> GF \pJava Excel API v2.6 Ba==h\:#8X@"1Arial1Arial1Arial1Arial + ) , * `fDC, 4$titletitle[alternative]title[translated]author[google(nobreakline)] contributor[author(nobreakline)]"contributor[scopusid(nobreakline)]date[issued]relation[journaltitle]identifier[issn] citationsidentifier[major]subject(nobreakline) publisheridentifier[thesisdegree]contributor[advisor]relation[ispartofseries]relation[index(nobreakline)]typeabstractidentifier[govdoc]identifier[uri]identifier[doi(doi)]identifier[isbn]identifier[ismn]
identifier$ x貨X ȩ Q, T ) X ` DPThe application of multi-group cognitive diagnostic models and The comparative study on the recovery rate between validation method of Q matrix : the application of emotionalbehavioral problem inspection2015Y P!YtTŐYP YDoctor1Doctoral Thesis0-The purpose of the study investigates the application possibility of Multi-Group Cognitive Diagnosis Models on actual data and the characteristics of the models. This study applies multi-group cognitive diagnostic model for the youth emotionalbehavioral problem measurement data from The Korean Children and Youth Panel survey (KCYPS). It also attempts to compare the features of the validation method of Q matrix. The validation method of Q matrix compares the accuracy through simulation by selecting empirical method and RSS method. The recovery rate is measured in simulation by intentionally conducting misspecification of Q matrix in simulation and checking whether each validation method of Q matrix properly appoints the factor of misspecification. The misspecification conditions are appointed with varied number of misspecification factor, item centeredness, case number.
The main analysis of simulation are as follows. First, the recovery rate of suggestion factor of RSS method is found to be higher than that of method in general conditions. RSS method particularly has higher recovery rate when the misspecification factors are spread into many items. Second, the recovery rate on misspecification of method factors is found to be higher in general conditions. For another 6 conditions excepting the 2 conditions which have the lowest number of case number and misspecification factor number, the recovery rate of method is higher. Third, the frequency of suggestion factors is concentrated on specific items and factors in both methods, and the concentration tendency of RSS method is higher. In addition, to explore the rate between each method of the total frequency of the factors which the modification is suggested throughout 20 sets and the actual misspecified factors among them, RSS method has higher rate in general conditions.
As the results of analysis, RSS method tends to suggest a small number of correct suggestion factors but not to suggest all modification among misspecification factors. method tends to have low accuracy by suggesting a large number of modification factors but to allow the modification of a variety of factors. The accuracy of method highly increases as the number of cases increases; RSS gets less influences from the number of cases but from the number of misspecification factors of Q matrix.
On the other hand, this study applies multi-group cognitive diagnostic models for actual data based on the Q matrix conducted through statistics validation. In order to determine the application suitability of multi-group G-DINA, the rates of G-DINA model, model goodness, and the rate of positive response per group are compared. It is found that the model utilizing gender and self-reported academic achievement as group classification variables has better goodness than G-DINA model. It is also found that the different characteristics exist according to gender and self-reported academic achievement in the results of analyzing the positive responses of the items per group. Therefore, the application of multi-group G-DINA is demonstrated to be appropriate.
The main results from the subject parameter assumed by applying multi-group cognitive diagnostic model to utilize gender and self-reported academic achievement are as follows. First, the applicat< ion of multi-cognitive diagnostic model according to gender shows that male students have the highest rate of possessing excessive behavior, and the rate of possessing attention deficit and aggressive behavior is followed in the rate of possessing emotional behavioral problems. In contrast, female students have difference in the rate of possessing depression and the rate of other factors, and the rate of depression is the highest.
Although externalizing tendency is easy to be observed because it is expressed into interfering with others or causing conflicts or aggressive or disruptive behaviors, internalizing tendency which does not appear as external aspects is relatively difficult to be classified as emotional behavioral problems, and it is likely to be marginalized in relevant programs or services. Therefore, teachers should change the awareness that the behaviors treated as problematic are not the only emotional behavioral problems in classroom. They should observe carefully to understand how the symptoms of depression and anxiety are expressed.
Second, the result of applying multi-group cognitive diagnostic model according to self-reported academic achievement shows that all the factors of the rate of possessive subject per each group increase as the level of academic achievement generally becomes lower. However, the case of anxiety is exceptional; the possessing rate of lower than 70 points group is the highest while the possessing rate of higher than 90 points group is the second-highest. The emotional behavioral problems and the related low academic achievement can cause constant learning deficit, being held back, and school dropout. Such problems in learning can lead to social maladjustment in adulthood that includes a variety of lifestyle problems or crimes (Bullock, Gable, & Melloy, 2005; Lane, 2004). Therefore, discovering emotional behavioral problems in adolescent and working to supplement the academic achievement of low-achievement youths is very significant. In addition, the fact that the rate of possessive subjects of emotional behavioral problems among the group with high self-reported academic achievement is low does not mean that the problems do not exist. It is noticeable that a lot of youths with learning-related anxiety is particularly included.
This study has following implications; it expands the applying rage of cognitive diagnostic model by applying multi-group cognitive diagnostic model for psychological test and demonstrates that the model can provide meaningful results on new field. It also confirms the advantage of assuming process of the model including group effect control, result comparison per group and result utilization. However, applying multi-group cognitive diagnostic model should be conducted after checking whether the collected data contains heterogeneous subgroups or is acquired from homogeneous groups. Analyzing the data from homogeneous groups does not require the application of multi-group cognitive diagnostic model. Rather it decreases the degree of simplicity by increasing the number of parameters that should be assumed. In addition, group classification variables should be the valid thing based on theoretical foundation. Utilizing the variables which are not relevant with the factors to be measured by investigation as group classification variables or classifying the group according to diverse variables and conducting the analysis several times and then reporting the results by selecting the variables in which the differences between groups exist makes the result interpretation invalid and useless for learning improvement or problem-solving.;x貨(Cognitive Diagnosis Models; CDMs)@ ( p 8p tD \(Gierl & Cui, 2008). tǔ x X P ɔ\ D h\. ɔ p 8p tD t ! ƌ \ ŀ| E` 䲔 tp, P ɔ @ 8mt ƌ| i<\ !` 䲔 t. x X P @ 0 \ D t $ ٳ8| XՔ ¬ĳ ǩX. XX @ ¬ ` ٳ 0xXp 9@ i ǩ<\ ` 0 L8t. ¬ x貨D ȩXt 0t \T X\ ļ п DȲ| ƌ \ U` ŀ| \ D .
\Ŕ x貨(Multi-group Cognitiv< e Diagnosis Models)t ȩ . x貨@ ȴ <Ր tȬXՔ X D X D <\ <Ր | \. t| t, @ t 8mը ` X ƥ%D 0\ \ U`X t X. <Ր X DP X 1D E` 0 L8 \ 0\ ɬHD ` .
\ո, x貨X ȩ@ Q,(Tatsuoka, 1983) 1 \. 췘 8m \ t t)X (t <\ xt 8X XǬD X\ <ǔ @ } J. t\ U1D DX0 t Q,D P` ǔ 䲑\ Ĭ )ĳ H , ¬$ Ĭ ) X\ Q,X @ ܴ p|XՔ Ĭt. 췘 )D XՔ @ } J. t )@ Ĭ 1 Lବt \ tp tȬXՔ Q, )D ȩ\ | DPX Q,D T XՔ @ <\ X.
l ̸(real data) \ x貨X ȩ 1 X 1 D <\ \. t| t, \m Dٳ0̌D (pȬ(The Korean Children and Youth Panel Survey; KCYPS)X ̌D $ ٳ8 ! ̸| <\ x貨D ȩ\. \ Q, T )X 1D DPXଐ \. Q, T )@ ) RSS)D X XD t U1D DP\. X@ Q,D Xĳ<\ $X Q, T )t $ ƌ| \ XՔ `D !\. $ pt@ $ ƌX , 8m 1, @ | 첬Xp X.
< XD \ Ȕ @ L . , H ƌ \ `@ pt RSS)t ) Dt X. RSS)@ ҈ $ ƌ 8m |8ȈD L, $ ƌ D L `t X. X, $ ƌ \ `@ pt )t X. @ @ $ ƌX @ P ptD xX 8 6pt )X `t X. K, H ƌX Hĳ P ) P 8m ƌ D . 췘 U1 DP| t \ XX pt@ 䲌 \t. ҈ ) ȩ \ D tǩXƔp D \֩Xt )X 1D 8 E` D t. lX pt @ X | U֥\t \͌ X @ | H` , @ 0x \ U1@ Ő ĳx Ux` D t.
\ո, Ĭ T| p 1\ Q,D <\ ̸ x貨D ȩX. G-DINAX ȩ 1D X0 t | G-DINA iĳ ļ
Q D(D DPX. 1ļ 0 X\ Y1| l\ \֩\ t | G-DINA iĳ

@
!"#$%&'()*+,-./012345789:;<=?@ABCDERoot EntryWorkbooklSummaryInformation(6DocumentSummaryInformation8>