Authors:Patrick Mair; Helmut Strasser Abstract: Rasch-type item response models are often estimated via a conditional maximum-likelihood approach. This article elaborates on the asymptotics of conditional maximum-likelihood estimates for an increasing number of items, important for modern data settings where a large number of items need to be scaled. Using approximations of the variance–covariance matrix based on Edgeworth expansions, the problem is studied theoretically as well as computationally. In a subsequent step, these results are used to split the large-scale estimation problem into smaller sub-problems involving blocks of items. These item blocks are estimated separately from each other and, finally, merged back into the full parameter vector (divide-and-conquer). By means of simulation studies, and in conjunction with the asymptotic results, it was found that block sizes in the range of 30–40 items approximate the full-scale estimators with a negligible loss in precision. It is also shown how varying block sizes affect the running time needed to fit the model. PubDate: 2018-02-20 DOI: 10.1007/s41237-018-0046-z

Authors:Yoshio Takane; Heungsun Hwang Abstract: With the advent of consistent partial least squares (PLSc), an interest has surged in comparing the quality of various estimation methods in structural equation models. Of particular interest are, beside PLSc, Bentler’s non-iterative confirmatory factor analysis, Hägglund’s instrumental variable (IV) estimation method, and Ihara–Kano’s non-iterative uniqueness estimation method. All of these methods yield consistent estimates of parameters in measurement models (factor loadings and unique variances), but require additional steps to estimate parameters in structural models [covariances among latent variables (LVs) and path coefficients]. These additional steps typically involve calculating LV scores, either correlating them or applying regression analysis, and correcting possible “biases” incurred by the use of LV scores as proxies of true LVs. In this paper, we conduct a Monte Carlo study to evaluate parameter recovery capabilities of the above LV extraction methods in conjunction with subsequent LV score construction and bias correction methods. We also compare these methods against more conventional estimation methods, such as the full least squares and maximum likelihood methods, that estimate parameters in both measurement and structural models simultaneously. In addition, we examine three methods of estimating standard errors (SEs) of estimated parameters from a single data set, the bootstrap method, ordinary least squares regression, and the inverse Hessian method. The SEs are important in assessing the reliability of parameter estimates and in testing their significance. It was found that Hägglund’s method used to extract one LV at a time from each block of observed variables, combined with Croon’s bias correction method, worked best in both parameter recovery and resistance to improper solutions, and that the bootstrap method provided the most accurate estimates of SEs. PubDate: 2017-11-29 DOI: 10.1007/s41237-017-0045-5

Authors:Masahiro Morii; Takayuki Sakagami; Shinya Masuda; Shigetaka Okubo; Yuki Tamari Abstract: Unfortunately the Figure 7 was published incorrectly in the original publication of the article. The corrected version of Figure 7 and figure caption should be as below. PubDate: 2017-11-27 DOI: 10.1007/s41237-017-0044-6

Authors:Sébastien Louvigné; Masaki Uto; Yoshihiro Kato; Takatoshi Ishii Abstract: Contemporary learning theories and their implementations associated with information and communication technologies increasingly integrate social constructivist approaches in order to assist and facilitate the construction of knowledge. Social constructivism also highlights the important role of culture, learning attitude and behavior in the cognitive process. Modern e-learning systems need to include these psychological aspects in addition to knowledge construction in order to connect with long-standing pedagogical issues such as the decrease and lack of motivation for education. Current Social Networking Services (SNS) provide a platform where peers can express their passion, emotion, and motivation toward learning. Therefore, this research utilizes this platform to recommend motivational contents from peers for learning motivation enhancement (i.e., learners’ perception of their goal and purpose for learning). The proposed system consists of an SNS platform for learners to (1) express and evaluate their own goals for learning, (2) observe motivational messages from peers recommended from an LDA-based (latent Dirichlet allocation) model, and (3) evaluate their perceptions on motivational and psychological attributes. The LDA-based model recommends messages expressing diverse purposes for a shared goal by maximizing the topic divergence of Twitter messages. Learners’ self-evaluations show the positive and significant impact of observing diverse learning purposes from peers on intrinsic motivational attributes such as goal specificity, attainability, and on the confidence to achieve the desired outcome. PubDate: 2017-11-16 DOI: 10.1007/s41237-017-0043-7

Authors:Fei Gu; Hao Wu Abstract: In canonical correlation analysis (CCA), the substantive interpretations of the canonical variates are of primary interest to the applied researchers. However, there are two different interpretive approaches used by different researchers—the weight-based approach and the loading-based approach, of which the latter is favored by the majority of researchers in practice. For those who choose the loading-based approach and apply CCA simultaneously to multiple samples, they may wish to test the invariance of the canonical loadings. In this paper, three covariance structure models are defined for CCA. In particular, the first model (i.e., the CCA-W model) corresponds directly with regular CCA, including the canonical correlations and canonical weights as the parameters, while the third model (i.e., the CCA-L model) is in alignment with the loading-based interpretive approach, including the canonical correlations and canonical loadings as the parameters. The CCA-L model is further extended to the unrestricted and restricted SCCA-L models, of which the latter allows one to test the invariance of the canonical loadings. A real example drawn from the sociological literature is provided to illustrate the restrictive SCCA-L model, and some strategies to calculate good starting values for the restrictive SCCA-L model are discussed. PubDate: 2017-10-25 DOI: 10.1007/s41237-017-0042-8

Authors:Augustus Jayaraj; Oluseun Odumade; Stephen Sedory; Sarjinder Singh Abstract: In this paper, we propose a new weighted squared distance while minimizing a distance between the true proportions and the observed proportions of (Yes, Yes), (Yes, No), (No, Yes) and (No, No) answers in the setup of Odumade and Singh (Commun Stat Theory Methods 38: 439–446, 2009) model. The resultant estimator is shown to be an unbiased estimator of the proportion of the sensitive attribute of interest in a population and has smaller variance than the estimator of Odumade and Singh (Commun Stat Theory Methods 38: 439–446, 2009) with the same protection of the respondents. A real data application is also considered. PubDate: 2017-10-24 DOI: 10.1007/s41237-017-0041-9

Authors:Shuzo Abe; Yoshiyuki Okuse Abstract: When two alternatives are compared, the one that becomes the focus of comparison has a relative advantage (resp., disadvantage) when two alternatives share positive (resp., negative) unique features. This “direction of comparison effect” (DCE) has been found to manifest clearly when three conditions are met: when information is retrieved from memory rather than being externally presented, when a consumer has low product involvement but a high need for cognition, and when the attribute set size is large. The authors argue that at least two additional conditions must be met when applying the DCE in marketing: (1) the presence of unique features of market offerings and (2) the absence of situational factors such as commitment to an alternative. Experimental data on choice behavior regarding business hotels provide support for these two conditions. Also, in discussing the results, it has become clear that unnoticed DCE bias should be carefully canceled out in surveys by designing questionnaires to ensure that each alternative randomly becomes the focal alternative. PubDate: 2017-10-11 DOI: 10.1007/s41237-017-0039-3

Authors:Hanako Ohmura Abstract: While egotropic evaluations of economic policy have long been analyzed as reflecting the absolute individual income situation per se, economic assessments should also comprise relative perceptions of others’ status; in other words, voters’ perceptions of economic inequality matter. First, using an internet survey, we examined how perceptions of economic disparity affect income tax rate preferences founded on absolute income. We then tested whether egotropic and sociotropic evaluations separately affect voters’ retrospection on government, or if the easier egotropic evaluation functions as an information cue for the more difficult sociotropic evaluation. The results demonstrated the following: (1) those who perceive economic inequality to exist tend to be punitive toward voters in other income brackets; (2) this punitive tendency is stronger among higher and middle earners; and (3) for the middle earners, absolute income status and for the higher earners, perception of economic inequality mediates the effect of sociotropic evaluations and ultimately influences government approval. PubDate: 2017-10-11 DOI: 10.1007/s41237-017-0040-x

Authors:Tomohito Okabe; Daisuke Nogiwa Abstract: Political party preference is a crucial element in the analysis of economics and political science. However, it is often difficult to investigate the dynamic properties of the individual partisanship due to inaccessibility to panel data. This study proposes a Bayesian approach for estimating Markov dynamics of individual-level partisanship with repeated cross-section data in which the history of respondents’ choice of favored party cannot be observed. The proposed approach identifies individual heterogeneities that affect transitional patterns of partisanship, and replicates the dynamic patterns of individual partisan mobility. Using the proposed method with American survey data, the study shows that age, education and race significantly influence partisan dynamics among Americans over a three-decade period from 1972. PubDate: 2017-09-19 DOI: 10.1007/s41237-017-0034-8

Authors:Masahiro Morii; Takayuki Sakagami; Shinya Masuda; Shigetaka Okubo; Yuki Tamari Abstract: Internet surveys are currently used in many academic and marketing research fields. However, the results for these surveys occasionally show traces of response bias. In our study, we analyzed how response bias appears in lengthy preference judgments. 1042 respondents participated in lengthy sequential preference judgments. Three stimuli series were used: scene pictures, Attneave nonsense shapes, and point-symmetric figures. One hundred stimuli were selected for each series and individually displayed on a computer screen, with presentation order randomized for each respondent. Respondents were then asked to rate their degree of preference for each stimulus. Mean preference scores increased over the first 10–20 trials, then, gradually decreased from the middle to the last trial. Furthermore, participants tended to produce the maximum and minimum score during early trials. These results demonstrated that response bias can be a function of presentation order. PubDate: 2017-09-18 DOI: 10.1007/s41237-017-0036-6

Authors:Koji E. Kosugi Abstract: Using Bayesian inference, this study aims to estimate the magnitude of the cognitive load when a person perceives asymmetric social relations. Some empirical evidence relating to balance theory has shown that a balanced state is comparatively easier to memorize than an unbalanced one. In this study, since a balanced state is defined by structural complexity, an experimental hypothesis was set whereby asymmetric social relationships have different difficulty levels depending on structural complexity. The balanced state of an asymmetric relation as structural difficulty is formally derived from the eigenvalue structure of a Hermitian matrix. Asymmetric triadic relations are modeled as featuring three kinds of structural difficulties according to the eigenvalue decomposition of the Hermitian matrix and pattern-specific difficulties. The differences among the structural difficulties were not sufficiently significant to exceed pattern-specific difficulties, but the Bayes factor of the informational hypothesis of this research yielded positive effects. PubDate: 2017-09-14 DOI: 10.1007/s41237-017-0033-9

Authors:Marcus Selart; Svein Tvedt Johansen Abstract: This study tested a model of the relationship between work performance, employee personality, and perceived career development. The model hypothesized that employee personality (extraversion) modifies the relationship between work characteristics (task, diversity, work feedback) and perceived career development. Results revealed that work characteristics (task diversity, work feedback) was positively related to perceived career opportunities and that extraversion positively moderated the relation between work characteristics (task diversity, work feedback) and perceived career opportunities. Issues related to the common method bias are discussed (see e.g. Podsakoff et al. in J Appl Psychol 88(5):879–903, 2003). PubDate: 2017-09-12 DOI: 10.1007/s41237-017-0035-7

Authors:Haiyan Liu; Zhiyong Zhang Abstract: Measurement error in continuous, normally distributed data is well known in the literature. Measurement error in a binary outcome variable, however, remains under-studied. Misclassification is the error in categorical data in which the observed category is different from the underlying one. In this study, we show, through a Monte Carlo simulation study, that there are non-ignorable biases in parameter estimates if the misclassification is ignored. To deal with the influence of the misclassification, we introduce a model with false-positive and false-negative misclassification parameters. Such a model can not only estimate the underlying association between the dependent and independent variables, but it also provides information on the extent of the misclassification. To estimate the model, a maximum likelihood estimation method based on a Fisher scoring algorithm is utilized. A simulation study is conducted to evaluate the performance, and a real data example is given to demonstrate the usefulness of the new model. An R package is developed to aid the application of the model. PubDate: 2017-08-23 DOI: 10.1007/s41237-017-0031-y

Authors:Tsuyoshi Hatori; Satoshi Fujii; Kazuhisa Takemura Abstract: The objective of this study was to examine the effects of people’s prior choice making on their weights of decision attributes. According to theories on post-decision processes, preferences are likely to be reconstructed in line with a preceding choice. The present study hypothesized that preferences are constructed so that superior (inferior) attributes of a chosen alternative are weighted higher (lower) than before. This hypothesis was tested in a field survey of university students in Japan. The data supported the hypothesis that the weights of the superior (inferior) attributes of the chosen alternative increased (decreased) as time passed since the decision was made. This result suggests the validity of the causal relationship in which choice shapes preference, which is the reverse of what is generally assumed in decision research. PubDate: 2017-07-27 DOI: 10.1007/s41237-017-0028-6

Authors:Shinya Masuda; Takayuki Sakagami; Hideaki Kawabata; Nobuhiko Kijima; Takahiro Hoshino Abstract: Some studies have shown low proportions of Japanese people rating their lives as happy. However, most surveys in Japan revealed that response distributions for subjective well-being items indicated excessive blips in middle categories. In this study, participants responded to four items that were designed to detect careless or unmotivated respondents; these items instructed respondents to choose a certain response category. Of the study’s respondents, about 40% failed to follow instructions for at least one of the four items. Besides, they often chose middle categories for well-being items. When those who did not follow instructions were excluded from the complete sample, middle categories showed low or no blips, and the proportion of respondents with high rating scores increased. Thus, this study suggested that one reason for reportedly low proportions of happy people in Japan was respondents who were not diligent in reading items. PubDate: 2017-07-01 DOI: 10.1007/s41237-017-0026-8

Authors:Maarten Marsman; Lourens Waldorp; Gunter Maris Abstract: Large-scale prediction problems are often plagued by correlated predictor variables and missing observations. We consider prediction settings in which logistic regression models are used and propose a novel approach to make accurate predictions even when predictor variables are highly correlated and only partly observed. Our approach comprises three steps: first, to overcome the collinearity issue, we propose to model the joint distribution of the outcome variable and the predictor variables using the Ising network model. Second, to render the application of Ising networks feasible, we use a latent variable representation to apply a low-rank approximation to the network’s connectivity matrix. Finally, we propose an approximation to the latent variable distribution that is used in the representation to handle missing observations. We demonstrate our approach with numerical illustrations. PubDate: 2017-06-14 DOI: 10.1007/s41237-017-0024-x