http://hdl.handle.net/1903/2262 Sun, 26 May 2013 06:05:58 GMT 2013-05-26T06:05:58Z http://hdl.handle.net/1903/13795 Title: The Impact of Model Selection on Loglinear Analysis of Contingency Tables Authors: Gao, Jing Abstract: It is common practice for researchers in the social sciences and education to use model selection techniques to search for best fitting models and to carry out inference as if these models were given a priori. This study examined the effect of model selection on inference in the framework of loglinear modeling. The purposes were to (i) examine the consequences when the behavior of model selection is ignored; and (ii) investigate the performance of the estimator provided by the Bayesian model averaging method and evaluate the usefulness of the multi-model inference as opposed to the single model inference. The basic finding of this study was that inference based on a single "best fit" model chosen from a set of candidate models leads to underestimation of the sampling variability of the parameters estimates and induces additional bias in the estimates. The results of the simulation study showed that due to model uncertainty the post-model-selection parameter estimator has larger bias, standard error, and mean square error than the estimator under the true model assumption. The same results applied to the conditional odds ratio estimators. The primary reason for these results is that the sampling distribution of the post-model-selection estimator is, in actuality, a mixture of distributions from a set of candidate models. Thus, the variability of the post-model- selection estimator has a large component from selection bias. While these problems were alleviated with the increase of sample size, the interpretation of the p-value of the Z-statistic of the parameters was misleading even when sample size was quite large. To avoid the problem of inference based on a single best model, Bayesian model averaging adopts a multi-model inference method, treating the weighted mean of the estimates from each model in the set as a point estimator, where the weights are derived using Bayes' theorem. Thu, 01 Jan 2009 00:00:00 GMT http://hdl.handle.net/1903/13795 2009-01-01T00:00:00Z http://hdl.handle.net/1903/13661 Title: The processing of past-tense inflection in first language (L1) and second language (L2) Authors: Kim, Say Young Abstract: The present dissertation research investigates how morphologically complex words are processed in isolation and in sentential context by native speakers and second language learners, and how four critical factors in morphological processing (regularity, stem frequency, whole-word frequency, and orthographic similarity) influence this processing. For comparisons between different first languages (Korean L1 and English L1) and between first and second languages (English L1 and English L2), Native Korean Speakers (Exp.1 and 3), Native English Speakers (Exp. 2a and 4a), and Korean Learners of English (Exp. 2b and 4b) were tested. In order to compare the priming effects from words in isolation and words in sentences, sets of inflectional prime and target pairs, one for each language, were used both in a masked priming lexical decision task (Exp.1 and 2) and a self-paced reading task with mask priming (Exp. 3 and 4). The results showed priming effects from inflectional prime and target pairs in both Korean L1 and English L1 when the pairs were presented in isolation, showing morphological sensitivity in both L1 groups. However, when the pairs were embedded in sentences, the priming effect was found only in native English speakers but not in native Korean speakers, implying language-specific differences between Korean and English in processing of inflectional words in sentences. Moreover, even though a similar pattern of priming effects was found for words in isolation, English L2 showed no significant priming effect for words in sentences, consistent with past literature demonstrating less sensitivity to morphological structure in L2. The different patterns of priming effects between the two tasks as well as across the three language groups in the present research were also analyzed in terms of the four morphological factors, and discussed from the perspective of language-specific characteristics. In summary, the present dissertation research examined morphological processing of two typologically different languages in two different reading contexts. The results suggest the importance of language-specific characteristics in various reading conditions in enhancing our understanding of morphological processing in the human mind. Sun, 01 Jan 2012 00:00:00 GMT http://hdl.handle.net/1903/13661 2012-01-01T00:00:00Z http://hdl.handle.net/1903/13660 Title: Group Norms and Intimacy Among Best Friends: A Normative Cross-Sectional Developmental Study Authors: Menzer, Melissa Mary Abstract: The focus of the current dissertation was on 1) the prevalence and nature of observed gossip behavior in the friendships of children in grades five and six, and 2) the associations of observed gossip behavior and perceptions of friendship quality. Scholars have argued that gossip is a normal part of communicative development and it also has been linked to perceptions of close and positive friendship (Gottman & Mettetal, 1986). The findings of the current dissertation indicated that gossip was prominent in children's conversations with their best friends, and that different forms of gossip behavior were evident. Results also confirmed the association of gossip and perceptions of friendship quality (Parker & Gottman, 1989; Sullivan, 1953), and that these relations were stronger for girls than for boys (Foster, 2004; Gottman & Mettetal, 1986; Leaper & Holliday, 1995). In addition, these relations varied depending on other contextual factors, such as whom the gossip was about. Generally, gossip functioned in two apparently contradictory ways for the friendships of girls. On the one hand, gossip was associated with positive aspects of friendship quality. On the other hand, gossip was also associated with negative aspects of friendship quality, such as conflict. Perhaps gossip was more important for the friendships of girls due to the motivations and importance of friendships for girls. In other words, girls are argued to focus their relationship efforts on building close dyadic relationships that involve high levels of disclosure and conversation whereas boys are argued to engaged in more activities that do not require as much disclosure or conversation (e.g., sports, video games; Findlay & Coplan, 2008; Leaper & Smith, 2004; Schneider & Tessier, 2007). Moreover, it may be that conflict resulted from greater engagement and higher frequencies of interaction within the friendship and thus may not necessarily indicate relationship difficulties. The results of the current dissertation highlighted the complexity of the ways in which gossip and perceptions of friendship quality were inter-related in the friendships of children, as well as provided direction for further investigations of the general functions of gossip. Sun, 01 Jan 2012 00:00:00 GMT http://hdl.handle.net/1903/13660 2012-01-01T00:00:00Z http://hdl.handle.net/1903/13616 Title: A SYSTEMATIC INVESTIGATION OF WITHIN-SUBJECT AND BETWEEN-SUBJECT COVARIANCE STRUCTURES IN GROWTH MIXTURE MODELS Authors: Liu, Junhui Abstract: The current study investigated how between-subject and within-subject variance-covariance structures affected the detection of a finite mixture of unobserved subpopulations and parameter recovery of growth mixture models in the context of linear mixed-effects models. A simulation study was conducted to evaluate the impact of variance-covariance structure difference, mean separation, mixture proportion and sample size on parameter estimates from growth mixture models. Data were generated based on 2-class growth mixture model framework and estimated by 1-, 2-, and 3-class growth mixture models using Mplus. Bias, precision and efficiency of parameter estimates were assessed as well as the model enumeration accuracy and classification quality. Results suggested that sample size and data overlap were key factors influencing the convergence rates and possibilities of local maxima in the estimation of GMM models. BIC outperformed ABIC and LMR in identifying the correct number of latent classes. Model enumeration using BIC could be improved by increasing sample size and/or decreasing overall data overlap, and the latter had more impact. Relative bias of parameters was smaller when subpopulation data were more separated. Both the magnitude of mean and variance-covariance separation and variance-covariance differences impacted parameter recovery. Across all conditions, parameter recovery was better for intercept and slope estimates than variance and covariances estimates. Entropy values were as high as the acceptable standards suggested by previous studies for any of the conditions even when data were very well-separated. Class membership assignment was more accurate when mean growth trajectories were more different among subpopulations and mixing proportions were more balanced. Sun, 01 Jan 2012 00:00:00 GMT http://hdl.handle.net/1903/13616 2012-01-01T00:00:00Z