File Name: exploratory and confirmatory factor analysis understanding concepts and applications .zip
Exploratory factor analysis EFA is a very popular statistical tool that is used throughout the social sciences. Unable to display preview.
- Exploratory Factor Analysis
- Exploratory and Confirmatory Factor Analysis PDF
- Exploratory and Confirmatory Factor Analysis: Understanding Concepts and Applications
Skip to search form Skip to main content You are currently offline.
In statistical terms, factor analysis is a method to model the population covariance matrix of a set of variables using sample data. Factor analysis is used for theory development, psychometric instrument development, and data reduction. Figure 1. Example of factor structure of common psychiatric disorders. Common disorders seem to represent two latent dimensions, internalizing and externalizing disorders.
Exploratory Factor Analysis
The sample was heterogeneous with regard to occupation, sociodemographic data, mental health history, and exposure to traumatic events. Temporal reliability test-retest was high and consistent for different cutoffs. Maximum likelihood exploratory factor analysis EFA was conducted and oblique rotation Promax was applied. The analysis yielded three symptom clusters which accounted for Results are discussed in relation to PTSD theoretical models. Keywords: Posttraumatic stress disorders, validity of tests, reproducibility of results, statistical factor analysis.
A confiabilidade temporal foi alta e consistente para diferentes pontos de corte. Posttraumatic stress disorder PTSD is an anxiety disorder associated with exposure to traumatic events. The most common PTSD symptoms are clustered into a three-dimensional structure: reexperiencing criterion B , avoiding and numbing criterion C , and increased arousal criterion D.
Factor studies have yielded inconsistent results regarding PTSD structure. Inconsistencies are attributed to differences such as: 1 research designs, especially the factor analysis approach; 2 sample characteristics; 3 the types of traumatic events considered; and d PTSD symptom measures. Early PTSD factor structure studies were carried out using exploratory factor analysis EFA , a descriptive technique widely used to group large covariance matrices into a smaller number of latent factors.
More recent studies, in turn, have employed confirmatory factor analysis CFA , an inferential technique. However, CFA provides limited information about factor loadings.
PTSD factor studies can be based on structured psychiatric interviews or self-report measures. Self-report instruments are preferred as a result of their fast application, convenience, and lower cost. Early validation studies conducted by the authors indicated satisfactory levels of test-retest reliability, internal consistency, convergent validity, and diagnostic accuracy.
Further studies using the original PCL version, applied to different samples, also showed adequate levels of reliability and diagnostic accuracy. In addition, it is suitable for any type of traumatic event. Exploratory and confirmatory approaches used in factor studies evaluating the original PCL, however, have yielded conflicting results.
Studies using EFA identified two to four factors and different loading items for each factor. The authors of the PCL presented a model with only two dimensions: a first cluster including reexperiencing, avoidance, and hyperarousal items, and a second cluster comprising numbing and hyperarousal items.
Later, two other exploratory studies conducted with cancer patients also yielded results that were inconsistent with the DSM-IV. Both studies suggest a four-factor structure, however with a non-homogeneous item distribution.
Studies using CFA found a more consistent structure and indicated the presence of three to four factors. The most common solution among such studies is the four- factor. These studies also suggest underlying patterns of symptoms non-congruent with the DSM.
To date, only three studies have investigated the psychometric characteristics of the Portuguese language version of the PCL used in Brazil. In the first study, firefighters and police officers from different regions of Brazil were evaluated.
Results obtained for the firefighters suggest a structure including reexperiencing, avoidance, and numbing-hyperarousal clusters. Police officer results were similar, except for the component hypervigilance D4 , which loaded as a single factor. The difference between the two groups was attributed to the nature of police work, i. The second study used a sample of victims of motor vehicle accidents and unexpectedly found a one-dimensional configuration.
Given the increased variability of responses, it is plausible that this change may have yielded a non-comparable version of the PCL. The third study was conducted with primary care patients living in a poor area of Rio de Janeiro. Results suggested a four-factor structure including the following dimensions: 1 reexperiencing; 2 avoidance; 3 numbing; and 4 hypervigilance.
For the Portuguese language version used in Brazil, difficulties in integrating different contributions have been attributed to the homogeneity of samples included in the two exploratory studies available. Therefore, the objectives of the present study are: 1 to investigate the factor structure of the PCL using EFA in a heterogeneous sample with regard to sociodemographic background, occupational characteristics, and exposure to traumatic events; and 2 to investigate the internal consistency and temporal reliability of the instrument.
The final score may range from 17 to 85 points. According to Weathers et al. The Portuguese language version of the PCL used in Brazil was developed in specifically for civilian contexts. Background and general information about mental health history and exposure to potentially traumatic events was collected using a structured self-report questionnaire. Sociodemographic background information included sex, age, marital status, number of children, years of education, ethnic group and monthly income.
The following information was collected regarding mental health in the last 12 months: psychological or psychiatric treatment, use of psychiatric medications, and diagnosis of mood or anxiety disorders. The questionnaire also inquired about exposure to potentially traumatic events and non-traumatic stressors in the prior 12 months; these questions were framed in accordance with DSM criteria.
Stressful life events non-traumatic stressors included severe financial problems, relationship breakup, unwanted change of address, and social prejudice. Participants were informed about the aims and the voluntary nature of the study and were asked to sign an informed consent form prior to assessment. In both application sessions, participants responded to the questionnaires in the same classrooms where they would subsequently have class health professionals and participants in first aid courses or training activities bank employees.
During the first application test , respondents were given the general questionnaire and the PCL. At the second application retest , which occurred within 15 days of the test, participants were asked to answer the PCL only. Statistical analysis.
An exploratory approach was adopted for factor analysis, to investigate the performance of each item of the Portuguese language version of the PCL used in Brazil. This measure may vary between 0 and 1; high values of sphericity mean that the variables are correlated and the analysis is feasible.
EFA was used to assess the structure of the scale, using the maximum likelihood method with oblique rotation Promax. Oblique rotation is suitable for the investigation of latent constructs for which one expects to extract correlated dimensions.
The conservative criterion suggested by Hair et al. The test-retest reliability of the PCL was assessed using the repeatability index and the Bland-Altman plot. Most participants were female Sixty-three participants Most participants reported mixed ethnicity In relation to mental health history, Exposure to potentially traumatic events was reported by Stressful life events were reported by Mean interval between the two application sessions was 7.
The KMO index 0. Factor analysis indicated three factors with eigenvalues greater than 1. The analysis of the scree plot also indicated a solution with three dimensions. Factor loadings for each factor are presented in Table 2. EFA results indicated three factors. Nevertheless, examination of each item loading showed that the underlying structure differed from that originally presented in the DSM-IV.
The dimensions obtained can be named as follows: 1 reexperiencing, 2 avoidance, and 3 numbing-hyperarousal. An analysis of individual factor loadings also showed that items B5 physical reactions when something reminded you of a stressful experience from the past and D4 being superalert had low factor loadings.
Reliability evaluation considered the total scores obtained for the PCL and different cutoffs described in the literature. Differences in test vs. Figure 1 shows the Bland-Altman plot with the difference between each pair of test and retest scores compared with the mean of the same two scores. The second strategy employed to analyze reliability considered the PCL as a dichotomous variable.
The kappa index was calculated for different cutoffs used in the literature for total scale score, 12,17 minimum number of symptoms according to the DSM-IV, 12 and a combination of both. The kappa indices for the different cutoffs indicate good reproducibility. This study investigated the psychometric characteristics of the Portuguese language version of the PCL used in Brazil, and the results suggested compatibility of the Brazilian instrument with the original version.
Factor validity, internal consistency, and temporal reliability were adequate, adding new information to semantic analyses performed earlier. A positive aspect of this study was that it included individuals heterogeneous not only with regard to their sociodemographic background and occupations, but also in relation to their mental health history, including PTSD symptoms.
This variability avoided the biases discussed by Passos 21 , present when the samples are homogeneous with regard to background and occupational characteristics. Heterogeneous samples are important to guarantee the validity of any psychometric instrument. Our inclusion criteria ensured the possibility to generalize the evidence of validity here reported for the PCL. The evaluation of potentially traumatic events congruent with DSM-IV criterion A and non-traumatic stressful events proved to be advantageous to skirt the limits of uncertainty related to trauma definitions.
Reliability was high for the different cutoffs considered, especially the cutoff suggested by the authors of the PCL. The analysis of reliability of psychometric measures using the approaches described above is not common in PTSD symptom scales. Previous studies employing PCL used only parametric correlations Pearson between different applications. Considering the lack of analyses in the literature, it is not possible to compare the Bland-Altman plot and repeatability index results obtained for the Portuguese language version of the PCL used in Brazil with other versions of the PCL.
Current DSM criteria describe reexperiencing and hyperarousal separately, and avoidance and numbing symptoms in a combined cluster. In this study, symptoms of avoidance and reexperiencing loaded into two dimensions, whereas items related to numbing and hyperarousal formed the third factor.
Taylor et al. For those authors, this two-dimensional solution is related to the theoretical model proposed by Foa et al. The results from our sample, which included subjects, evoke the theoretical model mentioned above, which considered conscious mechanisms vs.
A second possible interpretation is related to the findings of studies that specifically used the original PCL version. Confirmatory studies with four-factor models have indicated more consistent results. However, those studies suggest two models characterized by distinct patterns of symptoms: an emotional numbing model and a dysphoria model. The dysphoria model includes the clusters reexperiencing B1-B5 and avoidance C1-C2 , and combines components of hyperarousal D1-D3 and numbing C3-C7 into one dimension of symptoms observed in several mood and anxiety disorders, called either negative affect, general stress, or dysphoria.
Therefore, divergence is present only with regard to the arrangement of numbing and hyperarousal items.
Exploratory and Confirmatory Factor Analysis PDF
Click here to access. Scholars Portal Books. Table of contents. Internet Archive. Open Library. Please choose whether or not you want other users to be able to see on your profile that this library is a favorite of yours.
Exploratory and Confirmatory Factor Analysis: Understanding Concepts and Applications
The sample was heterogeneous with regard to occupation, sociodemographic data, mental health history, and exposure to traumatic events. Temporal reliability test-retest was high and consistent for different cutoffs. Maximum likelihood exploratory factor analysis EFA was conducted and oblique rotation Promax was applied.
Exploratory and confirmatory factor analysis pdf Exploratory factor analysis EFA could be described as orderly simplification of. Confirmatory factor analysis CFA is a statistical technique used to verify the. Exploratory and Confirmatory Factor Analysis. Items 5 -
Сьюзан взяла себя в руки и быстро подошла к монитору Хейла. Протянула руку и нажала на кнопку. Экран погас. ГЛАВА 39 Росио Ева Гранада стояла перед зеркалом в ванной номера 301, скинув с себя одежду. Наступил момент, которого она с ужасом ждала весь этот день.
Его обстановка напоминала домашнюю - мягкий ковер, высокотехнологичная звуковая система, холодильник, полный напитков и всяческой еды, маленькая кухня и даже баскетбольное кольцо. В отношении шифровалки в АНБ сложилась своеобразная философия. Нет смысла вбухивать миллиарды долларов в дешифровальный компьютер и одновременно экономить на тех, кто работает на этой превосходной технике.
Вы заместитель директора АНБ и обязаны победить. Стратмор медленно поднял голову и как человек, принимающий самое важное решение в своей жизни, трагически кивнул. Сьюзан решительно шагнула во тьму. ГЛАВА 87 Веспа выехала в тихий переулок Каретерра-де-Хуелва.
У меня черный пояс по дзюдо. Беккер поморщился.