Patient-Reported Outcomes: Development and Validation
Demissie Alemayehu, Joseph C. Cappelleri, Birol Emir, Kelly H. Zou in Statistical Topics in Health Economics and Outcomes Research, 2017
How does a researcher determine the number of factors that account for the relationships among the observed variables? That is, how many underlying dimensions (or subscales or domains) are there in the instrument? A widely used approach to determining how many domains or factors to retain is the scree test. The scree test is a rule-of-thumb criterion that involves the creation of a plot of the eigenvalues (i.e., the amount of incremental variance that is accounted for by a given factor) associated with each factor. The objective of a scree plot is to look for a break or dissociation between factors with relatively large eigenvalues and those with smaller eigenvalues; factors that appear before the break are taken to be meaningful and are retained. Researchers often refer to the break as the elbow in the curve of the scree plot.
Self-esteem scale: Translation and validation in Malaysian adults living with asthma
Elida Zairina, Junaidi Khotib, Chrismawan Ardianto, Syed Azhar Syed Sulaiman, Charles D. Sands, Timothy E. Welty in Unity in Diversity and the Standardisation of Clinical Pharmacy Services, 2017
Factor analysis was conducted to group the items of RSES-M sharing same dimensions. The value of KMO test was 0.681 suggesting sample size sufficiency, while Barlett’s test of sphericity x2 (45) = 419.37, p < 0.001 indicated that correlations between items were sufficiently large for factor analysis (Comrey & Lee 2013). The analysis of scree plot supported to retain two factors. These two factors were categorised as positive SE items and negative SE items on the basis of nature of the questions in each cluster as shown in Table 3.
Exploratory Factor Analysis
Douglas D. Gunzler, Adam T. Perzynski, Adam C. Carle in Structural Equation Modeling for Health and Medicine, 2021
A scree plot is a line graph that depicts the relationship between the eigenvalues on the vertical axis and number of factors on the horizontal axis. A scree plot has an elbow shape, with higher values on eigenvalues spread farther apart than lower values as one moves across the x-axis. The number of factors above the elbow indicates the optimal number of factors for the model.
Questionnaire for Turn-on Initiation Preference: Development and Initial Reliability and Validation
Published in The Journal of Sex Research, 2021
Petra Zebroff
The initial factor structure of the 41 QTIP items was examined with an exploratory factor analysis. The Kaiser-Meyer-Olkin measure of sampling adequacy was .943, above the commonly recommended value of .7, and Bartlett’s test of sphericity was significant (χ2 (820) = 28636.03, p < .001). Principal Axis Factoring was used following best practices of EFA for sexuality factor extraction (Sakaluk & Short, 2017). Visual assessment of the scree plot indicated that the eigen values leveled off at 4 factors (see scree plot in online supplementary file Figure S1A). Parallel analysis was conducted using MAP syntax in SPSS, resulting in five components in the original MAP test and four components in the revised MAP test (O’Connor, 2000). The first four factors explained 44% of the variance, were interpretable, and were congruent with the current theory of sexual excitation. Because gender was expected to be a mitigating factor, an EFA was conducted for men and women separately and factor loadings are available in online supplementary file Table S1.
The Internalized Stigma of Substance Abuse Scale for Caregivers: Measuring Substance Use Stigma Experienced by Caregivers
Published in Alcoholism Treatment Quarterly, 2022
Carissa D’Aniello, Rachel Tambling, Beth Russell
Researchers and theorists contended that the commonly accepted cutoff of eigenvalues greater than one may be arbitrary, and result in the over-acceptance of spurious factors (Watson, 2017). An acceptable approach to determining the number of factors is to use the scree test (Cattell, 1966; DeVellis, 2012; Watson, 2017). In the present study, researchers created a scree plot (see Figure 1) in which eigenvalues were plotted in descending order. We then identified the break, or elbow, where the steep slope of larger eigenvalues ends, the smaller eigenvalues plateau, and factors to the left of the elbow are retained (Cattell & Jaspars, 1967). One component had an eigenvalue greater than 10, while others were less than 2, indicating that this component structure fits these data. The one component was extracted that accounted for 41.59% (λ = 12.06) of variance in perceived substance use stigma for caregivers. Items 14 and 26 showed a component loading less than .4 and were marked for removal. Component loadings are presented in Table 1.
Differences between staff groups in perception of risk assessment and risk management of inappropriate sexual behaviour in patients with traumatic brain injury
Published in Neuropsychological Rehabilitation, 2020
Marie Holland, Christine Norman
Principal components analysis with Varimax rotation was conducted on questionnaire items of “causal explanations” in order to reduce the number of critical variables. All 15 items loaded on to one of the five factors which accounted for 70% of the variance (Table 1 highlights all factor loadings of >0.3 for clarity). Visual inspection of the scree plot highlighted eigenvalues >1. The factor labels are as follows: sexual motivation (1), poor management (2), attention seeking (3), negative emotion and unstable mental health (4), and lacks education (5). Two items “trying to seek affection” and “feeling bored at the time” loaded onto two factors; “sexual motivation” and “attention seeking.” Further interpretation resulted in the retention of the items due the distinct cross-over of behaviour. “Not being supervised properly” also loaded onto two factors; “poor management” and “lacks education”. A lack of supervision indicates “poor management” in the context of the environment, but this could also impact the opportunity for new learning, particularly if there is a lack of staff guidance. This item, therefore, remained within the analysis.
Related Knowledge Centers
- Exploratory Factor Analysis
- Factor Analysis
- Reliability
- Biplot
- Parallel Analysis
- Determining The Number of Clusters In A Data Set