In: Statistics and Probability
Source: Dingley, C., & Roux, G. (2014). The role of inner strength in quality of life and self-management in women survivors of cancer. Research in Nursing & Health, 37(1), 32-41.
Introduction
Dingley and Roux (2014) studied inner strength, depression symptoms, and selected demographic variables to predict quality of life (QOL) and self-management in women survivors of cancer. This predictive correlational study was conducted with a convenience sample of 107 women with cancer. “The strongest predictors of QOL were depressive symptoms, inner strength, and time since diagnosis. The strongest predictors of self-management were depressive symptoms and inner strength” (Dingley & Roux, 2014, p. 32).
Relevant Study Results
Selected instruments used to measure the study variables in the Dingley and Roux (2014) study are presented in this section with a focus on reliability testing.
“Depression. The Center for Epidemiological Studies Depression Scale (CES-D; Radloff, 1977) was used to assess depressive symptomatology. The CES-D is one of the most widely used self-report instruments for epidemiologic studies of depression and has been used in primary care, psychiatric, and related clinical and forensic settings. The 20-item instrument measures depressive affect, somatic symptoms, positive affect, and interpersonal relations. For each experience related to depression, the respondent selects the value (0, 1, 2, or 3) that best describes how frequently the experience occurred during the previous week. Total scores of 15 to 21 indicate mild to moderate depressive symptoms; scores over 21 indicate the respondent has experienced major depressive symptoms. Cronbach’s alpha in this study was 0.90.
Inner strength. The Inner Strength Questionnaire (ISQ) is a 27-item self-report instrument written at a fourth grade level (by Flesch Kincaid Grade Index). Respondents are asked to indicate their level of agreement with each item statement using a 5-point Likert-type scale (strongly agree, agree, slightly agree, disagree, strongly disagree). The ISQ assesses four factors representing dimensions of the theory (i.e., Anguish and Searching, Connectedness, Engagement, and Movement). Total scores can be calculated, as well as scores for each subscale. The maximum possible total score is 135, with higher scores indicating a higher presence of inner strength. Each sub-scale of the ISQ had a Cronbach’s alpha > 0.80 (Anguish and Searching 0.85, Connectedness 0.95, Engagement 0.85, and Movement 0.83). Internal consistency reliability of the total ISQ was α = 0.91. Cronbach’s alpha for the present study was 0.89.
Quality of life and spiritual well-being. The tool selected to measure QOL was the Functional Assessment of Cancer Therapy—Spiritual WELL-Being (FACTSp), one instrument from the FACIT Measurement System, a collection of QOL questionnaires targeted at the management of chronic illness. The FACIT measurement system is considered appropriate for use with patients with any form of cancer as well as other chronic illness conditions (e.g., HIV/AIDS, multiple sclerosis) and in the general population using a slightly modified version. The FACT-Sp incorporates the domain of spiritual well-being (SpWB) in addition to the four primary domains of physical (PWB), social/family (SWB), emotional (EWB), and functional well-being (FWB). The SpWB scale is 12 questions that measure a sense of meaning and peace and the role of faith in illness… The SpWB had a Cronbach’s alpha of 0.93 in the study sample. Cronbach’s alphas for the subscales were PWB = 0.85, SWB = 0.80, EWB = 0.79, FWB = 0.88, and SpWB = 0.83” (Dingley & Roux, 2014, pp. 35-36).
Based on the information provided from the Dingley and Roux (2014) study, which scale has the lowest reliability or Cronbach’s alpha coefficient? What random error did this scale have for this study? Was this a study strength or weakness?
This is a simple problem related to analysis of the scale validity /acceptability with relation to the Cronbach’s coefficient.
But before getting into the details of the validity of good of fit of a scale ,we need to understand as to what exactly is a Cronbach’s coefficient and how is it related to scaling ?
Cronbach’s coefficient also known as Cronbach’s alpha is a measure of internal consistency of a measurement, that is, how closely related a set of items are as a group.
It is considered to be a measure of scale reliability.
Technically Cronbach’s alpha is not a statistical test – it is a coefficient of reliability (or consistency).
Cronbach’s alpha can be written as a function of the number of test items and the average inter-correlation among the items.
the formula for the Cronbach’s alpha is given as:
Here N is equal to the number of items, c-bar is the average inter-item covariance among the items and v-bar equals the average variance.
Now what can we interpret from its value ?
Clearly we can see from its formula that if we increase the number of items, it will directly increase Cronbach’s alpha.
Also, if the average inter-item correlation is low, alpha will be low.
As the average inter-item correlation increases, Cronbach’s alpha increases as well (holding the number of items constant).
Ideally The range of Cronbach’s alpha normally is between 0 and 1.11
However ,the value of Cronbach’s alpha could take a negative values well if there will be negative inter-item covariance with a large absolute value.
Thus we can interpret that
The closer Cronbach’s alpha coefficient is to 1.0, the greater the internal consistency of the items in the scale.
Usually we follow the following rule of thumb for it
Now that we have got a firm background of the Cronbach’s alpha /coefficient ,let us try to understand the question given and answer it.
Here we need to find the scale having least Crohbach’s alpha .
We plot the table of Cronbach’s alpha for all scales and sub scales as follows
Clearly from the table we can infer that,
Now we need to know as to what is the contributor to the random error in these measurements?
The main reasons for random error are
Clearly in our case ,for all three scales, we have limitation of instruments , change in procedure and environmental factor as well. Thus all these contribute to random error in the scale measure.
Now we have already shown above that as per rule of thumb
So higher the better , Our cronbach’s alpha for each scale is greater than 0.80 (excluding FWB) . Hence we can say that our scale is reliable with good remarks.
Depression scale is excellent while Inner strength and quality of life have good reliability