Stress Profiler Validity Information

Below is a letter from Dr. Kenneth Linfield from Spalding University which states his research for the validation of The Stress Profiler...

Spalding University Logo


Mr. James Porter
Audio Vision
3 Morningside Place
Norwalk, CT 06854

Dear Mr. Porter,

I am pleased to report the following preliminary psychometric information on the Stress Profiler. The temporal stability (test-retest reliability) and the internal consistency (Cronbach's alpha) of the overall score on stress are excellent (r12=0.92 and a=0.95).

Further, the temporal stability of the ten subscales ranges from very good to excellent (0.78 to 0.93), and the internal consistency of those subscales ranges from good to very good (0.69 to 0.84).

Please find enclosed a more detailed description of these psychometric properties, including a complete table of the estimates for each subscale and the total score, as well as my abbreviated curriculum vitae.

As we have discussed, it is clear that the Stress Profiler is a very useful measure of stress because it is so user-friendly. In my use of it with students and in occupational consulting projects, those who have taken it have consistently reported finding it to be a very helpful measure partly because of the clear descriptions of stress levels and suggested coping strategies. Further, the data I have summarized suggests that it is a psychometrically sound measure of stress as well. As such, I plan to continue to collect data on its reliability and validity, and to report that information to you when it is available.

Best wishes for continued success with your fine measure.

Sincerely,

Kenneth J. Linfield, Ph.D.
Assistant Professor
School of Professional Psychology

Description of Psychometric Properties of the Stress Profiler

Kenneth J. Linfield, Ph.D.

The reliability estimates for the Stress Profiler listed in Table 1 are based on the responses by students enrolled at Spalding University during the 2003 calendar year. The temporal stability estimates (test-retest reliability correlations) are based on the responses by 20 graduate students enrolled in Fall 2003 and 20 undergraduate students enrolled in Spring 2003 who completed two administrations of the Stress Profiler two weeks apart.

The internal consistency estimates (Cronbach alpha) are based on the responses from the first administration of the above groups as well as a single administration to 28 undergraduate students enrolled in a class during Fall 2003. The test-retest group was predominantly female (n=29, 72.5%), and Caucasian (n=30, 75%) with a range of ages from 17 to 38. >The internal consistency group was also predominantly female (n=47, 69.1%) and Caucasian (n=53, 77.9%), with an age range also from 17 to 38.

Table 1

Scale

Test-Retest Reliability
Pearson Correlation Coefficients, n=40

Internal Consistency
Cronbach alpha, n=68

1

0.84

0.73

2

0.88

0.81

3

0.82

0.69

4

0.81

0.81

5

0.88

0.83

6

0.78

0.84

7

0.78

0.79

8

0.85

0.74

9

0.87

0.69

10

0.93

0.76

Total score

0.92

0.95

Temporal stability is generally seen as the primary indication of the reliability or consistency of any measure.It is based on the assumption that measures of important constructs (like stress or intelligence or personality) should yield similar results when the construct is assessed at two times, as long as the times are close enough that the underlying construct would not be expected to change substantially.

Although there is no single length of time that is ideal for test-retest reliability for the construct of stress measured by the Stress Profiler, the span of two weeks was chosen both to be short enough that stress level should not have changed substantially and long enough to reduce the degree to which respondents relied on their memory of their previous response.

The estimate of 0.92 for the total Stress Profiler score is an indication of excellent temporal stability, suggesting that the test is a highly consistent measure.The generally slightly lower estimates for the subscales, ranging from 0.78 to 0.93, show very good to excellent temporal stability, suggesting that even breaking the general construct of stress into the ten components of the Stress Profiler, something that normally reduces reliability, still results in consistent measures.

Description of Psychometric Properties of the Stress Profiler

Internal consistency is generally seen as another important indication of the consistency of measures.It is based on the assumption that the various items used to measure particular aspects of any construct should show similar patterns of results.There are a variety of specific ways to calculate estimates of internal consistency, although they all indicate the degree to which respondents show similar response to the various items.

That is: the degree to which people who score high on any given item also score relatively high on other items in the same scale, while those who score low on the first item also score relatively low on the other items.

One simple way to calculate an estimate of internal consistency is to split a scale into half, calculate individuals' scores on each half, and then calculate the correlation between the scores on each half.This is called "split-half reliability".

A more common measure that follows this principle recognizes that any given method of dividing the items into two halves of the scale may result in a stronger or weaker correlation than an alternate method with the same items.For example, in a scale of ten items, one could use the first five and the second five, or the five odd and the five even items.

One of those methods would likely give a slightly higher estimate with the other one giving a slightly lower one simply because of the particular responses to the various items.To avoid the error of a given method, all possible methods of splitting the scale into two halves are used, and the average of all correlations is calculated.This is called the Cronbach alpha.

The Cronbach alpha for the entire scale of the Stress Profiler is 0.95, indicating that the entire scale is highly internally consistent.People who score high on half of the 100 items tendvery strongly to score high on the other half, and those who score low on one half likewise tendvery strongly to score low on the other.The alpha estimates for the ten subscales would be expected to be lower because a greater number of items in a scale leads to higher Cronbach alphas even when the tendency of people to respond in similar ways across items remains the same.

In general, however, the internal consistency estimates of the subscales from 0.69 to 0.84 show good to very good internal consistency.These estimates, especially in combination with the very good estimates of test-retest reliability, show that the Stress Profiler is a very consistent measure.

Biography

Kenneth J. Linfield earned his Ph.D. in clinical psychology at the University of Illinois at Urbana-Champaign in 1999, and now serves as an assistant professor in the School of Professional Psychology at Spalding University in Louisville, Kentucky. With his minor in quantitative psychology, he teaches graduate courses in psychometrics of psychological assessment and one of the sections of the research sequence in addition to a range of undergraduate courses.He is the principle investigator on multiple projects (Assessment of Spiritual Well Being in a Community Mental Health Center population, Temple Care II an action research project on health psychology interventions, and Development of Screening and Career Assessment Measures for Clergy and Missionaries).In addition, he has served as the statistical consultant on several others (The Early Intervention Project, CARE I and CARE II, Development of the Operating Style Assessment) as well as assessing the psychometric properties of the Stress Profiler.

Related Items

The Stress Profiler Student Version
A 100 question stress test designed for students.
The Stress Profiler
A 100 question stress test.

Popular Products


Biodots

Stress Testing Cards

The Stress Profiler

A Day Away From Stress CD

The Perfect Day DVD