4) Cross-validate

Question 25

What statistical procedure do you use to compare group responses with the empirical keying method?

Answer 25

A t-test or compare effect sizes.

Question 26

Do you have to explain why the groups are different with the empirical keying approach?

Answer 26

No, it doesn’t require any theory. You just have to show that the groups are different – this means you could have really weird items. For ex: depressed people typically report not liking the color green compared to those who are not clinically depressed.

Question 27

What is cross-validation?

Answer 27

A procedure used to assess the stability of a statistical model

Question 28

How is cross-validation done?

Answer 28

A data set is randomly divided into two subsets, the first is used to develop the model and the second is used to test it.

In regression analysis, for example, the first subset would be analyzed in order to develop a regression equation, which would then be applied to the remaining subset to see how well it predicts the scores that were actually observed.

Question 29

What is the most well-known empirically keyed test?

Answer 29

MMPI (Minnesota Multi-phasic Personality Inventory)

Psychometric test used to assess personality traits and psychopathology.

Question 30

What are the [two] advantages to the empirical keying approach?

Answer 30

1) no assumptions about capability to self-report behavior
2) researchers can assess poorly defined concepts (help identify groups)

Question 31

What are the [three] disadvantages to the empirical keying

Answer 31

1) test taker don’t like it
2) tends to be long
3) there is little influence of psychological theory (actually better today)