[A factor analysis method for contingency table data with unlimited multiple choice questions].

The purpose of this study is to propose a method of factor analysis for analyzing contingency tables developed from the data of unlimited multiple-choice questions. This method assumes that the element of each cell of the contingency table has a binominal distribution and a factor analysis model is applied to the logit of the selection probability. Scree plot and WAIC are used to decide the number of factors, and the standardized residual, the standardized difference between the sample, and the proportion ratio, is used to select items. The proposed method was applied to real product impression research data on advertised chips and energy drinks. Since the results of the analysis showed that this method could be used in conjunction with conventional factor analysis model, and extracted factors were fully interpretable, and suggests the usefulness of the proposed method in the study of psychology using unlimited multiple-choice questions.

Magnifying endoscopy for diagnosis of residual/local recurrent gastric neoplasms after previous endoscopic treatment.

BACKGROUND: Incomplete resection of gastric neoplasms by endoscopic treatment could lead to residual/local recurrence, which may be difficult to identify. This study aimed to evaluate the usefulness of magnifying endoscopy for identifying and demarcating residual/local recurrent gastric neoplasms after endoscopic treatment. METHODS: Between December 2004 and November 2010, magnifying endoscopy was performed in 15 patients with residual/local recurrent gastric neoplasms. All patients underwent conventional magnifying endoscopy (CME) and enhanced-magnification endoscopy with acetic acid instillation (EME) after conventional endoscopy (CE). Eleven patients additionally underwent magnifying endoscopy using narrow-band imaging (NBI-ME) and a combination of narrow-band imaging and acetic acid instillation (NBI-EME). For each procedure, it was recorded whether the location and circumferential demarcation of the lesions were identified. All lesions were resected by endoscopic submucosal dissection. RESULTS: Eleven lesions were identified using CE. However, two and four additional lesions were identified using CME and EME, respectively. In 11 cases, NBI-ME and NBI-EME were performed and all lesions were identified. Three lesions, which were identified by CME, were not demarcated circumferentially. All 15 lesions were well demarcated by EME and 11 by NBI-ME and NBI-EME. Of the resected specimens, histopathology indicated that ten lesions were differentiated tubular adenocarcinomas and five lesions were adenomas. The histopathological diagnosis of the location and demarcation of all neoplasms corresponded to endoscopic findings. CONCLUSIONS: Magnifying endoscopy techniques (CME, EME, NBI-ME, and NBI-EME) may be useful for identifying and demarcating residual/local recurrent gastric neoplasms after previous endoscopic treatment.

[Effect sizes, statistical power and sample sizes in "the Japanese Journal of Psychology"].

This study analyzed the statistical power of research studies published in the "Japanese Journal of Psychology" in 2008 and 2009. Sample effect sizes and sample statistical powers were calculated for each statistical test and analyzed with respect to the analytical methods and the fields of the studies. The results show that in the fields like perception, cognition or learning, the effect sizes were relatively large, although the sample sizes were small. At the same time, because of the small sample sizes, some meaningful effects could not be detected. In the other fields, because of the large sample sizes, meaningless effects could be detected. This implies that researchers who could not get large enough effect sizes would use larger samples to obtain significant results.

Using non-normal SEM to resolve the ACDE model in the classical twin design.

One of the biggest problems in classical twin studies is that it cannot estimate additive genetic (A), non-additive genetic (D), shared environmental (C), and non-shared environmental (E) effects, simultaneously, because the model, referred to as the ACDE model, has negative degrees of freedom when using Structural Equation Modeling (SEM). Therefore, instead of the ACDE model, the ACE model or the ADE model is actually used. However, using the ACE or ADE models almost always leads to biased estimates. In the present paper, the univariate ACDE model is developed using non-normal Structural Equation Modeling (nnSEM). In SEM, (1st- and) 2nd-order moments, namely, (means and) covariances are used as information. However, nnSEM uses higher-order moments as well as (1st- and) 2nd-order moments. nnSEM has a number of advantages over SEM. One of which is that nnSEM can specify models that cannot be specified using SEM because of the negative degrees of freedom. Simulation studies have shown that the proposed method can decrease the biases. There are other factors that have possible effects on phenotypes, such as higher-order epistasis. Since the proposed method cannot estimate these effects, further research on developing a more exhaustive model is needed.

[An improved method using k-means to determine the optimal number of clusters, considering the relations between several variables].

In this article, we propose a non-hierarchical clustering method that can consider the relations between several variables and determine the optimal number of clusters. By utilizing the Mahalanobis distance instead of the Euclidean distance, which is calculated in k-means, we could consider the relations between several variables and obtain better groupings. Assuming that the data are samples from a mixture normal distribution, we could also calculate Akaike's information criterion (AIC) and the Bayesian information criterion (BIC) to determine the number of clusters. We used simulation and real data examples to confirm the usefulness of the proposed method. This method allows determination of the optimal number of clusters, considering the relations between several variables.

[Scheffé-type paired comparison models for examining correlations between individual preferences for alternatives].

It is difficult to estimate and examine correlations between individual preferences for alternatives using the present Scheffé-type paired comparison models. In this paper, we propose two models that address individual preferences for alternatives. One is a simple model that makes it possible to estimate correlations between individual preferences. The other is an improved model that makes it possible to extract independent components from those correlations. Paired comparison data were collected in a survey about preferences for several new product names. Analysis of this data shows that the proposed models enabled the estimation not only of average preferences for alternatives, but also correlations between individual preferences and loading matrices for independent components. The effectiveness of the proposed methods was confirmed by the interpretations of those estimates.

[New assessment scale based on the type of person desired by an employer].

In many cases, aptitude tests used in the hiring process fail to connect the measurement scale with the emotional type of the person desired by an employer. This experimental study introduced a new measuring method, in which the measurement scale could be adjusted according to the type of person an employer is seeking. Then the effectiveness of this method was verified by comparing the results of an aptitude test utilizing the method and the results of the typical hiring process carried out by the new method in hiring.

Ease of early gastric cancer demarcation recognition: a comparison of four magnifying endoscopy methods.

BACKGROUND AND AIM: Various techniques using magnifying endoscopy (ME) have been developed to enhance images of early gastric cancer (EGC) demarcations, which are often obscure. We investigated four ME methods to determine which is most effective in enhancing the recognition of EGC demarcations: conventional ME (CME), ME with narrow band imaging (NBI-ME), enhanced-magnification endoscopy with acetic acid (EME), and ME with NBI and acetic acid (NBI-EME). METHODS: Thirty-seven successive patients having a total of 40 EGCs participated in the investigation. The endoscope was fixed and magnification images of EGC demarcations in each patient were recorded using four different ME methods (CME, NBI-ME, EME and NBI-EME). Eight experts and eight non-experts scored each of the four images of each lesion for ease of recognition of demarcation (1 to 4, with 4 being easiest). RESULTS: The mean scores of expert and non-expert judges, respectively, for images acquired using each technique were: CME 1.23, 1.24; NBI-ME 2.61, 2.95; EME 2.62, 2.32 and NBI-EME 3.54, 3.50. There were significant differences between the mean scores for the four techniques (P < 0.0001) using one-way repeated-measures anova. In a Bonferroni's multiple comparison, the average scores (expert and non-expert) of images acquired using NBI-EME were significantly higher than those acquired using other methods; images acquired by NBI-ME or EME also scored significantly higher than those by CME. Non-experts also scored NBI-ME images significantly higher than CME and EME images. CONCLUSIONS: Early gastric cancer demarcations were recognized most easily using NBI-EME, and more easily using EME or NBI-ME than CME.

[A endoscopic treatment of penetration of the colon by an ingested fish bone].

An 81-year-old woman was admitted complaining of left lower abdominal pain. Computed tomography on admission revealed foreign a body penetrating descending colon and free air. We extracted the foreign body endoscopically. The foreign body was a sharp-edged fish bone. She was treated by conservative medication without complication.

Gastric adenocarcinoma coexisting with a reddish semipedunculated polyp arising from Helicobacter pylori-negative normal gastric mucosa: a report of two cases.

We experienced two cases of adenocarcinoma coexisting with a hyperplastic polyp arising from Helicobacter pylori-negative normal gastric mucosa. The first case was of a 59-year-old man. Esophagogastroduodenoscopy revealed a protruding lesion measuring 4 mm in diameter on the greater curvature of the middle gastric body. The second case was of a 47-year-old man. Esophagogastroduodenoscopy showed a protruding lesion measuring 5 mm located at the greater curvature of the upper gastric body. The absence of atrophic changes in the entire gastric mucosa was confirmed endoscopically in both cases. Multiple H. pylori tests were all negative. Endoscopically, these protruding lesions appeared as reddish, semipedunculated polyps. Hyperplastic polyps were suspected based on white light endoscopic findings. Magnified endoscopy with narrow band imaging could not predict the coexistence of cancer. However, histopathological examination of the resected specimen revealed focal cancer at the surface of the gastric hyperplastic polyps. Considering the possibility that cancer may be coexisting in polyps, when reddish polyps are detected in H. pylori-negative normal gastric mucosa, it is important to perform a biopsy, or a careful follow-up endoscopy.

[Graphical modeling for constructs: an integrated solution for factor analysis and GM].

This research develops a new method for analyzing relation for factors which combines graphical modeling (GM) and factor analysis. In this method, estimation of the inverse of the variance-covariance matrix is done in the framework of factor analysis, and then the data-model fit is investigated using GM. The partial correlation coefficients of the estimated model are calculated, and the estimation of parameters is repeated until discovery of the worst fit index. In order to confirm the effectiveness of this method, three correlation matrices were analyzed as a real data study. In first and second case, intelligence models of Harman and Turstone were restructured using this method. In third case, EQ model was structured using it. The results show that this method can be apply GM for latent variables and a good assistant to set up path models for factors.

Features of early gastric cancer and gastric adenoma by enhanced-magnification endoscopy.

BACKGROUND: Changes to the mucosal surface of early gastric carcinomas and gastric adenomas as viewed by enhanced-magnification endoscopy with acetic acid have not been investigated thoroughly. Using this technology, we investigated the appearance of the gastric surface patterns of neoplastic and surrounding nonneoplastic mucosa. METHODS: Forty-seven consecutive patients with early gastric carcinomas or gastric adenomas underwent enhanced-magnification endoscopy following 1.5% acetic acid instillation. All biopsy specimens were taken from the area at which the enhanced-magnified endoscopic image was obtained. RESULTS: Surface patterns of gastric tumors and the surrounding mucosa were classified into five types: type I, small round pits of uniform size and shape; type II, slit-like pits; type III, gyrus and villous patterns; type IV, irregular arrangements and sizes of pattern types I, II and III; type V, destructive patterns of types I, II and III. The predominant pattern of the surrounding mucosa was type III, and most type III mucosa had characteristics of intestinal metaplasia. Although all elevated adenomas showed type II or type III surface patterns, both depressed adenomas showed type IV. Elevated carcinomas showed type III (42.9%) or type IV (57.1%) surface patterns, while depressed carcinomas showed type IV (70%) or type V (30%). Although differentiated tubular adenocarcinomas showed type III (10.3%), type IV (86.2%), or type V (3.5%) surface patterns, all of the signet-ring cell carcinomas and poorly differentiated tubular adenocarcinomas showed type V. CONCLUSIONS: Enhanced-magnification endoscopy may be useful for identifying gastric tumors and determining the extent of horizontal spread, especially in tumors of the depressed type.

[The ranking model which estimates the factor scores of residual variables].

In this research, a new directions for utilizing the residual variable score in structural equation modeling was proposed. The residual variable score is created by removing the influence of the comprehensive index, which is strongly related to brand cognition from each trait factor. By using it as a new index, we tried to detect low cognition and high evaluation Brands which are not ranked high in the comprehensive index. It was shown that the ranking by the residual variable score on the same trait has very high correlation between years and same trait, and low correlation with the comprehensive index. Thus, the discriminant validity and stability of the new index were identified. Concerning the interpretation of the new index, it became clear that the brand which shows an increasing trend with the new index has the uniformly low degree of cognition as compared with the descending brand. It indicates that the new index has the capability of detecting brands with low cognition and high evaluation Brand.

Intestinal Anisakiasis Treated Successfully with Prednisolone and Olopatadine Hydrochloride.

The clinical characteristic of gastrointestinal anisakiasis is severe abdominal pain after eating raw fish. Intestinal anisakiasis is more uncommon than gastric anisakiasis. Most patients with intestinal anisakiasis need hospitalization because anisakiasis can cause intestinal obstruction, ileus, peritonitis or intestinal perforation. We report a case of intestinal anisakiasis. A 43-year-old woman presented with symptoms of intermittent abdominal pain 2 days after eating raw fish. Her brother had eaten the same food and had been suffering from gastric anisakiasis. Abdominal ultrasonography in this patient showed localized jejunal wall thickening with dilated lumen of proximal jejunum and ascites. According to the clinical course and examinations, she was diagnosed with intestinal anisakiasis. Administration of prednisolone 5 mg/day and olopatadine hydrochloride 10 mg/day improved her symptoms quickly without hospitalization. Prednisolone was administered for 10 days, and olopatadine hydrochloride was administered for a total of 6 weeks according to ultrasonographic findings. Six months after the treatment, the abdominal ultrasonography demonstrated normal findings. This case demonstrates that ultrasonography was quite useful for the diagnosis and surveillance of intestinal anisakiasis. Furthermore, treatment with corticosteroid and an antiallergic agent could be an option for patients with intestinal anisakiasis.

[Comparison of several expressions of two-way ANOVA for a latent factor: analysis of the relation between music and mood].

The purpose of this research is to provide a two-way analysis of variance (ANOVA) model for a latent factor. Typical psychological studies measure mental states with questionnaires and analyze the variance of the measures into the portions attributable to various sources. This type of research, when conducted under regular ANOVA designs, uses total score as the dependent measures. However, this method is based on the unrealistic presumption that every item on the questionnaire has the same factor loading on the attribute being measured. In this research, we incorporated factor analysis model, and used a latent factor instead of total score as the dependent measure, thereby applying ANOVA under a more realistic assumption. Structural Equation Modeling (SEM) was used to express statistical models. This paper also examined a relation between music and mood, which is a quite popular area of research in psychology of music. To study the possible effects of tonality and key-signature on mood, music was chosen to represent tonality and key-signature conditions.

[Parameter estimation for Rasch type items that have never been used].

This study proposes a new conjoint model with orthogonal array using the notation of the structural equation modeling. It is shown that commonly available computer programs such as LISREL (Jöreskog & Sörbom, 1993) and LISCOMP (Muthen, 1984) can be used to estimate the worth values of conjoint measure and item response curves of the Rasch type and graded response type. One remarkable feature of this method is that the response curves of items that have never been used, can be constructed. The constructed distribution functions were compared with the empirical distribution functions of cross validation data. The proposed method seemed to work well, because these functions were very similar to each other.

[Reliability of students' evaluations of university teaching: an analysis of four-facet data by a generalized model and structural equation modeling].

In our study, we examined the reliability of students' evaluation of university teaching. First, we analyzed four-facet data (teacher x rater x viewpoint x order) by the generalizability theory approach, and estimated the variance components for the facets as a G study. Then, we evaluated the reliability from the point of view of the generalizability coefficient (a reliability-like coefficient that is used when a decision concerns the relative ordering of individuals) and the index of dependability (an index that is used when a decision focuses on the absolute level of an individual's performance independent of others' performance), and examined how those indices change as the number of raters and viewpoints are changed. With these analyses we found conditions needed to maintain sufficient reliability of the evaluation for various situations. We also introduced a new method to evaluate the reliability at each level of facet by using structural equation modeling. By this method, one can examine the reliability of students' evaluation of university teaching more specifically.

[Structural equation modeling using ability parameters: analysis of the situation where item parameters have been estimated by item response theory].

One of the problems with structural equation modeling (SEM) is that the estimation of measurement equation is not separated from the estimation of structural equation. The main aim of this study was to propose a new method to overcome that problem by using ability parameters estimated by item response theory (IRT) as data. According to IRT, the error variance of measurement equation can be easily computed as the reciprocal of the information function. By using the estimates of the error variance, we can fix all parameters in measurement equation and can separate the estimation of structural equation from that of measurement equation. This method also allows us to estimate relations among factor scores with improved precision, because the errors of estimating factor scores are taken into account. The article concludes with a simulation result for verifying the efficacy of this method and an actual numerical.

[Estimation of Cronbach's alpha coefficient in the context of item response theory].

Cronbach's alpha coefficient is an index of reliability for the total score of a set of measurements. The alpha takes a value of 1.0 when the total score variance is perfectly attributable to the common factors running through the test items. In this research, a method of estimating Cronbach's coefficient alpha is proposed for the case where the parameters of dichotomous item response model are known and the ability distribution of subjects is given. In addition, the case of polytomous response model where the item parameters are estimated by graded response model (Samejima, 1969) is also considered, and a method of estimating the alpha coefficient is proposed for that case.