Initial Evaluation 247 Evaluation of the Matrix 248 Evaluation of the Deterministic Test 250 for the Identity Matrix 251 .
PREFACE
Chapter three presents a discussion of the types of measurement error and strategies to reduce error. Chapter six begins with a discussion of scale development, which is the focus of the rest of the book.
Acknowledgments
Chapter 1 presents the definition of measurement along with a brief history of measurement in psychology and the health sciences and a discussion of the benefits of measurement. Many thanks also to those who read drafts of the book and provided feedback, especially Stanley A.
THE AUTHOR
Measurement in Health Behavior
Role of Measurement in Health Education and Health Behavior Research
CHAPTER ONE
INTRODUCTION TO MEASUREMENT
She helped with the development of the focus group interview guide and the writing of items for the questionnaire; both tasks required an understanding of measurement principles. The aim of the study is to evaluate an intervention designed to increase physical activity among older men and women.
Brief History of Psychosocial Measurement
In the early 20th century, Spearman (1904) introduced the concept of reliability to refer to the consistency of a measure. Advances in computer technology made possible significant advances in psychometrics in the second half of the twentieth century.
Conceptualization of Measurement
Therefore, alternative approaches to the conceptualization of measurement developed in the second half of the century, along with new evaluation techniques. Without the measurement process, we would be forced to use a conceptual system to discuss the children's characteristics, limiting our statements to broad conceptual comparisons.
Reasons for Measuring Concepts
The health educator might also suggest that seminar participants would be more likely to adopt strategies to control hypertension than control group participants. Using data collected from the pretests and posttests, the health educator can answer the research questions or test the hypotheses.
Scales of Measurement
The numbers assigned to the categories may only reflect the order of the categories. Variables measured on the interval scale have a meaningful order: higher levels have more of the attribute.
Major Concepts of Measurement: Reliability and Validity
The major area of controversy for behaviorists is the interpretation of the summed rating scales. Nunnally and Bernstein (1994), along with others (Pedhazur & Schmelkin, 1991), note that there is little harm in the use of higher-level statistics for the analysis of data using summed rating scales.
Summary
In this book, we will address reliability related to instruments designed to measure health behaviors and factors associated with health behaviors. The purpose of this chapter is to familiarize the student with the types of measures available for assessing health behaviors and factors associated with health behaviors.
CHAPTER TWO
At the beginning of the first chapter, we briefly described a number of roles filled by graduates of schools of public health. We can broadly classify instrument types as self-report, observations, bio-behavioral measures, and electronic monitors.
TYPES OF MEASURES
You will notice that the graduates were involved in several projects that focused on different health issues that required data collection. To make data collection decisions, graduates must first decide which variables to assess; Then they must decide how to measure it.
Self-Report
Each respondent's numerical score is calculated by measuring the distance from one point of the line to the respondent's point on the line. The values of the ranking given by the judges to the items are placed in order from high to low.
Observations
Biobehavioral Measures
Biobehavioral measures are generally considered more objective measures of health status than self-report measures. Because biobehavioral measures are seen as more objective measures in some cases, researchers are encouraged to include them in their studies whenever possible.
Electronic Measures
The enormity of measurement error is captured by the proverb "Measure twice, cut once." And the dilemma faced when errors do occur is reflected in Brennan's axiom (2001, p. 2): “A person with one watch knows the time of day. A person with two watches is never quite sure what time it is." Because measurement error is inherently ubiquitous,.
CHAPTER THREE
MEASUREMENT ERROR
Either this percentage of body weight change is due to gross measurement error or it reflects a serious medical condition. We begin this chapter by defining measurement error in the context of psychometric theory.
Definition
It is important for health professionals to understand the sources of measurement error and learn strategies to reduce or eliminate it when assessing health-related concepts. Random and systematic errors can be attributed to three broad factors: respondents, instruments, and situations.
Classification of Measurement Error
In any case, the instrument may be reliable; investigators may obtain the same results with repeated use of the instrument. Systematic error causes results to be inaccurate and, if extreme, can limit the usefulness of the measuring instrument.
Measurement Error Factors
At the extreme end of the continuum are people who use a negative attitude to respond to items (Silverman, 1977). In the tailored design method, researchers pay attention to the introduction of the survey (either in a written cover letter for paper and pencil and mail surveys or in the introductory pattern for in-person and telephone surveys).
CHAPTER FOUR
SURVEY DEVELOPMENT
In the course of their work, health educators and behavioral scientists often create surveys to collect information on a variety of health-related topics. Health behavior researchers use surveys to collect data on health behavior or the variety of factors associated with health behavior.
Basic Principles of Survey Construction
Daniel, as mentioned earlier, also wants to know how often unit staff wash their hands each day, whether or not they have developed contact dermatitis, and what they perceive as barriers to hand hygiene practices. What percentage of unit personnel developed irritant contact dermatitis due to common hygiene practices.
General Principles of Item Writing
Write items related to the purpose of the survey and addressing the objectives. That is, the items should reflect the purpose of the survey and be transparent to the respondent.
Survey Format
Second, people tend to interpret consecutive questions as related to the same topic unless they are told otherwise. They can fill in topics about which they are most knowledgeable or most comfortable and then move in a seemingly random fashion through the survey.
Survey Administration
In this chapter, we presented a brief description of the survey development and administration procedures. The objectives determine the content of the survey and direct the development of individual items.
CHAPTER FIVE
Likewise, sun worshipers are likely to understand the relationship between sun exposure and skin cancer risk. Despite the lack of a consistent link between knowledge and health behavior, behavioral scientists occasionally want to know what people know about health behavior.
KNOWLEDGE TESTS
The development and evaluation of these tests receive little attention in the health behavior literature. In this chapter, we present a brief introduction to some of the principles and practices that are useful to behavioral researchers and health educators in developing and evaluating knowledge tests.
Test Construction
Often little attention is paid to the primary use of the test and the quality of the items. Instead, the researcher must carefully consider the content of the test and the population for which it is being developed.
Types of Items
In the following example of bad answers, one of the confounding factors (wrong answers) is reading about HIV. Better Answer Choice: Which of the following lists the three main modes of transmission of HIV.
Item Analysis
Therefore, consider the full range of ratings below when evaluating each item-lens combination. The calculation of the item-objective congruence index is complex and will not be presented here.
CHAPTER SIX
Identify ways in which a selected concept is the same or different from another concept. Matt, whom we met in Chapter One, is the project director of a recently funded study to evaluate the results of a nutrition and exercise program for people over sixty-five.
THEORY AND MEASUREMENT
However, he has not been able to find a scale to measure OE for healthy eating. The principal investigator was pleased that Matt and the staff had taken the initiative to achieve the goal of obtaining scales to measure the study variables.
Linking Measurement to Theory-Based Health Practice and Research
Women who report higher levels of self-efficacy to perform breast self-examination will report more frequent breast self-examination. For example, we can say that the concept of self-efficacy is associated with the health behavior of breast self-examination.
Conceptualization Issues in Scale Development
Two researchers may use the same theoretical definition of a concept, but choose different empirical indicators of the concept. The concept analysis enables the researcher to develop a good understanding of the concept and the ways in which people use and interpret the concept.
Concept Analysis
If the concept is embedded in a theory, the theorist often identifies dimensions of the concept. Consequences are those events, actions, processes, cognitions or feelings that follow the expression of the concept.
CHAPTER SEVEN
This information is then used to guide the article writing process, which we discuss in this chapter. In this chapter, we present the steps (or plan) used to develop a set of items that together measure a variable as defined by the researcher.
ITEM WRITING AND SCALING
We have learned that a thorough concept analysis must conclude with an explanation of the theoretical definition and a description of the attributes associated with the concept. Many of the features of writing survey, multiple choice, and true/false items are also part of writing items for scales.
Item Development Process
To begin the item writing process, the researcher refers to the theoretical definition and attributes of the concept derived from the concept analysis. Thus, he and staff members relied on their own interpretations of the concept to create the items for the scale.
Summated Rating Scales
The researcher must select the type of scale (agreement, frequency, evaluative) that corresponds to the explanation of the variable. The novelty of the scale is likely to hinder accurate responding to the items.
CHAPTER EIGHT
Once a scale is developed, it must be evaluated for reliability and validity before widespread use. In order to evaluate instruments, researchers must have a good understanding of how and when to perform correlational analysis.
REVIEW OF STATISTICAL CONCEPTS
Before delving into specific psychometric tests, this chapter presents a brief overview of the principles of correlation and ANOVA. In presenting the material, we assume that most students have taken a statistics course and can use this chapter as an overview of the material.
Basic Statistical Concepts
When an item is negatively worded, agreement with the statement represents a lower level of the measured variable. The procedure creates a new variable (e.g., “appreciation”), which is the sum of participants' responses to each of the items on the scale.
Review of Correlation
The first number given in each cell of the matrix is the correlation coefficient (1). In the calculation below, we show the calculation of the percentage of shared variance for the number of hours of exercise and VO2.
Analysis of Variance
An F test that is statistically significant indicates that the difference between the mean scores of at least two groups is greater than chance. P <.05 means that there is a statistically significant difference between at least two groups.
CHAPTER NINE
Use the reliability coefficient to determine the percentage of the variance of the observed score due to the variance of the true score and the error score.
FUNDAMENTALS OF RELIABILITY
Suppose the first histogram is a plot of the scores that a student, Josh, gets on 100 attempts to take the same research methods test. Notice that the scores he gets from the attempts on the theory test are closer to the average than on the research methods test.
Classical Test Theory
Thus, the best estimate of the true score is the average of the observed scores. Error scores were obtained by subtracting the true score from each corresponding observed point.
Reliability Coefficient