Developing and Testing Self-Report Scales & Descriptive Statistics in nursing

Polit10e_PPT_Ch15.ppt

Home >Nursing homework help >Developing and Testing Self-Report Scales & Descriptive Statistics in nursing

Chapter 15

Developing and Testing Self-Report Scales

Question

Tell whether the following statement is true or false:

Scale development begins with a sound conceptualization of the construct (the latent variable) to be measured, excluding its dimensionality.

Answer

False

Scale development begins with a sound conceptualization of the construct (the latent variable) to be measured, including its dimensionality.

Conceptualizing the Construct

Scale development begins with a sound conceptualization of the construct (the latent trait) to be measured, including its dimensionality.
Unobservable latent trait
Unidimensional

Type of Scale

Scale selection is done before items can be generated.

Two broad categories of multi-item reflective scales

Traditional summated rating (Likert)
Latent trait

Generating Item Pool

Sources
Existing instruments
Research literature
Concept analyses
In-depth studies
Clinical observations

Decisions about Item Features

Number of items
Response items
Positive and negative stems
Item intensity
Item time frames

Number of Items

Domain sampling model: homogeneous set of items from a hypothetical universe of items
Goal is to provide appropriate redundancy so idiosyncrasies can be cancelled out.
Recommends starting with 3 to 4 times as many items as you will have in your final scale

Response Options

Scale items contain:
A stem
Response options
For example, strongly disagree, disagree, agree, strongly agree

Positive and Negative Stems

The objective was to include items that would minimize the possibility of an acquiescence response set—the tendency to agree with statements regardless of their content.
Many experts currently advise again including negative and positive items on a scale.

Item Intensity

The intensity of the statements (stems) should be similar and fairly strongly worded.
Statements should not be so extremely worded as to result in universal rejection.

Item Time Frames

A time frame should not emerge as a consequence of item development.
It should be decided upon in advance, based on the conceptual understanding of the construct and the needs for which the scale is being constructed.

Wording the Items—Readability

Clarity
Jargon
Length
Double negatives
Double-barreled items

Question

Once content validity has been established at a satisfactory level, the scale must be administered to a development sample. What is the typical number of respondents who are representative of the target population?

100

150

250

300

Answer

Once content validity has been established at a satisfactory level, the scale must be administered to a development sample—typically 300 or more respondents who are representative of the target population.

Content Validity

Satisfactory level
Development sample
300 or more respondents representative of target population

Field Testing the Instrument

Developing the sampling plan
Developing a data collection plan
Preparing for data collection

Question

Tell whether the following statement is true or false:

Exploratory factor analysis is used to reduce a small set of variables into a larger set of underlying dimensions, called factors.

Answer

False

Exploratory factor analysis is used to reduce a large set of variables into a smaller set of underlying dimensions, called factors.

Analysis of Scale Development Data

Basic item analysis

Inter-item correlation
Item–scale correlations

Exploratory factor analysis (EFA)

Factor extraction
Principal components analysis (PCA)
Factor rotation—orthogonal or oblique

Illustration of Factor Rotation

Scale Refinement and Validation

Revising the scale
Scoring the scale
Conducting a validation study

Confirmatory factor analysis

Confirmatory Factor Analysis

CFA is a second study often undertaken to validate the scale.
A subset of sophisticated statistical techniques (structural equation modeling)
Involves tests of a measurement model
Stipulates the hypothesized relationship between latent traits and manifest variables (items)

Interpretability of Scale Scores

Interpretability: the degree to which one can assign qualitative meaning
Percentile
Standard scores
Norms
Cutoff points

Critiquing Scale Development Studies

Methods used
Whether the evidence regarding the scale’s psychometric adequacy is sufficiently sound
See Box 15.1