No full-text available
To read the full-text of this research,
you can request a copy directly from the authors.
Discriminant Analysis
Abstract
Discriminant function analysis (DA) allows the researcher to examine and describe simultaneously the differences between 2 or more mutually exclusive groups with respect to several continuous variables. Although there are numerous studies in the literature in which DA would have been appropriate, the technique is seldom used in leisure and recreation research. The present authors review the nature, scope, and use of DA. The topics covered include a description of DA, its relevance to leisure research, data requirements of DA, evaluation of DA results, and the uses of DA. DA is well suited for describing and/or classifying many aspects of leisure behavior, especially regarding 2 or more variables and how 2 or more are similar or dissimilar with respect to those variables. Data requirements are discussed in terms of the nature of the groups used, variable selection, sample size, and other conditions. The results of a study by H. E. A. Tinsley and R. A. Kass (see record
1981-09305-001) are reviewed to provide an example of the uses and concepts of DA. (14 ref) (PsycINFO Database Record (c) 2012 APA, all rights reserved)
... The variables are those the researcher views as potentially important in understanding the nature of group differences; usually they are measured as continuous variables, but discrete variables may also be used on occasion. The variables are called discriminant or discriminator variables (e.g., Brown & Tinsley, 1983), but they are equivalent to predictor variables or independent variables when used for the prediction of group membership. ...
... The example that follows provides an example of this situation. See Brown and Tinsley (1983), Cronbach and Gleser (1965), Meehl and Rosen (1955), Taylor andWeiss (1972), andWiggins (1973) for more extensive discussions of these and other issues involved in classification. ...
... If the discriminant function is to be used for predictive purposes in new populations, it is essential that the sample specificity of the discriminant analysis, and thus its tendency to overestimate the accuracy of classification, be considered. There are several methods of cross validation, including the following (Dillon & Goldstein, 1984;Brown & Tinsley, 1983): (a) cross-validation using a holdout sample; (b) double crossvalidation; and (c) what has been called the jackkm'fe, Umethod, or "leaving-one-out" method. ...
Discriminant analysis is a technique for the multivariate study of group differences. More specifically, it provides a method of examining the extent to which multiple predictor variables are related to a categorical criterion, that is, group membership. Situations in which the technique is particularly useful include those in which the researcher wishes to assess which of a number of continuous variables best differentiates groups of individuals or in which he or she wishes to predict group membership on the basis of the discriminant function (analogous to a multiple regression equation) yielded by the analysis. The method is also useful as a follow-up to a significant analysis of variance. In this article, I describe the method of discriminant analysis, including the concept of discriminant function, discriminant score, group centroid, and discriminant weights and loadings. I discuss methods for testing the statistical significance of a function, methods of using the function in classification, and the concept of rotation functions. The use of discriminant analysis in both the two-group case and the multigroup case is illustrated. Finally, I provide a number of illustrative examples of use of the method in the counseling literature. I conclude with cautions regarding the use of the method and with the provision of resources for further study. (PsycINFO Database Record (c) 2012 APA, all rights reserved)
... The sample size is also required to be checked in this stage. Brown and Tinsley [25] suggested that the ratio between the sample size and number of independent variables (input variables) should be a minimum of 10:1. ...
... According to Brown and Tinsley [25], the ratio between the sample size and number of independent variables should be a minimum of 10:1. Pituch and Stevens [32] recommended a ratio of 20:1, with a minimum of 20 members in the group having the least number of objects. ...
... However, small sample sizes are not recommended either because the idiosyncrasies in the sample will unduly influence the statistical results. Brown and Tinsley (1983) recommended that the total sample size be at least ten times the number of discriminator variables. Stevens (1996) argued that the ratio of cases to variables should be more on the order of 20 to 1. Brown and Wicker (2000, p.214) suggested that good sample sizes be within these two recommendations, with care given to ensuring that the sample size of the developmental sample (used to develop the discriminant function) meets the requirements for the number of cases in each group. ...
... The computation process starts with calculating the probability of classifying cases by chance. This calculation is made using the following equation: P1a1+ p2a2+.....+ pkak where p is proportion of the total sample actually belonging to a group, a is the actual proportion of the cases classified by discriminant analysis into a particular group, and k is the number of the groups (Betz, 1987;and Brown & Tinsley, 1983). In our classification using the developmental sample, p1 and p2 were equal to 0.533 and 0.473, respectively; a1 and a2 were equal to 0.527 and 0.467, respectively; and the chance rate = (0.533 (0.527) + (0.473) (0.467)= 0.502. ...
Two types of hypermarket spenders with multi-patronage behavior were identified; namely, “Concentric” and “Sprinkled.” The objective of this study is to examine which of hypermarket store attributes differentiate between the two types of spenders, and to determine the differences of their demographic characteristics. Six store attributes including the depth of the product assortment, store services, location convenience, sales promotion, prices, and store reputation were examined. Also, five different demogrpaphics were tested including gender, nationality, marital status, education and monthly income.. A cross sectional design with an intercept survey was used. Three hundered cusdtomers were intercepted at diffeerent hypermarket store locations and asked to fill out the survey instrument. Two research hypotheses were tested using the survey data. The interpretation of the discriminant function showed that “concentric” spenders score quite high on store services, moderately on convenience and sales promotion, and low on prices. Both product assortment depth and store reputation were not important to the discriminant function interpretation. Results also indicated that only two demographics were significantly differentiating between the two types of spenders. Several recommendations were made based on the study findings to enable each hypermaqrket store in Kuwait to increase its share of a consumer’s wallet.
... Since Minitab ™ does not offer a test of significance for the proportion of classified observations in each group, the proportion of observations correctly classified based on the climate variables were compared with the expected proportion of correctly classified observations based on chance alone. The expected proportions were calculated using formula (1) that allows for different sample sizes within the groups(Brown & Tinsley 1983). This gives a more accurate estimate of the calculated expected proportions(Brown & Tinsley 1983).where: ...
... The expected proportions were calculated using formula (1) that allows for different sample sizes within the groups(Brown & Tinsley 1983). This gives a more accurate estimate of the calculated expected proportions(Brown & Tinsley 1983).where: PI .... k = the proportion of actual observations belonging to each group al .... k = the proportion of observations classified by the discriminant analysis as belonging to each group k = the number of groups 5: Site discrimination ...
To refine our knowledge and to adequately test hypotheses concerning theoretical and applied aspects of invasion biology, successful and unsuccessful invaders should be compared. This study investigated insect establishment patterns by comparing the climatic preferences and biological attributes of two groups of polyphagous insect species that are constantly intercepted at New Zealand's border. One group of species is established in New Zealand (n = 15), the other group comprised species that are not established (n = 21). In the present study the two groups were considered to represent successful and unsuccessful invaders. To provide background for interpretation of results of the comparative analysis, global areas that are climatically analogous to sites in New Zealand were identified by an eco-climatic assessment model, CLIMEX, to determine possible sources of insect pest invasion. It was found that south east Australia is one of the regions that are climatically very similar to New Zealand. Furthermore, New Zealand shares 90% of its insect pest species with that region. South east Australia has close trade and tourism links with New Zealand and because of its proximity a new incursion in that analogous climate should alert biosecurity authorities in New Zealand. Other regions in western Europe and the east coast of the United States are also climatically similar and share a high proportion of pest species with New Zealand. Principal component analysis was used to investigate patterns in insect global distributions of the two groups of species in relation to climate. Climate variables were reduced to temperature and moisture based principal components defining four climate regions, that were identified in the present study as, warm/dry, warm/wet, cool/dry and cool/moist. Most of the insect species established in New Zealand had a wide distribution in all four climate regions defined by the principal components and their global distributions overlapped into the cool/moist, temperate climate where all the New Zealand sites belong. The insect species that have not established in New Zealand had narrow distributions within the warm/wet, tropical climates. Discriminant analysis was then used to identify which climate variables best discriminate between species presence/absence at a site in relation to climate. The discriminant analysis classified the presence and absence of most insect species significantly better than chance. Late spring and early summer temperatures correctly classified a high proportion of sites where many insect species were present. Soil moisture and winter rainfall were less effective discriminating the presence of the insect species studied here. Biological attributes were compared between the two groups of species. It was found that the species established in New Zealand had a significantly wider host plant range than species that have not established. The lower developmental threshold temperature was on average, 4°C lower for established species compared with non-established species. These data suggest that species that establish well in New Zealand have a wide host range and can tolerate lower temperatures compared with those that have not established. No firm conclusions could be drawn about the importance of propagule pressure, body size, fecundity or phylogeny for successful establishment because data availability constrained sample sizes and the data were highly variable. The predictive capacity of a new tool that has potential for eco-climatic assessment, the artificial neural network (ANN), was compared with other well used models. Using climate variables as predictors, artificial neural network predictions were compared with binary logistic regression and CLIMEX. Using bootstrapping, artificial neural networks predicted insect presence and absence significantly better than the binary logistic regression model. When model prediction success was assessed by the kappa statistic there were also significant differences in prediction performance between the two groups of study insects. For established species, the models were able to provide predictions that were in moderate agreement with the observed data. For non-established species, model predictions were on average only slightly better than chance. The predictions of CLIMEX and artificial neural networks when given novel data, were difficult to compare because both models have different theoretical bases and different climate databases. However, it is clear that both models have potential to give insights into invasive species distributions. Finally the results of the studies in this thesis were drawn together to provide a framework for a prototype pest risk assessment decision support system. Future research is needed to refine the analyses and models that are the components of this system.
... Study 4 (Brown and Tinsley 1983;Costello and Osborne 2005:4). 42 Semantic differential scales were coded such that lower scores were more representative of Democrats and higher scores were more representative of Republicans (Burke and Tully 1977;Burke and Stets 2009;Stets and Carter 2006). ...
For nearly a century, the claim that expertise leads to influence has been axiomatic in social psychology, yet this tradition of research cannot explain why people sometimes resist expert influence. I attempt to fill this gap by synthesizing communications research on the message and its properties, political science research on identity, and the perceptual control system branch of identity theory. I argue that identity is a "missing link" in research on social influence-one that helps explain resistance to expert influence. Focusing on the role of the political identity, I propose that politicized messages convey underlying, implicit meanings relevant to one's political identity. Discrepancies between these meanings and the political identity standard contribute to the resistance of expertise by impacting perceptions of message favorability, negative emotion, and the likelihood of deprecating the message source. To test these claims, survey respondents (50% Democrats, 50% Republicans) used a novel measure of political identity to rate themselves as well as evaluate ten partisan statements, each of which were attributed to an expert source and focused on "hot-button" political issues. Overall, identity-message discrepancies significantly impacted participants' perceptions of message favorability, negative emotions, and likelihood of deprecating expert sources; however, they did not always do so in a manner consistent with identity theory, and effects sometimes differed for Democrat and Republican respondents. With the exception of the impact of identity-message discrepancies on source deprecation, my findings were more nuanced than expected. I discuss the implications of my research for public policy and for future work in social psychology and political science.
... Linear Discriminant Analysis (Brown & Tinsley, 1983) is a method that helps researchers explore and explain the differences between two or more distinct groups by looking at several continuous variables at the same time. Although gender is a demographical variable, it is not advised to use it as a dependent variable for classification. ...
Although it seems that the roles of leaders in sustainable development are extremely important, it is not entirely clear whether the obstacles for women leaders in this context have really changed and to what extent there is still gender discrimination in this domain. In line with this, the study investigates the relationship between the gender of the leaders and the alignment of organizations with the 17 Sustainable Development Goals (SDGs) set by the United Nations. With the growing emphasis on sustainability in business practices, understanding how leadership, particularly CEO gender, influences corporate sustainability initiatives is crucial. The research adopts a quantitative approach to analyze data from organizations led by both male and female CEOs across various industries. By considering the 17 SDGs as independent variables, the study aims to discern whether organizations led by female CEOs exhibit a stronger commitment to specific sustainability goals compared to those led by male CEOs. The analysis seeks to uncover patterns in goal prioritization and explore whether gendered leadership affects sustainability outcomes. The findings are expected to provide insights into how leadership characteristics impact an organization’s sustainability strategies and could suggest future policies and practices aimed at enhancing gender diversity in leadership roles.
... DFA is a standard multivariate analysis for classification and dimensionality reduction, used to identify a linear combination of variables that best separates two or more groups or classes. Its main objective is to maximize the between-class and within-class variance ratio, finding a linear combination of variables that maximizes the differences between the means of the groups while minimizing the variation within each category [41]. We performed DFA to analyze residuals generated in comparisons (1) and (2). ...
The evolutionary development and phylogenetic division between Asian and African cercopithecoids (Cercopithecidae) have attracted significant attention in genetics, molecular biology, behavior, and morphology. However, less emphasis has been placed on how they have evolved morphologically after divergence, approximately 10 million years ago (mya) for Colobinae and 5–7 mya for Cercopithecinae, corresponding to the significant variation and diversity in landscape, climate, habitat, and ecologies between the two continents. This study examines whether such variation and diversity have been reflected in dental morphology. Our findings reveal substantial differences between Hylobatidae and Cercopithecidae, as well as between Colobinae and Cercopithecinae, indicating that size-adjusted dental variation mainly reveals the diversity associated with evolution and phylogenetic inertia. Interestingly, despite the earlier divergence of Afro-Asian colobines, their Euclidean Distance is comparable to that of Afro-Asian cercopithecines. This implies that latecomers (macaques) demonstrate equivalent diversity to colobines due to their extensive dispersion and broader adaptative radiation on the same continent. Colobinae exhibit more developed premolar and molar regions. However, when post-canine teeth are considered alone, Colobinae present a significantly larger molar size than Asian Cercopithecinae but not with the African Cercopihecinae. This contradicts the hypothesis that folivorous primates (Colobinae) have larger post-canine molars than frugivorous ones (Cercopithecinae). The considerable molar size in African Cercopithecinae must be associated with their more protrusive and larger facial structure rather than a specific dietary preference, being less diverse than their Asian counterparts—a trait that has evolved phylogenetically. This study also paves the way for further exploration of facial and cranial differences between the continental groups of Cercopithecinae and Colobinae, delving deeply into diversity variation due to geographical and climatic adaptations.
... The response to be predicted was 0 for pixels in control areas and 1 for pixels in symptomatic areas. The scores produced by the ROSA model were then input into a LDA classification method (Brown and Tinsley, 1983) to produce the probabilities of membership in the two classes, "control" and "infected." The performance of the models was characterized based on a confusion matrix. ...
This article proposes a generic framework to process jointly the spatial and spectral information of hyperspectral images. First, sub-images are extracted. Then each of these sub-images follows two parallel workflows, one dedicated to the extraction of spatial features and the other dedicated to the extraction of spectral features. Finally, the extracted features are merged, producing as many scores as sub-images. Two applications are proposed, illustrating different spatial and spectral processing methods. The first one is related to the characterization of a teak wood disk, in an unsupervised way. It implements tensors of structure for the spatial branch, simple averaging for the spectral branch and multi-block principal component analysis for the fusion process. The second application is related to the early detection of apple scab on leaves. It implements co-occurrence matrices for the spatial branch, singular value decomposition for the spectral branch and multiblock partial least squares discriminant analysis for the fusion process. Both applications demonstrate the interest of the proposed method for the extraction of relevant spatial and spectral information and show how promising this new approach is for hyperspectral imaging processing.
... For H1, H2, and H5, we used the adjusted R² to show how much of the variation of the dependent variable is based on the independent variable (Marill, 2004). Wilk's Lambda (Λ) was used via MANOVA to measure the error variance between groups (e.g., LiKnow compared to LiExp or IPlor compared to IPloi; see results for details) (Brown and Tinsley, 1983). For the leadership level as a moderating variable (H2), we used ordinary least square (OLS) regression via SPSS AMOS (Hayes and Matthes, 2009;Memon et al., 2019). ...
Organizations must align their strategies and renew their capabilities to create business value from advances in artificial intelligence (AI). Ambidextrous leadership is relevant here, as leaders must decide how, when, and where to deploy AI, balancing between its use for increased efficiency (exploitation) and for innovation (exploration). Making such strategic choices calls for AI literacy, comprising certain skills and knowledge. Through an online survey, we identified that leaders’ AI literacy is a strong predictor of achieving ambidextrous leadership and making profound choices about which organizational capabilities should be developed to achieve AI transformation successfully. Our findings demonstrate the importance of balancing investments in both tangible organizational resources, especially data governance, and those that are intangible, such as having an open culture with dynamic workforce capabilities. Notably, leaders’ AI knowledge is more important than their AI experience for making balanced AI-related decisions.
... Overall, 80.6% of the cases were correctly classified by the discriminant analysis. The a priori chance rate to classify candidates correctly according to the formulas provided by Brown and Tinsley (1983) is 66.5%. Comparing the obtained classification rate with the chance rate, the z-test shows that the classification rate significantly differs from chance (z = 19.68; ...
ABSTRACT
Objective
This paper aims at identifying the factors influencing performance in basic flight simulators in the selection of ab initio pilot candidates and to determine the incremental value of these flight simulators in predicting performance in flight training.
Background
Several factors such as cognitive ability, personality, and demographic information are discussed to be associated with performance in basic flight simulators. Many commercial flight operations tend to refrain from using flight simulators in the selection of ab initio pilot candidates in favor of less expensive test batteries.
Method
Study 1 uses 4,340 candidates in the selection of major European airlines to investigate the factors influencing performance in the basic flight simulator. Study 2 then uses 1,753 candidates to evaluate the incremental validity of the basic flight simulator in predicting issues in flight training.
Results
Multiple task ability, gender, and previous flight experience are the strongest predictors of performance in the basic flight simulator. Still, the flight simulator shows significant incremental validity in predicting problems in flight training.
Conclusion
These findings seem to justify the additional use of a basic flight simulator in the selection of ab initio pilot candidates. Despite a strong correlation with more economical multiple task ability tests, the flight simulator still shows incremental validity in predicting success in flight training.
... The response to be predicted was 0 for pixels in control areas and 1 for pixels in symptomatic areas. The scores produced by the ROSA model were then input into an LDA classification method [26] to produce the probabilities of membership in the two classes, "control" and "infected". The performance of the models was characterized based on a confusion matrix. ...
... Joy and Tollefson (1975) give an idea on the financial application of the discriminant analysis and discuss several obstacles on applying discriminant analysis (p.723). According to Brown and Tinsley (1983), Discriminant function analysis is "a technique which allows the researcher to examine and describe simultaneously the differences between two or more mutually exclusive groups with respect to several continuous variables", which is also the definition of discriminant analysis in this paper. The result of discriminant analysis, named Mahalanobis distance, will measure the power of discriminating objects. ...
Companies in different industries have their own operation situation. Then, if companies in a same industry are facing some general same problems or advantages, will some financial performance of companies in an industry be distinguishable from companies in a different industry? Through using discriminant analysis, the paper gives out the outcome of differentiating two industries based on the combination of two ratios. Each industry is represented by a number of companies that is listed and successful. Five rations have been adopted and combined, including Current Ratio, Operating Expense Ratio, Gross Profit Margin and EBIT Margin. The paper finds that the combination of Gross Profit Margin and Operating Expense Ratio have the highest Mahalanobis distance, which means the best way to distinguish two industries However, the number is too small to effective, so people are hard to decide the industry of a company only based on its business performance.
... DFA is a linear method, but it differs from PCA in that it utilizes the cluster information that was given during the training (supervised method), while the PCA does not care about the relationship of the data points with the specified clusters. The DFA method helps to have the best discrimination by reducing distances between samples (variance within classes) and maximizing distances between clusters (variance between classes) [18]. ...
Urbanization is causing people to live close to chicken houses. Their large number leads to a deterioration of air quality, which in turn leads to an increase in complaints from the population. In order to counteract the adverse effects of chicken farming, malodorous air from poultry farms needs to be characterized using appropriate tools. This would give an idea of the degree and source of pollution in order to reduce the impact on the environment. This study aimed to test the ability of the developed e-nose based on six gas sensors to analyze odorous emissions from three poultry farms located in Meknes (Morocco) and Berlin (Germany). This pilot study was also carried out on odorous air samples in one week at different times of the day. Pattern recognition methods such as Principal Component Analysis (PCA), Hierarchical Cluster Analysis (HCA), Support Vector Machines (SVM), and Discriminant Function Analysis (DFA) were used to process the dataset. Moreover, the gas sensors' sensitivity towards hydrogen sulfide, ammonia, and ethanol was also investigated. The finding results reveal that the developed system is able to differentiate the volume fractions of the analyzed gases. Furthermore, the relative humidity values have an effect of less than 1.6% on the gas sensor responses when the relative humidity increases from 15% to 67%. Data processing, using PCA, HCA, and SVM, shows clear discrimination between the odorous air samples collected from the three chicken farms, without any overlap with clean air. The same trend is obtained between odorous air samples collected at different days and times in a poultry farm using DFA and SVMs methods. From the relevant results, it can be concluded that the developed artificial sensory system can clearly classify and assess odorous air from poultry farms.
... This morphological differentiation may be relative to the inherent genetic potential of each breed, geographical isolation and ecological variation (Gizaw et al., 2007;Agaviezor et al., 2012). Wilks' Lambda test, which is the ratio of within-group variability to total variability on the discriminating variables, is an inverse measure of the importance of the discriminant functions (Brown and Tinsley, 1983;Betz, 1987;Huberty, 1994). In this case, the value of Wilks' Lambda was 0.09. ...
... Similarly, Yakubu et al. (2010) was able to correctly allocate more than 99% into their different groups. Additionally, Wilks' Lambda, the ratio of within-group variability to total variability on the discriminating variables, is an inverse measure of the importance of the discriminant functions (Brown and Tinsley, 1983;Betz, 1987;Huberty, 1994). In this case, the value of Wilks' Lambda was 0.29. ...
... The multivariate statistics for differences between the districts was also significant (P<0.0001) in all of the four multivariate tests (Wilks' lambda, Pillai's trace, Hotelling-Lawley trace and Roy's greatest root) for female sample population (Table 7). Wilks' lambda, the ratio of withingroup variability to total variability on the discriminator variables, is an inverse measure of the importance of the discriminant functions [10][11][12][13]. The Wilks' lambda test for the female sample populations was 0.31 (Table 6). ...
... Sample size requirements for RLR and stepwise DFA can be considered as roughly similar. As a minimum criterion for stepwise DFA, it has been suggested to ensure at least one sample per variable in each group, while the total number of samples involved in the training of the discriminant function should be about 10 times the number of variables (Brown and Tinsley 1983). RLR was successfully applied in genetic studies on datasets with >2000 variables containing <70 training samples and a minimum of only 8 observations per group (Wu 2006). ...
Purpose
Soils and sediments can be distinguished based on “composite fingerprints”, i.e., sets of physical and chemical properties that are suitable for discrimination. At present, statistical stepwise variable selection methods are frequently applied to identify composite fingerprints, although they have been seriously criticized. Here, we test regularized logistic regression (RLR) as an alternative approach in the context of a reservoir siltation study where the post-dam facies is to be distinguished from the pre-dam facies.
Materials and methods
The pre- and post-dam facies of four reservoirs located in the Kruger National Park were examined with respect to grain size composition, color, and content of calcium-lactate leachable phosphorus (PCAL). A composite fingerprint was identified applying RLR to training data. The fitted regression model was used for the classification of samples not involved in the training dataset. For comparison, variable selection was performed with stepwise discriminant function analysis (DFA) and samples were classified by applying linear discriminant analysis (LDA). Both approaches were validated by comparing field interpretation and classification results. The analysis was extended based on Monte Carlo simulations and synthetic datasets to quantify uncertainties and to enhance the method comparison.
Results and discussion
RLR and stepwise DFA identify grain size parameters and PCAL content to be particularly useful for the facies discrimination. Neglecting and taking into account a potential sampling bias, both approaches lead to ≤3 and 5% misclassifications, respectively. RLR outperforms stepwise DFA/LDA in Monte Carlo simulations, although misclassification rates do not significantly differ (p = 0.84). RLR uses on average 12% less fingerprint properties. Moreover, RLR-derived probabilities of group membership represent a more reliable measure for classification conclusiveness than probabilities calculated from LDA, which is evident in significantly lower (p < 0.001) probability residuals for misclassified samples. Stepwise DFA/LDA reveals lower misclassification rates than RLR when data fulfill multivariate normality in each group and equal within-group covariance matrices.
Conclusions
RLR is an innovative tool for the discrimination of sediment facies in reservoirs and, more generally, for studies requiring the discrimination of soils and sediments. Although stepwise procedures will in practice often perform similarly well, we discourage their use for the identification of composite fingerprints due to the risk of suboptimal variable selection involving variables with spurious discriminatory power.
... 5 As noted in Footnote 4, the audio search landscape has evolved since we first collected our web data and we were limited to 90 and 127 tokens in our training and test sets, respectively. 6 The sample size should be 10 times the number of attributes according to Brown & Tinsley (1983), 20 times the number of attributes according to Stevens (2002). 7 The methods of regularized discriminant analysis (Friedman (1989) ) or shrinkage discriminant analysis (Ahdesmäki & Strimmer (2010)) have been proposed to improve performance of simple discriminant analysis when the number of attributes exceeds the size of the dataset. ...
We present a new methodological approach which combines both naturally-occurring speech “harvested” on the web and speech data elicited in the laboratory. This proof-of-concept study examines the phenomenon of focus sensitivity in English, in which the interpretation of particular grammatical constructions (e.g. the comparative) is sensitive to the location of prosodic prominence. Machine learning algorithms (support vector machines and linear discriminant analysis) and human perception experiments are used to cross-validate the web-harvested and lab-elicited speech. Results confirm the theoretical predictions for location of prominence in comparative clauses and the advantages using both web-harvested and lab-elicited speech. The most robust acoustic classifiers include paradigmatic (i.e. un-normalized), non-intonational acoustic measures (duration and relative formant frequencies from single segments). These acoustic cues are also significant predictors of human listeners’ classification, offering new evidence in the debate whether prominence is mainly encoded by pitch or by other cues, and the role that utterance-normalization plays when looking at non-pitch cues such as duration.
... Pair-wise squared Mahalanobis distances between sites ( Wilks' Lambda, the ratio of within-group variability to total variability on the discriminator variables, is an inverse measure of the importance of the discriminant functions (Tatsuoka, 1970;Huberty, 1994;Betz, 1987;Brown and Tinsley, 1983). In this case, the value of Wilks' Lambda for the female sample populations was 0.046. ...
This study was conducted in Awi, East and West Gojjam Zones of Amhara region, Ethiopia, to identify and describe the cattle genetic resources and production systems and assess the relative importance of on-going genetic improvement programmes in the area. General linear model procedure employed separately for male and female sample populations showed significant (p<0.0001) differences. The level of association of site with most categorical variables was medium except for a few cases where it was found higher. The Mahalanobis distances between sites were highly significant (p<0.0001). The maximum and minimum distances were observed respectively between Ankasha and Enemay (2.49) and Gozamen and Metekel (42.26) for female sample population and again between Ankasha and Enemay (1.38) and Gozamen and Metekel (32.37) for male sample population. Discriminant analysis was run to classify sample cattle populations from all sites into their respective sites with an overall matching rate of 80.9 % and 79.9% for females and males, respectively. The cluster analysis led to identification of two cattle breed types. These are the Gojjam Highland Zebu (cattle from Gozamen, Ankasha and Enemay) and the Fogera (cattle from Bahir Dar Zuria and Metekel ranch). The Gojjam Highland Zebu cattle are small body sized animals adapted to the highland climatic conditions. On the other hand, the Fogera are relatively bigger animals that belong to the Zenga (Zebu x Sanga) cattle breed group. While the dominant coat colour pattern among the Gojjam Highland Zebu is plain, spotted animals are more common among the Fogera. Results of focus group discussions revealed that most farmers give first priority to adaptive traits. The major functions of cattle according to their importance are supply of traction power, milk production, income generation and beef production. Observed variations in the performances of animals of the same breed possibly indicate the influence of environment. Most of the farmers in the study area practiced natural, unplanned and uncontrolled mating system. Selection of male or female breeding animals is rarely practiced. Under some circumstances, farmers select animals for breeding purposes considering body size, hump size, coat colour, sheath width, udder size and pelvic width. Except in few urban and peri-urban areas where crossbreeding is practiced, pure breeding is the most common breeding system throughout the study area. No breeding programme to improve indigenous cattle has been operational. The average herd size per household in the different sites included in this study varied from 3.0 to 4.4. The male to female ratios were found to be almost equal for the two sexes. The number of adult breeding bulls in each herd gave a ratio of one bull to less than three breeding females. Indicative reproductive and productive performances of cattle in the study area were found low for many of the parameters. While the estimated milk productions for first and second lactations were found to be higher for the Fogera, the lactation periods for first and second parities were longer for the Gojjam Highland Zebu. The average values for all reproductive parameters except lifetime calf production were found to be larger for the Fogera breed. The average ages at first calving were found to be about 49 and 63 months for the Gojjam Highland Zebu and the Fogera, respectively. Indicative average calving intervals were about 24 and 37 months for the two breeds, in that order. The overall average daily milk yield was predicted to be 1.5 and 1.6 liters for the first and second parities respectively with an overall lactation length of 4 and 3.7 months for first and second lactations, respectively. The overall reported average lifetime calf production was estimated to be 5.6. The reported overall average age at culling of breeding animals during normal years was 9.5 and 12 years for male and female animals, respectively. Disease and genetic admixture by other breed types have posed serious threat to the declining population of the Fogera. In many areas, success from introducing improved cattle genotypes seems unlikely unless the problems of feed shortage and disease are overcome. Purebreeding of indigenous cattle types such as the Fogera is a necessity to succeed in crossbreeding ventures. Village breeding prgrammes need to be run side by side with the selection activity.
... To determine the predictor variables that contribute most to the distinction between the two travel experience levels, discriminant analysis was performed on the two identified groups with age, domestic travel level, and international travel level as discriminant variables. Discriminant analysis is a statistical technique that enables researchers to investigate and simultaneously describe the significant differences between two or more mutually exclusive groups on several dependent variables (Brown and Tinsley 1983;Diekhoff 1992). The basic assumption for conducting a discriminant analysis is that the predictor variables, termed discriminant variables, are low in multicollinearity to achieve optimum results (Diekhoff 1992). ...
The purpose of this study lies in the conceptual adjustment of the travel career ladder (TCL) approach to travel motivation. In this context, the study examined the relationship between patterns of travel motivation and travel experience. This research was conducted through two studies: an interview phase to guide the further conceptual development of the travel career approach and a major survey phase for further empirical exploration of the ideas. Overall results suggested that host-site-involvement motivation (e.g., experiencing different cultures) and nature-related motivation (e.g., being close to nature) were more important factors to the more experienced travelers, whereas motivations such as stimulation, personal development, relationship (security), self-actualization, nostalgia, romance, and recognition had a higher priority for the less experienced ones. Importantly, a core of travel motivation factors including escape, relaxation, relationship enhancement, and self-development seem to comprise the central backbone of motivation for all travelers.
... The structure matrix table (Table 2) presents standardized discriminant function coefficients (adjusted for scaling differences), which can be compared to determine the relative strength of discriminator variables in differentiating among groups (M. T. Brown & Tinsley, 1983). Table 3 presents the group centroid output with which discriminant function plots were created. ...
Discriminant function analysis assessed the predictive relevance of nine characteristics measured in sixth grade for differentiating among social identities claimed 4 years later by 616 participants in the Michigan Study of Life Transitions. For females, the first discriminant function, associated with academic motivation, self-esteem, and appearance, accounted for 47% of between-group variability, and the second (sports competence and social skills) accounted for 36%. For males, the first discriminant function (academic ability and self-concept of appearance, in opposite directions) accounted for 54% of variability, and the second (sports competence) accounted for 30%. Findings suggest that differences among individuals with particular high school social identities predate adolescence and point to differences in the primary predictors of male and female identity categories.
Discriminant analysis is a powerful statistical technique used in plant biotechnology for various applications. In this research, five NaCl concentrations as supplements to the sugarcane culture medium in temporary immersion bioreactors (TIBs) were evaluated. Shoot multiplication rate, shoot cluster fresh weight, and levels of aldehydes, carotenoids, phenolics, and chlorophylls were evaluated. Additionally, soluble phenolic concentration was evaluated in the culture medium. Observations indicated that control TIBs (0 mM NaCl) did not exhibit stress symptoms, whereas TIBs with 200.0 mM NaCl were clearly stressed. These two contrasting treatments were utilized to derive Fisher’s linear discriminant function. In contrast, treatments with 50.0, 100.0, and 150.0 mM NaCl displayed intermediate phenotypes. The goal of this report was to determine the NaCl concentrations that induced stress in sugarcane plantlets using discriminant analysis. Fisher’s linear discriminant function successfully distinguished between the two categorical groups: non-stressed plantlets (0 mM NaCl) and stressed plantlets (200.0 mM NaCl). Based on the magnitude of the coefficients in Fisher’s function, increases in the contents of malondialdehyde and other aldehydes were strong indicators of plant stress. Conversely, increases in soluble phenolics, shoot multiplication rate, and fresh weight of shoot clusters indicated non-stressed plant conditions. Results showed that 50.0 and 100.0 mM NaCl caused only mild stress, classifying these treatments as non-stressed plants. However, sugarcane plantlets cultivated in 150.0 mM NaCl TIBs were classified as stressed plants. The application of discriminant function analysis in this report highlights the saline eustress observed in sugarcane cultivated in TIBs.
This study aimed to identify the underlying factors explaining the nature of decisions related to the issuance of the Supplementary Finance Law during the period from 2000 to 2020, This was based on a set of factors, including Gross Domestic Product (GDP), budget deficit, exchange rate, oil price, inflation, and external debt, We employed the discriminant analysis methodology, which is one of the multivariate analysis methods, by focusing on the Stepwise Discriminant Analysis (SDA) method, then estimating the discriminant function to distinguish between two groups, issuing and not issuing, some testing to measure the quality of the classification. The study found a single discriminant function, with 85.7% of the years in both groups correctly classified, the budget deficit was found to have a significant impact on the decision to issue or not issue the Supplementary Finance Law, in addition to the lack of influence of some other economic variables such as gross domestic product and inflation. Therefore, this calls for the development of financial forecasting tools and strengthening governance to improve the accuracy of original financial estimates and reduce reliance on supplementary financial laws to ensure the stability of financial policies in the long term.
This study examined the degree to which parental contextual factors and infant characteristics predicted whether parents read aloud to their 8-month-old infants. Discriminant function analysis revealed that mothers with higher family incomes and those who reported less parenting stress and fewer general hassles were more likely to read to their infants. Gender and temperament of the infant did not significantly predict whether mothers would engage in shared reading. Furthermore, there was no evidence that mothers who reported reading aloud to their infants display more enriching parenting practices in the laboratory. Paternal contextual factors did not discriminate readers from nonreaders, but infant temperament did. Fathers who read aloud had infants who were less soothable and who displayed longer durations of orienting. The possibility that book reading could serve as 1 mediator of the temperament-cognition relationship is discussed.
Sex estimation is an essential step for identification of unidentified human remains. Pelvis, skull, and long bones are commonly used because of their distinct sexual dimorphism. However, these bones are incomplete or missing in some circumstances and other potential bones are needed. Several studies have shown the sexual dimorphism of the patella in specific populations by using an anthropometry method with statistical modeling. The patella has been recognized to be resistant to post-mortem changes. We developed discriminant function equations for sex estimation from measurements of the patella in a modern Central Thai population. Six variables of the patella were measured on 130 skeletons derived from Central Institute of Forensic Science, Thailand. Results showed prediction accuracies of 90% on the left side and 93% on the right side. This study illustrated the potential usage of the patella for sex estimation and indicated size variations of the patella among Thai sub-populations.
Purpose
Like all other natural fibers, the physical properties of cotton also vary owing to changes in the related genetic and environmental factors, which ultimately affect both the mechanics involved in yarn spinning and the quality of the yarn produced. However, information is lacking about the degree of influence that those properties impart on the spinnability of cotton fiber and the strength of the final yarn. This paper aims to discuss this issue.
Design/methodology/approach
This paper proposes the application of discriminant analysis as a multivariate regression tool to develop the causal relationships between six cotton fiber properties, i.e. fiber strength (FS), fiber fineness (FF), upper half mean length (UHML), uniformity index (UI), reflectance degree and yellowness and spinning consistency index (SCI) and yarn strength (YS) along with the determination of the respective contributive roles of those fiber properties on the considered dependent variables.
Findings
Based on the developed discriminant function, it can be revealed that FS, UI, FF and reflectance degree are responsible for higher YS. On the other hand, with increasing values of UHML and fiber yellowness, YS would tend to decrease. Similarly, SCI would increase with higher values of FS, UHML, UI and reflectance degree, and its value would decrease with increasing FF and yellowness.
Originality/value
The discriminant functions can effectively envisage the contributive role of each of the considered cotton fiber properties on SCI and YS. The discriminant analysis can also be adopted as an efficient tool for investigating the effects of various physical properties of other natural fibers on the corresponding yarn characteristics.
For the performance analysis of Fast Moving Consumer Goods (FMCG) industry, discriminatory power of financial ratios are examined by using Wilks' lambda and Multiple discriminant function analysis. For this purpose sample of eighteen FMCG companies listed with Bombay Stock Exchange is taken in to account. Market capitalization is taken as basis for selecting these companies. Data is collected for twelve years ranges from 1 April 2006 to 31 March 2017. For effective implementation of discriminant analysis, firstly average stock market returns are computed from the annual stock prices of the selected companies and average stock market returns are classified in to three groups viz. 'Market Under-Performers', 'Market Average-Performers' and 'Market Out-Performers'. It has been found that revenue from operations/share is the most important ratio and having impact to assess the company's market performance. Debt equity ratio and inventory turnover ratio having moderate impact in assessing the company's stock market performance of companies and dividend payout ratio is the ratio having less impact in assessing the company's stock market performance.
The main focus of the paper is to investigate the relationship between financial ratios and stock returns and to find the ratio(s) which can discriminate between outperformers and underperformers in stock market. Multiple discriminant analysis models with Wilks' Lambda were used on fourteen selected companies from April 1, 2004 to March 31, 2016. Market capitalization was the basis for this selection. For this a basic model was developed to identify the potential Good stock market performer and the Poor stock market performer, based on the Predictor variable viz. Eight ratios which were identified by the Discriminant Analysis. The classification summary shows that a good number of original Groups were correctly classified in to "Good" performer and "Poor" performer. This indicates a very good predictive capacity of the selected ratios. Also it has been concluded that financial variables viz. financial ratios have impact on the Capital Structure of the Automobile companies In India. The Market cap/Net Operating Revenue, Current ratio are the important set of ratio, having impact on financial performance of the companies. Revenue from operations/share, Asset turnover Ratio, Cash earnings Retention Ratio, PBDIT/share having moderate impact on financial performance of companies and Quick ratio and EV/Net operating Revenue are the set of ratios, having less impact on financial performance of companies.
As a departure from previous approaches to the study of work/leisure definitions, perceptions of work and leisure were studied with respect to an activity relatively consistent in terms of task characteristics (high school and college/university basketball). Task attributes which distinguished perceptions of involvement as work, leisure or both were examined, with effects observed for scholarship, gender and year in school. The results are consistent with previous work in the area of intrinsic motivation in that when task factors are held relatively constant, the distinction between perception of participation in sport as work or leisure can be differentiated along the lines of intrinsic/extrinsic motivation.
A matched sample of men and women in the US Navy (N = 1,068) were examined in a study of shipboard medical-care use. Ss completed the Health Beliefs Questionnaire (P. Norman and M. Fitter, 1989, 1991). Results for each separate discriminant function analysis yielded a single statistically significant function for female crew members only. Univariate F-ratio results between female medical users and nonusers show that users have significantly higher ratings of health value, health comparison, and "reason" barriers (e.g., "I already see the doctor a lot"). Male users and nonusers differ in that users have significantly higher ratings of medical-care satisfaction, whereas nonusers have significantly higher "motivation" barriers (PsycINFO Database Record (c) 2014 APA, all rights reserved)
Pharmacists employed by the Northwest Region of Kaiser Permanente were surveyed in 1985 and again in 1987. A secondary analysis of the data from these surveys was undertaken to investigate factors associated with pharmacists' commitment to pharmacy as a career. The first objective was to examine the relationship between pharmacists' commitment to their pharmacy careers and perception of the degree to which their expectations are fulfilled and pharmacists' overall satisfaction with their jobs and pharmacy careers. The second objective was to investigate factors that discriminate between pharmacists who would choose pharmacy again as a career and those who would not. Factors significant in discriminating between those committed to pharmacy as a career and those who were not committed included the degree to which pharmacists' career expectations were fulfilled regarding their sense of doing important and creative work, opportunity for influencing and exercising leadership, personal fulfillment and achievement, recognition by their supervisors, and future prospects for financial security.
Recently, much attention has been given to participation in high risk sport and recreation activities. Psychologists attribute participation to innate personality traits and sociologists suggest that social structure is influential. This paper adopts a life span perspective to investigate patterns of preference for a leisure-based tourist role called the Thrill Seeker. A purposive sample of 1277 New England (USA) residents were surveyed; 124 males and 107 females report taking thrill seeking vacations. Crosstabulations were employed to identify patterns of preference for the Thrill Seeker role for men and women over the life course. Stepwise discriminant function analysis was used to develop profiles of male and female thrill seekers. The results suggest that preference for the Thrill Seeker role peaks for men and women in Early Adulthood and declines thereafter. The findings are discussed within the context of Levinson et al. (1978) model of the adult life course.
This study investigated how an analysis ofintake information, including psychological symptom status, would discriminate between male and female returners and non-returners to post-intake counselling. The sample consisted of261 clients (180 women, 81 men) at a university counsel-ling centre. The results revealed distinct psychological symptom statuses and demographic profiles of female and male non-returners to counselling. Implications of these results and directions for future research are discussed.
This study addresses the issue of data combination in personnel selection. In a pilot study for the selection of trainee pilots for the German Luftwaffe, 99 applicants were assessed using a comprehensive battery of tests that measured inductive thinking, spatial thinking, attentiveness, visual and verbal short-term memory, sensorimotor coordination, and reactive stress tolerance. The global evaluation of the applicants' performance in a flight simulator served as an external criterion. The predictive validity of this test battery was checked by carrying out a discriminant analysis as well as by calculating a neural network. The 2 methods were compared with regard to their classification rate, stability, and separation of correct and incorrect classifications. The results show that artificial neural networks are useful tools for improving the quality of selection procedures for trainee pilots.
Analysis of commitment in long distance running has developed rapidly over the past decade. No attention has been devoted, however, to different commitment patterns of runners. This paper identifies the commitments of ‘professional level’, ‘semi-professional level’ and ‘amateur level’ long distance runners. Discriminant analysis reveals that those of amateur level have relatively weaker personal and structural commitment in comparison with the professional level runners. Semi-professional level runners are in a marginal position: while they have the same personal commitment as the more serious runners, they do not have the parallel structural commitment. The implications of this position are analysed. The preliminary formulation of a theoretical model of sport and leisure commitment is presented, in terms of ‘a circle of commitment’. This involves commitment profile, self-concept, activity levels and achievements.
This study examined attitudes held toward the public and commercial sectors as variables relevant to classifying the recreation participants. The data were collected in a controlled experimental setting. A taxonomy was developed which can be used to categorize individuals on the basis of their attitudes toward the two sectors. Discriminant analysis was used to place individual subjects into one of nine groups in the taxonomy. The discriminant analysis identified statistically significant differences in the attitudes of individuals who were placed into the various groups. Further analysis revealed that the taxonomy may be simplified by collapsing the nine groups into three groups. The three groups include: (1) people who have more favorable attitudes toward the public sector than toward the commercial sector; (2) people who have more favorable attitudes toward the commercial sector than toward the public sector; and (3) people who have similar attitudes toward both sectors. Limitations of the study, implications for managers, and directions for future research are discussed.
We sought to extend the existing research on men who pursue female-dominated careers by examining three masculinity-related constructs which have received little attention in the literature: masculinity ideology, masculine gender-role conflict, and homophobia. Two groups of 50 men, classified as career-traditional or career-nontraditional, were compared in terms of their scores on the Male Role Norms Scale (MRNS), the Gender-Role Conflict Scale (GRCS), and the Index of Homophobia-Modified (IHP-M). Consistent with predictions, results revealed that career-traditional men endorsed antifemininity and toughness norms, reported difficulties concerning restrictive emotionality and restrictive affectionate behavior between men, and indicated homophobic attitudes to a greater extent than did career-nontraditional men. Contrary to prediction, career-traditional and career-nontraditional men did not indicate different endorsement of status norms; difficulties concerning success, power, and competition; or conflicts between work and family relations. Implications for future research and counseling are discussed.
This study suggests that the personal meaning of participation, previously referred to as commitment and involvement, may be appropriately represented by the concept of enduring involvement. The theoretical basis for the development of an instrument to measure enduring involvement in a recreation context is discussed and the results of the administration of the instrument in a beach camping setting in southeast Queensland are reported. These results demonstrate that enduring involvement is appropriately measured by three scales termed attraction, self-expression and centrality. In addition, the study demonstrates that, in the case of beach campers, the derived measure of centrality is moderately predictive of the choice of recreation setting.
The search for risk and danger in outdoor recreation activities is a phenomenon currently facing recreation managers. This study tested the validity of an adventure recreation model of participation. Using the level of engagement as the dependent variable, the model was effective in identifying the components of type of risk, level of risk, social orientation, locus of decision-making, frequency of participation, and preferred environment in the adventure recreation setting. It is suggested that the model can be used by managers and researchers in identifying the types of social, psychological, and physical environments that are preferred by adventure recreationalists relative to their level of experience and engagement in the activity. -Authors
For 102 third world countries, presence and absence of national parks is used as the dependent or classification variable in a discriminant analysis. Eleven variables suggestive of diverse tropical savanna habitats contribute to a single canonical discriminant function. Socioeconomic variables are poor contributors to the function. Seventy‐nine of the countries are properly classified in their group by the canonical discriminant function. Misclassification focuses on sub‐Saharan Africa where many countries have established parks under “improper”; conditions, and strong positive association with the function calls attention to Latin America.
Discriminant analysis (DA; also known as discriminant function analysis) is a powerful descriptive and classificatory technique developed by R. A. Fisher in 1936 to (1) describe characteristics that are specific to distinct groups (called descriptive DA); and (2) classify cases (i.e., individuals, subjects, participants) into pre-existing groups based on similarities between that case and the other cases belonging to the groups (sometimes called predictive DA). The mathematical objective of DA is to weight and linearly combine information from a set of
p-dependent variables in a manner that forces the
k groups to be as distinct as possible. In this chapter we give a thorough and complete discussion of what investigators need to know and do to use DA properly. A brief layout of the specific steps and procedures necessary to conduct a descriptive DA is followed by a more detailed discussion of each step. Information about how to properly interpret the results of a descriptive DA is provided, followed by a discussion of predictive DA. Finally, we describe reporting requirements for publishing results of either a descriptive or predictive DA. (PsycINFO Database Record (c) 2012 APA, all rights reserved)
This research was designed to develop low-fat sugar-free frozen sherbet products containing an acceptable level of soy protein (SP) recommended by the Food and Drug Administration and to determine consumer sensory profile driving consumer acceptance and purchase intent. Four orange-flavored sherbets were manufactured using four levels of SP isolate (6.0, 6.5, 7.5 and 7.9 g/serving) and evaluated by 140 consumers. Consumers evaluated each sample for acceptability of appearance/color, flavor, sweetness, sourness, texture/mouthfeel and overall liking. Consumers also evaluated overall acceptance and purchase intent of these products. All four formulations were overall different (multivariate analysis of variance, Pr > F = 0.0003). Appearance and sweetness were not used by the consumers to differentiate among the four sherbet formulations. Formulations with 7.9 g and 6.0 g SP were equally liked with the mean overall liking score of 5.60–5.66. Formulation with 6.0 g SP had the highest acceptance (65.0%) and purchase intent (55.7%) after consumers had been informed of soy health benefits. Specifically, overall liking and texture were identified as the two most critical attributes affecting overall acceptance and purchase intent of these products.
Consumer interest in healthy eating is increasing. Soy-based products have moved into the marketplace and are becoming increasingly popular. Consumers tend to associate consuming soy with healthy eating habits. In this study, consumer acceptability of low-fat sugar-free sherbets containing soy protein (SP) was evaluated. Development of a frozen dessert with SP would give consumers another venue to satisfy consumer's eating desire and provide them with the health benefits of soy.
American societal norms frequently link alcohol, dating, and sexuality. This cross-sectional study examined the role of alcohol and dating risk factors for sexual assault among a representative sample of female students at a large urban university. Over half of the 1,160 women had experienced some form of sexual assault. Ninety-five percent of these assaults were committed by someone the woman knew and almost half of these assaults involved alcohol consumption by either the man, the woman, or both. Discriminant function analyses indicated that dating, sexual, and misperception experiences and alcohol consumption during these experiences predicted assault group status. Furthermore, alcohol consumption during consensual sex and sexual misperceptions were positively related to alcohol consumption during the sexual assault. The predictors of assault group status were similar for African American and Caucasian women. Theoretical implications are discussed and suggestions are made for combining alcohol and sexual assault prevention programming.
This longitudinal study on 94 families examined the extent to which parent sensitivity, infant affect, and affect regulation at 4 months predicted mother – infant and father – infant attachment classifications at 1 year. Parent sensitivity was rated from face-to-face interaction episodes; infant affect and regulatory behaviors were rated from mother – infant and father – infant still-face episodes at 4 months. Infants' attachment to mothers and fathers was rated from the Strange Situation at 12 and 13 months. MANOVAs indicated that 4-month parent and infant factors were associated with infant – mother but not infant – father attachment groups. Discriminant Function Analysis further indicated that two functions, “Affect Regulation” and “Maternal Sensitivity,” discriminated infant–mother attachment groups; As and B1 – B2s showed more affect regulation toward mothers and fathers than B3 – B4s and Cs at 4 months, and mothers of both secure groups were more sensitive than mothers of Cs. Finally, the association between maternal sensitivity and infant – mother attachment was partially mediated by infant affect regulation.
Many sport organisations that rely on the services of volunteers experience difficulties in the retention of their volunteer labour force. Organisational commitment has been demonstrated to be a significant predictor of task performance, absenteeism and turnover among employees in work organisations. Using a time-lagged research design, the purpose of this study was to examine the temporal influence of organisational commitment and perceived committee functioning in predicting committee member turnover behaviour among volunteers in community sport organisations. Data from a one-year, three-wave longitudinal study of volunteer administrators (N = 262) were used in a discriminant function analysis. It was found that organisational commitment and perceptions about committee functioning measured closest to the time that turnover occurred, were significant but not strong predictors of whether a volunteer stayed with or left the committee of their organisation. It was concluded that organisational commitment was a stronger predictor of turnover than perceived committee functioning, particularly when measured at a point closer to when the turnover occurred, and that organisational commitment may moderate the influence of perceived committee functioning on volunteer turnover behaviour.
The decision to voluntarily quit or remain in a job is one that employees make on a regular basis. The decision process can
range from a job-related event that causes an individual to resign immediately to one that is agonized about for several years;
in limited cases it can be an impulsive act. Within a turnover decision framework, individuals engage in the cognitive process
of thinking about quitting which may or may not translate into quitting; it may also translate into alternative forms of withdrawal
(see Mobey, 1977, and Hanisch, 1995a). What occurs between the time individuals think about quitting until their resignation date? Do individuals who think about
quitting for long periods of time prior to quitting have different attitudes and engage in different behaviors than those
employees who think about quitting and then shortly thereafter resign? Are those individuals with a longer period of time
between thinking about quitting and quitting more disruptive to an organization than those with a shorter time between their
thoughts of quitting and resignation? This research evaluated whether the decision lag was related to behaviors employees
engage in prior to exiting. It also examined how the timing of the first thoughts of quitting relates to employees’ attitudes
in an organization prior to resignation.
The numbers of women choosing to become entrepreneurs has been steadily increasing for the past two decades. Recently, the Internet has provided a new venue for business that may be especially appealing for women. This study was conducted to explore gender- and business-related characteristics of women who own Internet businesses. Two hundred and eight women web entrepreneurs responded to an on-line survey. The results indicated that the types of businesses participants owned could be differentiated based on characteristics of the women web entrepreneurs. Specifically, women owning retail businesses reported less familiarity with computers and geared their Internet businesses more to women. In addition, participants defined the success of their Internet businesses in multiple ways that reflected their current life roles.
We present here a neural network applied to a universal business problem: the estimation of the future fiscal health of a corporation. The commonly used accounting and financial tool for such classification and prediction is a multiple discriminant analysis (MDA) of financial ratios. But the MDA technique has limitations based on its assumptions of linear separability, multivariate normality, and independence of the predictive variables. A neural network, being free from such constraining assumptions, is able to achieve superior results. Our neural network model is the Cascade-Correlation architecture recently developed by Scott E. Fahlman and Christian Lebiere at Carnegie Mellon University. This new approach solves the hidden architecture enigma encountered using other types of neural networks. Also, Cascade-Correlation manages error signals in a manner which significantly improves execution speed. Our research is the first to use Cascade-Correlation for corporate health estimation.
This study investigated the influence of climate variables on insect establishment patterns by using discriminant analysis to classify the climatic preferences of two groups of polyphagous insect species that are intercepted at New Zealand's border. One group of species is established in New Zealand, and the other group is comprised of species that are not established. The discriminant analysis classified the presence and absence of most species significantly better than chance. Late spring and early summer temperatures correctly classified a high proportion of sites containing the presence of both established and nonestablished species. Soil moisture and winter rainfall were less effective discriminating the presence of most of the species studied here. Cold winter air temperature was also a good classifier for the insect species that are not established in New Zealand. This study showed that multivariate statistical techniques such as discriminant analysis can help distinguish the climatic limits of insect distributions over large geographical scales.
When a test of multiple ANOVA is found to be significant, it must be followed by other analyses before a researcher can arrive at an accurate understanding of the data set. Two possibilities for follow-up analyses include univariate ANOVA and discriminant analysis. This article presents the results of a Monte Carlo study (
N = 450) wherein typical, but simple, multivariate data were analyzed by the 2 techniques. Results demonstrate that discriminant analysis is capable of showing the underlying dimensionality of the data as well as determining the contribution of individual variables to the underlying dimensions, whereas ANOVA is limited to specifying the contribution of each variable to group separation. It is argued that when researchers analyze multivariate data, primary goals become interpretation and understanding the data set. It is concluded that discriminant analysis is most suitable for this purpose. (29 ref) (PsycINFO Database Record (c) 2012 APA, all rights reserved)
This investigation has two primary concerns: first, to classify the units of the public and private urban recreation systems into discrete functional classes and identify their basic characteristics, and second, to analyze the distributions of the two systems in terms of absolute and relative locational patterns and compare and contrast their similarities and differences.
Data pertaining to 172 public units and 112 private establishments in Columbia, S.C., were subjected to multiple dimensional scaling, Ward's grouping algorithm, and discriminate analysis. The procedure produced a five-class categorization scheme for each system. The classes of the public system differ more in degree than in kind of facilities and serve more general recreation demands. The private system's classes are functionally more discrete because they offer a fairly narrow range of recreation activities. There is marked contrast in the spatial organization of the two systems. Public recreation sites are concentrated in densely populated areas whereas private recreation enterprises are uniformly distributed.
The public and private recreation systems are complementary and best understood in context of their purpose, structure, and distribution. The purpose of individual units or classes determines the types of facilities, programs, and activities established (i.e., structure). The structure, along users, the location of users, influences the spatial distribution of specific types or classes of recreation sites.
An empirical test of a “goal interference” theory of outdoor recreation conflict is reported. According to the theory, when the behavior of one group of recreationists is incompatible with the social, psychological, or physical goals of another group, a state of goal interference will occur. This will result in conflict perception. Weak support for the model was found in an analysis of conflict between water skiers and fishermen at a midwestern reservoir. Variations in conflict perception among fishermen were somewhat related to variations in recreation goals in a manner predicted by the goal theory. Fishermen who placed greater emphasis on tension release, various forms of escape, and nature enjoyment were more likely to define high-speed boating as “reckless.” Since support for the theory tested was modest, alternative theoretical explanations for recreation conflict are discussed, as is the relevance of conflict studies to recreation management and leisure theory.
This paper provides a brief overview of several lines of research which can now be said to constitute an 'area', called here motivations for leisure. The paper touches on past and current research and then describes the results of two conferences on motivations for leisure. A perspective on this developing area is provided as well as suggestions for future research. This is a blossoming area which should generate considerable basic and applied research, and can constitute part of the core knowledge of the leisure field.- Author
The need-satisfying characteristics of 10 popular leisure activities were examined by two independent factor analyses. The total domain of 45 need satisfiers was subjected to Rao's canonical factor analysis resulting in 10 replicable common factors. Subsequently, the total domain of need satisfiers was reduced to a subset of 27 “leisure activity specific” need satisfiers (i.e., needs characterized by differential satisfaction dependent on the specific leisure activity involved). A principal components factor analysis of the 27 specific need-satisfier dimensions resulted in eight replicable factors. The correspondence between the two sets of factors is discussed, and the 10 leisure activities are described in terms of their potential to satisfy each of the dimensions underlying the leisure activity specific needs.
The use and interpretation of factor analysis is discussed and an example from leisure research is presented. Topics include a conceptual understanding of factor analysis, appropriate data for factor analysis, communality estimates, methods of factor extraction (principal components, principal axis, MINRES, multiple groups and maximum likelihood), the number of factors to extract, factor rotation, factor interpretation, factor scores, and summary comments on the reporting and further uses of factor analysis in leisure research.
The Leisure Activity Questionnaire (LAQ) was administered to a developmental sample of 418 undergraduate college students and to a cross-validation sample of 209 students. The Paragraphs About Leisure (PAL) questionnaire, an alternative form of the LAQ, was administered to a second cross-validation sample of 215 students. Subsequent to the completion of a discriminant function analysis, the hit rates obtained in the two cross-validation samples were determined. It was concluded that the use of the PAL with results reported in terms of factor scores is the most valid and parsimonious measurement strategy of those investigated.
Discriminant analysis Introductory Multivariate Analysis
- L Sanathanan
Discriminant analysis: The study of group differences. Champaign, Ill.: The Institute for Personality and Ability Testing
- M Tatsuoka
Discriminant analysis Handbook of Marketing Research
- D G Morrison