Chapter 13: Scaling
Applied Social Research A Tool for Human Service
Terms in this set (11)
What is a Scale?
A scale is a number of items that combine to form a composite score on a variable.
May involve multiple dimensions of a variable
May involve degree to which a variable is present
Advantages of Scaling
Increased Level of Measurement
Increased Efficiency in Data Handling
Developing Scales I
First seek out existing scales
Scale development principles:
Obtain or develop a pool of possible items.
Eliminate redundant, ambiguous or inappropriate items
Pretest for validity and reliability
Screen out items based on pretest
Repeat process to select number of items desired.
Developing Scales II
Sources of Scale Items
Judges—individuals knowledgeable about the topic
Target population of the research
Characteristics of Scale Items
Range of variation
Unidimensionality—all items measure one variable
A Likert scale consists of a series of statements, each followed by a series of response alternatives.
The numbers associated with each response are totaled to provide the overall score .
The Likert is a summated rating scale.
*Note: mid and neutral point
Pros and Cons:
-Offers range of choices
-Provides at least ordinal level measurement
-Summary score loses information
Constructed with equal-appearing intervals—assumes that the distance between any two adjacent points on the scale is the same.
Items range from 1 (most unfavorable) to 11, (most favorable). Respondents select the item(s) with which they most agree.
Thurstone vs. Likert
-Interval level data
-Quick response by users
-Higher reliability with fewer items
-Ease of construction
Semantic Differential Scales
The semantic differential format presents the respondent with a stimulus, such as a person or event, that is to be rated on a scale between a series of polar opposite adjectives.
Pros and Cons:
-Requires only four to eight adjective pairs per dimension for reliability of .80 or better.
-Easy and less time consuming to construct.
-Adjective pairs can be adapted from prior studies
-Identifying the abstract dimensions is somewhat subjective and judgmental.
In a Guttman scale, items have an inherently progressive nature relating to the intensity of the variable being measured.
There is only one pattern of responses that will yield any given score on the scale.
Reproducibility: the ability of each individual's composite score to predict exactly the items agreed and disagreed with.
*Note: What is the breaking point of the idea/question/concern?
Response bias is the tendency for people's answers to questions to be influenced by things other than their true feelings, beliefs, and behaviors.
Sources of response bias:
-response pattern anxiety
-social desirability effect
Scaling of the Human Services
Example of linkage between research and practice.
Strengthen intake process
Aid to assessment for intervention planning
Document progress of intervention
Overall agency quality assurance monitoring system