This study was an examination of the effect of type of correlation matrix on the robustness of LISREL maximum likelihood and unweighted least squares structural parameter estimates for models with categorical manifest variables. Two types of correlation matrices were analyzed; one containing Pearson product-moment correlations and one containing tetrachoric, polyserial, and product-moment correlations as appropriate. Using continuous variables generated according to the equations defining the population model, three cases were considered by dichotomizing some of the variables with varying degrees of skewness. When Pearson product-moment correlations were used to estimate associations involving dichotomous variables, the structural parameter estimates were biased when skewness was present in the dichotomous variables. Moreover, the degree of bias was consistent for both the maximum likelihood and unweighted least squares estimates. The standard errors of the estimates were found to be inflated, making significance tests unreliable. The analysis of mixed matrices produced average estimates that more closely approximated the model parameters except in the case where the dichotomous variables were skewed in opposite directions. However, since goodness-of-fit statistics and standard errors are not available in LISREL when tetrachoric and polyserial correlations are used, the unbiased estimates are not of practical significance. Until alternative computer programs are available that employ distribution-free estimation procedures that consider the skewness and kurtosis of the variables, researchers are ill-advised to employ LISREL in the estimation of structural equation models containing skewed categorical manifest variables. / Ph. D.

Les échanges d'information basés sur des nouvelles technologies d'information et de communication (NTIC) offrent des avantages stratégiques aux entreprises. Les pratiques effectives de ces NTIC ne sont cependant que peu connues, de même que leurs impacts sur la performance. L'objectif de la thèse est de présenter, dans le cas de l'échange de données informatisé (EDI), un modèle de recherche permettant une analyse causale entre les pratiques de cette NTIC et la performance. La méthodologie utilisée est de type mixte : une phase qualitative basée sur des études de cas d'entreprises et une phase quantitative, basée sur un modèle d'équations structurelles traité par LISREL. Les résultats montrent que, dans un contexte logistique, les capacités intrinsèques de l'EDI sont insuffisantes pour parvenir à une amélioration de la performance. Celle-ci peut cependant être atteinte lorsque ces capacités sont alignées sur des processus de gestion simples et routiniers, nécessitant peu de coordination organisationnelle.

The improvement of construction systems and management systems do not move together in construction industry, not using the effective management approaches decrease the productivity of the work. The firms both in national and international area in manufacturing industry use supply chain management to preserve their competitive force. As material costs account for a high percentage of the overall cost, the success of a project is heavily dependent on material related factors and processes. In Turkey because of extensively using in-situ concrete in construction, the supply chain for reinforcement and concrete is very important. One of the components of concrete is &ldquo / re-bar&rdquo / and its supply chain, rbSC (rebar supply chain) was investigated within the scope of this research. In this study a framework was developed on the basis of the information from a literature survey and a field-test consisting of face-to-face interviews and questionnaire to determine the factors affecting the rbSC externally and those related to the supply chain participants. The developed rbSC performance framework was analyzed with Structural Equation Modeling rules. LISREL software was used, and the outputs of LISREL were evaluated by relationing with the information obtained from literature survey. The developed framework can be used to maximize both effectiveness of the rbSC process and efficiency of time and cost related issues in the rbSC. rbSC performance was associated with the project performance at the end of the study to reach a higher-level scale. The difference between the performance ratings obtained by self-assessment of a participant and those obtained through assessment done by the others was used to determine the affects of cultural issues on performance. Common reasons for the lower performance were analyzed due to cultural context and recommendations were set down for rbSC participants to solve the problems.

Abstract The purpose of this paper is attempt to build and test the model of consumer behavior, using data communication on the mobile phone. Despite the explosive growth of 2nd generation mobile system (GSM) and rapidly increasing number of consumers who use media (such as photos or access the website) and try to get information they want. We adopt four dimensions to build the structural model of consumer behavior, including need for cognition, product attributes, price factors and purchase intention, and define them as unobserved variables in the LISERL. Finally, we perform relative strength in the Path analysis showing causal model. For model analysis, there are two phase to implement as follows: Firstly, we use Confirmatory Factors Analysis (CFA) to explore the correlation between unobserved (latent) and observed variables. Secondly, the relative relations of four unobserved variables will be confirmed by identification, estimation, test and model-fitting in the Structural Equation Modeling (SEM). We expect that the model of consumer behavior will be appropriately inferred and explain to 3rd mobile communication (3G) in the future.

近年來，由於全球經貿的日益開放、科技快速發展以及知識經濟所帶來的競爭新局，使得各國產業的競爭力皆受到嚴峻的考驗，特別是高科技產業，若無法快速地研發出新科技或新產品，滿足市場客戶的需求，企業勢必遭受淘汰。「研發」儼然已成為維持企業長期成長及競爭地位的不二利器。 然而面對快速成長的研發經費，產業環境的迅速變動、市場顧客需求的快速消長，企業經營者如何讓研發專案發揮其最大的效益，這對於企業未來的成長有長遠且重大的影響。在近來的許多的研究中發現，研發專案組織設計、專案規劃、執行與控制、良好的績效評估、人力資源與成果管理，皆是研發專案關鍵成功的因素，但國內學者並未對這些議題進行綜合性探討，因此，本研究將以國內IC設計產業為研究對象，對研發專案之工作環境、專案領導人能力、團隊運作與專案績效做一整合驗證性的研究，並利用線性結構相關模式（LISREL）來探討四個構面間的關係，所得到的結論如下： 1.在研發專案中，工作環境對專案領導人能力有顯著的正向關係。 2.在研發專案中，工作環境對團隊運作有顯著的正向關係。 3.在研發專案中，專案領導人能力對團隊運作有顯著的正向關係 4.在研發專案中，工作環境對研發專案績效有顯著的正向關係。 5.在研發專案中，團隊運作對研發專案績效有顯著的正向關係。 6.在研發專案中，專案領導人能力對研發專案績效不具顯著的正向關係。

A survey of the field revealed a need for a Confirmatory Factor Analysis of the Personal Orientation Inventory (POI). The two major objectives were to understand the concept of selfactualization and the use of the POI as a measuring instrument. The POI correlation coefficients were used as input data for the computer program, Linear Structural Relations (LISREL), to perform a Confirmatory Factor Analysis. A sample of 495 individuals was used in this research. This sample consisted of both male and female second-year Industrial Psychology students from various race groups. The research concluded that the Root Mean Square Error of Approximation (RMSEA) indicated a poor fit between the empirical model and the empirical reality. This determined that the POI's factor structure could not confirm the theoretical model of selfactualization. It is acknowledged that more research needs to be conducted, taking due consideration of bias variables and cultural diversity. / Industrial and Organisational Psychology / M. Com. (Industrial Psychology)

1980年代末期，美國摩托羅拉公司率先引進六標準差管理系統以改善其產品品質，使得成本大幅下降、獲利上升後，聯訊、德州儀器公司等公司也相繼在1990年代初期導入六標準差管理系統，也從中獲得了相當大的利益，而真正讓此一管理手法受到重視的則是於1995年時，奇異公司將六標準差全面性的導入奇異公司內部的所有流程中，總裁傑克•威爾許更視六標準差為奇異公司最重要的管理突破，也由於如此之重視，奇異公司所獲得的效益也最大，從此，六標準差管理手法盛名遠播，而從1999年起，國內廠商也陸續跟進採用六標準差管理系統。 由於六標準差管理系統擁有採取專案型式以改善流程的特性，本研究即針對此專案特性，配合過往學者所研究出之六標準差關鍵成功因素，發展出工作環境、專案領導人能力、團隊運作情形以及專案績效等四個構面，採用實證問卷調查方式，並利用線性結構關係模式（LISREL）來探討在六標準差架構下，四個構面之整合關聯性，所得到的結論如下： 1.工作環境對專案領導人能力有顯著之正向關係。 2.專案領導人能力對團隊運作有顯著之正向關係。 3.工作環境對專案績效有顯著之正向關係。 4.團隊運作對專案績效有顯著之正向關係。 5.工作環境對團隊運作以及專案領導人能力對專案績效不具顯著之正向關係。

Les modèles d'équations structurelles à variables latentes constituent des modèles statistiques complexes permettant de mettre en relation des concepts non observables. Les deux méthodes d'estimation de ces modèles, que sont, d'une part, l'analyse de la structure de covariance (méthode LISREL) et, d'autre part, l'approche PLS, offrent des solutions à la fois concurrentes et complémentaires. Dans ce travail, un certain nombre de questionnements liés à ce type de modèles et de méthodes sont abordés aussi bien d'un point de vue théorique qu'empirique. Nous étudions la construction du modèle initial par le biais d'algorithmes itératifs, au niveau des relations du modèle de mesure et du modèle structurel. Nous présentons des tests basés sur des permutations afin de comparer des échantillons non appariés sur un modèle donné. Des transformations dites optimales des variables sont utilisées afin de rechercher des relations non linéaires. Finalement, des méthodes permettant le traitement de données manquantes spécifiques induites par des filtres sont décrites. Pour chaque cas, la théorie existante est présentée et des approches novatrices sont introduites. Nous appliquons l'ensemble de ces méthodes dans le cadre de l'analyse de la satisfaction et de la fidélité des clients sur des données provenant d'Electricité de France.

Mémoire numérisé par la Direction des bibliothèques de l'Université de Montréal.

雖然允許變異數－共變數矩陣異質之LISREL-MANOVA模式能克服MANOVA在理論上的的限制（變異數－共變數矩陣同質性的假設），而帶有潛在變項的LISREL-MANOVA能解決MANOVA分析目的上的限制（不能整體考驗多組在潛在變項平均數向量上的差異），但仍有一些問題尚未解決與釐清。其一是MANOVA後續分析的問題，其二是帶有潛在變項之LISRES-MANOVA的統計特性以及切割現象所造成的影響。據此，本論文的主要研究目的有二：（一）發展允許變異數－共變數矩陣異質之LISREL-MANOVA的部份考驗模式以解決MANOVA後續分析中同質性假設的限制，並以模擬分析的方式驗證之；（二）利用模擬分析的方式來探討帶有潛在變項之LISREL-MANOVA的統計特性。 　　依據上述研究目的，本論文具體的研究問題有二：（一）在變異數－變異數－共變數數矩陣同質或變異數－變異數－共變數數矩陣異質的狀態下，允許變異數－共變數異質之LISREL-MANOVA的部份考驗和傳統MANOVA的後續分析（單變量F考驗、區別函數標準化加權和典型變量相關）何者較為優異？（二）是否有切割現象、切割的程度、樣本大小以及估計方法等因子對帶有潛在變項之LISREL-MANOVA的估計有何影響？ 　　針對上述研究問題，本論文主要的研究結果有二：（一）在變異數－共變數同質的狀態下，允許變異數－共變數異質之LISREL-MANOVA的估計結果與單變量F考驗、典型變量相關一致，但較區別函數標準化加權顯著的優異；在變異數－共變數異質的狀態下，允許變異數－共變數異異質之LISREL-MANOVA的估計結果較單變量F考驗、區別函數標準化加權和典型變量相關優異。（二）就帶有潛在變項LISREL-MANOVA而言：1.切割現象會使參數估計值較不精確因而使整體配適考驗容易達顯著。2.切割的比例越大會使整體配適考驗達顯著的機率急遽的上升。3.在切割比例固定下，樣本越大越容易使整體配適考驗達顯著，且會增加WLS估計法犯型Ⅰ錯誤的可能性及GLS和ML估計法的統計考驗力。以及4.在GLS、ML和WLS三種估計法中，WLS較不容易使整配適考驗達顯著，但較容易犯型Ⅰ錯誤；GLS和ML較容易使整體配適考驗達顯著且容易犯型Ⅱ錯誤。 　　最後，本論文根據上述研究結果做成結論，並對應用上及未來研究方向上提出建議。