On exploratory factor analysis: a review of recent evidence, an assessment of current practice, and recommendations for future use

Exploratory factor analysis (hereafter, factor analysis) is a complex statistical method that is integral to many fields of research. Using factor analysis requires researchers to make several decisions, each of which affects the solutions generated. In this paper, we focus on five major decisions that are made in conducting factor analysis: (i) establishing how large the sample needs to be, (ii) choosing between factor analysis and principal components analysis, (iii) determining the number of factors to retain, (iv) selecting a method of data extraction, and (v) deciding upon the methods of factor rotation. The purpose of this paper is threefold: (i) to review the literature with respect to these five decisions, (ii) to assess current practices in nursing research, and (iii) to offer recommendations for future use. The literature reviews illustrate that factor analysis remains a dynamic field of study, with recent research having practical implications for those who use this statistical method. The assessment was conducted on 54 factor analysis (and principal components analysis) solutions presented in the results sections of 28 papers published in the 2012 volumes of the 10 highest ranked nursing journals, based on their 5-year impact factors. The main findings from the assessment were that researchers commonly used (a) participants-to-items ratios for determining sample sizes (used for 43% of solutions), (b) principal components analysis (61%) rather than factor analysis (39%), (c) the eigenvalues greater than one rule and screen tests to decide upon the numbers of factors/components to retain (61% and 46%, respectively), (d) principal components analysis and unweighted least squares as methods of data extraction (61% and 19%, respectively), and (e) the Varimax method of rotation (44%). In general, well-established, but out-dated, heuristics and practices informed decision making with respect to the performance of factor analysis in nursing studies. Based on the findings from factor analysis research, it seems likely that the use of such methods may have had a material, adverse effect on the solutions generated. We offer recommendations for future practice with respect to each of the five decisions discussed in this paper.

Panicca: panic on cross-section averages

The cross-section average (CA) augmentation approach of Pesaran (2007) and Pesaran et al. (2013), and the principal components-based panel analysis of non-stationarity in idiosyncratic and common components (PANIC) of Bai and Ng (2004, 2010) are among the most popular “second-generation” approaches for cross-section correlated panels. One feature of these approaches is that they have different strengths and weaknesses. The purpose of the current paper is to develop PANICCA, a combined approach that exploits the strengths of both CA and PANIC.

Estimation of factor-augmented panel regressions with weakly influential factors

The use of factor-augmented panel regressions has become very popular in recent years. Existing methods for such regressions require that the common factors are strong, such that their cumulative loadings rise proportionally to the number of cross-sectional units, which of course need not be the case in practice. Motivated by this, the current paper offers an indepth analysis of the effect of non-strong factors on two of the most popular estimators for factor-augmented regressions, namely, principal components (PC) and common correlated effects (CCE).

Measuring community and service provider attitudes to child sexual abuse in remote indigenous communities in Western Australia

This study aimed to evaluate a scale to measure attitudes to child sexual abuse (CSA) in remote Australian Indigenous communities. The scale was developed to gauge attitudes that may be inhibiting the reporting of cases of CSA to police, as well as to evaluate whether interventions that focused on collaborative relationships between community members and police resulted in changes in attitudes. Participants included service providers living outside the community (58%), community members (living within the community; 9%), and service providers who were also community members (33%); 18% of participants identified as Indigenous. Principal components analysis revealed a nonintuitive six-factor solution that did not support the original four concepts. Four intuitive factors emerged from an abridged version of the scale: entrenched issues, personal understanding and knowledge, communication between community and government, and community action. The scale detected significant differences between community status and between Indigenous status groups on some factors.

Influence of intrinsic variation on foraging behaviour of adult female Australian fur seals

Phenotypic variation and individual experience can create behavioural and/or dietary variation within a population. This may reduce intra-specific competition, creating a buffer to environmental change. This study examined how intrinsic variation affects foraging behaviour of Australian fur seals. Foraging movements of 29 female Australian fur seals were recorded using FastLoc GPS and dive behaviour recorders. For each individual, body mass, flipper length and axis length were recorded, a tooth was sampled to determine age, and milk was collected for diet analysis. Clustering of fatty acid dietary analysis revealed 5 distinct groups in the population. Behaviour was described using 19 indices, which were then reduced to 7 principal components (>80% of the behavioural variation). Bayesian mixed effect models were developed to describe the relationship between these components and intrinsic variation. No association was found between diet and age or body shape; however, age had a negative relationship with component 1 (27% of variation). Older females spent less time at-sea and foraged nearer to the colony. Age had an effect on component 5 (7% of variation), which represented haul-outs and dive depth; older females made fewer visits to haul-out sites and dived deeper to the benthos. This suggests that as animals age they are able to utilise prior knowledge to exploit nearby foraging sites that younger animals are either unaware of, or have yet to gain the experience required to efficiently utilise. Mass had a negative effect on components representing the directedness of a foraging trip, suggesting heavier individuals were more likely to travel directly to a foraging site.

Antenatal dietary patterns and depressive symptoms during pregnancy and early post-partum

Perinatal depression is a debilitating disorder experienced during pregnancy and/or the first year post-partum. Recently, maternal dietary intake during pregnancy has emerged as a possible area of intervention for the prevention of mental disorders in women and their offspring. However, the relationship between antenatal diet quality and perinatal depressive symptoms remains poorly understood. The current study explored the predictive role of antenatal diet quality for antenatal and post-natal depressive symptoms. Pregnant women (n = 167) were recruited between February 2010 and December 2011. Women completed the Edinburgh Postnatal Depression Scale at time 1 [T1, mean weeks gestation = 16.70, standard deviation (SD) = 0.91], time 2 (T2, mean weeks gestation = 32.89, SD = 0.89) and time 3 (T3, mean weeks post-partum = 13.51, SD = 1.97) and a food frequency questionnaire at T1 and T2. Diet quality was determined by extracting dietary patterns via principal components analysis. Two dietary patterns were identified: 'healthy' (including fruit, vegetables, fish and whole grains) and 'unhealthy' (including sweets, refined grains, high-energy drinks and fast foods). Associations between dietary patterns and depressive symptoms were investigated by path analyses. While both 'healthy' and 'unhealthy' path models showed good fit, only one significant association consistent with study hypotheses was found, an 'unhealthy' diet was associated with increased depressive symptoms at 32 weeks gestation. Given that this association was cross-sectional, it was not possible to make any firm conclusions about the predictive nature of either dietary patterns or depressive symptoms. Dietary intervention studies or larger prospective studies are therefore recommended.

The associations of vegetable consumption with food mavenism, personal values, food knowledge and demographic factors

Background: Poor dietary choices, in particular low consumption of fruits and vegetables are associated with the prevalence of diet related diseases. Ways to increase consumption are urgently required. This paper examines the associations of demographic, psychographic and food knowledge variables with reported vegetable consumption. Methods: An online questionnaire was administered in late 2012 to a national sample 2146 Australians who were selected to represent the Australian population in terms of age, sex, education and location of residence. It was divided into sections which assessed food knowledge, food involvement, food mavenism, personal values and personality factors, demographic characteristics and reported consumption of 13 vegetables and the total number of servings of vegetables per day. Principal components analyses of the individual vegetable consumption ratings derived three forms of vegetable consumption scores. These and total serving per day were used as dependent variables in a structural equation model to identify pathways between them and their likely antecedents. Major findings: Three types of vegetable consumption were formed:. Salad vegetables (onion, tomato and lettuce);. Dinner vegetables (carrot, peas and beans); and 'Green' vegetables (cabbage, spinach broccoli and cauliflower). Food mavenism, food knowledge, food involvement and equality-universalist values mediated the relationships between demographics and conscientiousness and the vegetable consumption variables. Conclusions: The three types of vegetable consumption and total servings per day were associated with different antecedent pathways. The mediating roles of food mavenism, food knowledge, food involvement and equality-universalist values may present opportunities for health promotion and the horticultural industry to increase population vegetable intake. Further research is required to test these associations via experimental and longitudinal studies and qualitative investigation of the meaning and place of the three forms of vegetable consumption in people's daily lives is recommended.

Effects of season period, team quality, and playing time on basketball players' game-related statistics

The aim of this study was to identify within-season differences in basketball players' game-related statistics according to team quality and playing time. The sample comprised 5309 records from 198 players in the Spanish professional basketball league (2007-2008). Factor analysis with principal components was applied to the game-related statistics gathered from the official box-scores, which limited the analysis to five factors (free-throws, 2-point field-goals, 3-point field-goals, passes, and errors) and two variables (defensive and offensive rebounds). A two-step cluster analysis classified the teams as stronger (69 ± 8 winning percentage), intermediate (43 ± 5 winning percentage), and weaker teams (32 ± 5 winning percentage); individual players were classified based on playing time as important players (28 ± 4 min) or less important players (16 ± 4 min). Seasonal variation was analysed monthly in eight periods. A mixed linear model was applied to identify the effects of team quality and playing time within the months of the season on the previously identified factors and game-related statistics. No significant effect of season period was observed. A team quality effect was identified, with stronger teams being superior in terms of 2-point field-goals and passes. The weaker teams were the worst at defensive rebounding (stronger teams: 0.17 ± 0.05; intermediate teams: 0.17 ± 0.06; weaker teams: 0.15 ± 0.03; P = 0.001). While playing time was significant in almost all variables, errors were the most important factor when contrasting important and less important players, with fewer errors being made by important players. The trends identified can help coaches and players to create performance profiles according to team quality and playing time. However, these performance profiles appear to be independent of season period.

Measuring the stigma surrounding Type 2 diabetes: development and validation of the Type 2 Diabetes Stigma Assessment Scale (DSAS-2)

OBJECTIVE: To develop and validate a self-report measure of perceived and experienced stigma for use with adults with type 2 diabetes: the Type 2 Diabetes Stigma Assessment Scale (DSAS-2). RESEARCH DESIGN AND METHODS: An item pool was drafted based on qualitative data from 25 adults with type 2 diabetes and content from other health-related stigma questionnaires. Thirteen adults with type 2 diabetes completed 57 draft diabetes stigma items and participated in cognitive debriefing interviews. Based on participant feedback, the pool was reduced to 48 items with a 5-point Likert scale (strongly disagree to strongly agree). A total of 1,064 adults with type 2 diabetes completed a survey including these 48 items and other validated measures. Data were subject to principal components analysis and Spearman ρ correlations. RESULTS: The scale was reduced to 19 items, with an unforced three-factor solution indicative of three subscales: Treated Differently (6 items, α = 0.88), Blame and Judgment (7 items, α = 0.90), and Self-stigma (6 items, α = 0.90). A forced one-factor solution supported the calculation of a total score (α = 0.95). Satisfactory concurrent, convergent, and discriminant validity were demonstrated. CONCLUSIONS: The 19-item DSAS-2 is a reliable and valid measure of type 2 diabetes stigma. A rigorous design and validation process has resulted in a relatively brief measure of perceived and experienced stigma in type 2 diabetes. The novel scale has satisfactory psychometric properties and is now available to facilitate much-needed research in this field.

The relationship between house prices and demographic variables: an Australian case study

Purpose – The process for examining the value of house prices in an urban city has given limited attention, if any, to demographic variables associated with urban geography. Although the disciplines of property/real estate and demography have moved closer, little progress has been made when modelling house prices using population-related data in the field of urban geography to explain the level of house prices.

Design/methodology/approach – This paper proposes an innovative model to examine the influence of population variables on the level of house prices. It used a two-stage approach as follows: principal components analysis (PCA) identified social dimensions from a range of demographic variables, which were then retained for further analysis. This information was sourced from two Australian Bureau of Statistics censuses undertaken involving all Melbourne residents during 1996, 2001, 2006 and 2011; multiple regression analysis examined the relationship between the retained factor scores from the PCA (as independent variables) and established residential house prices (as the dependent variable).

Findings – The findings confirm the demographic profile of each household, which is directly related to their decisions about housing location and house prices. Based on a case study of Melbourne, Victoria, it was demonstrated that households with specific demographic characteristics are closely related to a certain level of house prices at the suburban level.

Originality/value – This is an innovative study which has not been previously undertaken for an extended period of time to facilitate an analysis of change over time.

A unified learning algorithm to extract principal and minor components

Recently, many unified learning algorithms have been developed to solve the task of principal component analysis (PCA) and minor component analysis (MCA). These unified algorithms can be used to extract principal component and if altered simply by the sign, it can also serve as a minor component extractor. This is of practical significance in the implementations of algorithms. Convergence of the existing unified algorithms is guaranteed only under the condition that the learning rates of algorithms approach zero, which is impractical in many practical applications. In this paper, we propose a unified PCA & MCA algorithm with a constant learning rate, and derive the sufficient conditions to guarantee convergence via analyzing the discrete-time dynamics of the proposed algorithm. The achieved theoretical results lay a solid foundation for the applications of our proposed algorithm.

Sustainable components of leadership effectiveness in organizational performance

Purpose – The objective is to describe a selection of sustainability components of leadership effectiveness in organizational performance.
Design/methodology/approach – The paper is based on a conceptual discussion. Findings – The paper contributes to descriptive models that address sustainability components of leadership effectiveness in organizational performance. The study highlights some common views that exist in the management literature and in prosperous management practice related to the direct impact of the relationship in organizational performance between leadership and effectiveness. In fact, it also highlights the critical or sceptical views of leadership effectiveness in organizational performance raised in the literature. Research limitations/implications – The paper contends that the actual leadership effectiveness in organizational performance varies over time and across contexts. At times, the achievement in organizational performance is the outcome of prosperous and conscious leadership, while it at other times may be the outcome of poor and deficient leadership. The topic at hand is positioned and limited to the interface that may describe and explain the connection between these two views. Furthermore, it is limited to corporate decision making and business behaviour in relation to leadership effectiveness and organizational performance. Practical implications – The leadership of an organization need not only to be successful today, but they also need to be successful tomorrow to stay in control and to flourish. Quality control and quality assurance are no longer enough for most organizations. They need to build an awareness of the sustainability components into processes of their management and business practices (i.e. internal and external ones) in order to be judged as successful in corporate decision-making and business behaviour in organizational performance in the long term. Originality/value – The principal contributions of the study are a model of timely leadership effectiveness, a model of contextual leadership effectiveness, and a typology of leadership effectiveness in corporate decision-making and business behaviour. These contributions provide theoretical and managerial ideas and insights into the sustainability components of leadership effectiveness in organizational performance.

Evolution of JAK-STAT pathway components : mechanisms and role in immune system development

Background
Lying downstream of a myriad of cytokine receptors, the Janus kinase (JAK) – Signal transducer and activator of transcription (STAT) pathway is pivotal for the development and function of the immune system, with additional important roles in other biological systems. To gain further insight into immune system evolution, we have performed a comprehensive bioinformatic analysis of the JAK-STAT pathway components, including the key negative regulators of this pathway, the SH2-domain containing tyrosine phosphatase (SHP), Protein inhibitors against Stats (PIAS), and Suppressor of cytokine signaling (SOCS) proteins across a diverse range of organisms.

Results
Our analysis has demonstrated significant expansion of JAK-STAT pathway components co-incident with the emergence of adaptive immunity, with whole genome duplication being the principal mechanism for generating this additional diversity. In contrast, expansion of upstream cytokine receptors appears to be a pivotal driver for the differential diversification of specific pathway components.

Conclusion
Diversification of JAK-STAT pathway components during early vertebrate development occurred concurrently with a major expansion of upstream cytokine receptors and two rounds of whole genome duplications. This produced an intricate cell-cell communication system that has made a significant contribution to the evolution of the immune system, particularly the emergence of adaptive immunity.