Office Online
    Online Submission
    Office Work
    Peer Review
    Editor Work
  Journal Online
    Forthcoming Articles
    Current Issue
    Advanced Search
    TOP Read
    TOP Download
    Email Alert
  • Table of Content
       , Volume 43 Issue 06 Previous Issue    Next Issue
    For Selected: View Abstracts Toggle Thumbnails
    Effect of Inter-Item Lag on Spoken Word Repetition: An Event-Related Potential Study
    HUANG Xian-Jun,ZHANG Qin,DING Jin-Hong,GUO Chun-Yan
    . 2011, 43 (06): 599-607.  
    Abstract   PDF (463KB) ( 1482 )
    Not all the knowledge or behaviors of human beings and animals are acquired from the first exposure. Repetitive study, as a more common learning approach, plays an important role in getting new information into long term memory. When stimulus presented repeatedly, repetition effect arises as a sign of learning and memory. Previous researches have found that inter-item lag is one of the most important determining factors of the effect. However, it is still not clear what kind of repetition effect can be produced when it comes to phonological materials. Thus, repetition effect of Chinese spoken word was investigated in this study.
    Using a continuous recognition paradigm, 180 bi-syllable spoken words were repeated after 0, 1 or 8~10 intervening items (lags). Electroencephalogram was recorded when 16 right-handed participants were required to judge whether a spoken word had been heard before in the experiment.
    The behavior results show that as inter-item lags increase, reaction time facilitation of repeated words decreased gradually till even disappeared. The ERP data show that compared to their first presentation, repeated words elicit widespread positive waveforms. And the ERP repetition effects occur later as lags increase: the old-new effect begins from about 200ms at Lag0, 300ms at Lag1 and 400ms at Lag8.
    These results show that the repetition effects of spoken word at different lags are modulated at least by three different neural mechanisms, which reflect facilitation of phonological process, semantic process and episodic information retrieving respectively.
    Related Articles | Metrics
    The Units Saccade Targeting Based on and Words Procession Style in Chinese Reading: Evidences from Disappearing Text
    LIU Zhi-Fang,ZHANG Zhi-Jun,ZHAO Ya-Jun
    . 2011, 43 (06): 608-618.  
    Abstract   PDF (371KB) ( 1093 )
    It is well known that words were most important for alphabetic scripts reading (O’Regan & Jacobs, 1992). However, unlike alphabetic languages, Chinese texts consist of characters which take the same amount of horizontal extent, and there are no explicit boundaries between the words. Yang and McConkie (2004) proposed that Chinese readers developed oculomotor strategies that did not involve words. But recently, there were strong evidences supporting that word was the basic unit not only for comprehension of text but also for saccade target selecting (Bai, Yan, Liversedge, Zang, & Rayner, 2008; Yan, Kliegl, Richter, Nuthman, & Shu, 2010). We employed two disappearing text experiments aimed to resolve following confuses: First, was target selection of saccade based on characters or words? Second, if words, whether or not words in foveal and parafoveal vision, were identified serially or in parallel?
    The sentences used in experiments were all composed of 7 to 10 two-character words. Our manipulations are as following: Experiment 1 contained four disappearing conditions and a control. In disappearing conditions, the fixated text disappeared after readers fixate on it 50~55ms. It reappeared immediately as the readers fixated on other text, and a refixation on the current text did not result in it reappearing. The disappearing conditions were (1) current character fixated disappeared, (2) current word fixated disappeared, (3) two adjacent characters (they were not belong to the same word) disappeared together when any of them was fixated, and (4) word n and word n+1 disappeared together when word n was fixated. We also conducted the Experiment 2, in which the word n-1, word n, word n+1, and the word n+2 disappeared respectively when word n was fixated after 50~55 ms.
    The effects of the disappearing text manipulation on total reading times, number of fixations, and total runs of fixations were not reliable in the first experiment. There were significant effects of the presentation on mean saccade length, number of word refixated and skipped in the first experiment. The results of Experiment 2 have showed that readers spent more time on sentence reading when word n-1 disappearing than the normal. However, word n+1 and word n+2 disappearing conditions did not have any influence on the total reading time. The words refixated on word n disappearing condition was also less and words skipped on word n+1 disappearing condition was more than the control.
    The gaze time, the number of words refixated and skipped supported the point of “saccade generating for target selection was word–based”, thus, the result of first experiment were compatible to cognitive control models rather than vision/oculomotor models. According serial computation models, word n-1 disappearing condition do not interrupt reading, word n+1 disappeared when word n was fixated after 50~55 ms does, however the results of both experiments showed were totally different from the predictions of serial computation models, such as E-Z Reader Model. The experiments were also largely been motivated by an important difference between serial and parallel computation models. So the results of both experiments provided evidences to support parallel processing models.
    Related Articles | Metrics
    Influence of Prototype Representation on Elicitation Effect in Creative Problem Solving
    TIAN Yan,LUO Jun-Long,LI Wen-Fu,QIU Jiang,ZHANG Qing-Lin
    . 2011, 43 (06): 619-628.  
    Abstract   PDF (440KB) ( 1298 )
    Heuristic strategy is one of the usual strategies in problem solving. Zhang Qinglin proposed the heuristic theory of prototypal matters in insight problem solving, which can be regarded as creative problem solving under experimental conditions. According to this theory, “prototypal events activation” and “key heuristic information activation” are two parts of the process of prototype elicitation. The present study was designed to explore the influence of the two conditions of prototype learning on elicitation effect in creative problem solving. Namely the influence of the mark of prototype and the illustration of prototype on “prototypal events activation” and “key heuristic information activation” in creative problem solving. We adopted the activation rate of prototypal events and the accuracy of problem solving as two dependent variable index in order to depart the process of“prototypal events activation”from the process of “key heuristic information activation”.
    This study selected thirty-two scientific inventions-creations problems as the material and adopted the “8 vs 8” learning and testing experimental paradigm. First, the participants were asked to study eight prototypes successively; second, the participants were required to answer eight qestions presented randomly through oral report. And the eight qestions should be answered according to the eight prototypes supplied earlier. In the experiment, a 2 (with marks vs. without marks) ×2 (with illustrations vs. without illustrations) experimental design was adopted. Marks were the bold key phrases marked out with green in the literal representations of prototypes and they can most unlimitedly embody the principle and method of science. Illustrations were the pictorial representations of prototypes which were presented together with the literal representations of prototypes. The instrument for the statistical analyses used in the experiment is SPSS 16. 0.
    The results showed that : 1)the activation rate of prototypal events was effected evidently by the presence of marks on the level without illustrations (F(1,35)=11.87, p<0.01), and it was higher with the presence of marks, the activation rate of prototypal events was not effected evidently by the presence of marks on the level with illustrations(F(1,35)=0.03, p=0.865); the activation rate of prototypal events was effected evidently by the presence of illustrations(F(1,35)=7.48, p<0.01),and it was higher with the presence of illustrations. 2)the accuracy of problem solving was effected evidently by the presence of marks(F(1,35)=17.71, p<0.001), and it was higher with the presence of marks; the accuracy of problem solving was effected evidently by the presence of illustrations(F(1,35)=31.84, p<0.001), and it was higher with the presence of illustrations. 3)the interaction effects of marks and illustrations were significant on the activation rate of prototypal events(F(1,35)=8.21, p<0.01) and the accuracy of problem solving(F(1,35)=13.15, p<0.001).This result showed that the thinking process of prototype elicitation was flexible to a certain degree.
    Based on these results above, we can work out corresponding teaching programs in teaching activities to improve students’ creativeness for now or for the future. What’s more, in this study, we first adopted the activation rate of prototypal events and the accuracy of problem solving as two dependent variable index to depart the process of “prototypal events activation”from the process of “key heuristic information activation”. Additionally, we gained a different image representation of prototype through the presence of the prototypal literal representation together with the illustration of the prototype. This will be very helpful to the study of phased elicitation effect and image representation of prototype in the future.
    Related Articles | Metrics
    Developmental Trajectories and Gender Differences of Aggression during Middle and Late Childhood
    CHEN Liang,ZHANG Wen-Xin,JI Lin-Qin,CHEN Guang-Hui,WEI Xing,CHANG Shu-Min
    . 2011, 43 (06): 629-638.  
    Abstract   PDF (357KB) ( 1998 )
    The development and control of childhood aggression have been among topics that received the most attentions from researchers for more than half a century. Especially in recent years, with violence being recognized as a primary and increasingly important public health priority, research in this area has been further promoted and the amount of research has kept rapidly growing.
    The last decade has witnessed a transformation of research interest from the average development to the individual difference in the development of aggression, and the extant research indicated that although during middle and late childhood, physical aggression decreases for most children and many refrain entirely, some continue to engage in high level of aggression. Recent person-centered longitudinal studies in western countries demonstrated that children followed different development trajectories in development of aggression and other forms of externalizing problem behaviors. However, empirical evidence on the developmental trajectories of children’s aggression in non-western cultures has been rare.
    In the view that systematic information regarding the development of aggression of Chinese children has been highly limited, the present study explored the development and gender differences of aggression during middle and late childhood among Chinese urban children by using both variable-centered methods and person-centered methods. Three specific questions were addressed: (1) the general developmental pattern, (2) the typical developmental trajectories that different children would follow, and (3) the possible gender differences in the above aspects.
    Approximately 2000 children from 36 classes of 11 primary schools in Jinan City, Shandong Province were followed from grade 3 (average age 9.26 ± 0.29 years) till grade 6. Children’s mothers reported their aggression on five items from the Aggression subscale of CBCL annually. 1618 students have complete records on at least 3 waves of data collection, among which 1253 students have complete records on all waves of data collection.
    The main findings of the study were as follows:
    (1) The average level of aggression decreased with age. (2) Three trajectories were identified using semi-parametric modeling, respectively: no aggression trajectory (68.7%), low-decreasing trajectory (26.8%), and persistently high aggression trajectory (4.5%). Both the low-decreasing and persistently high aggression trajectories differed from the no aggression trajectory on measures of peer acceptance and peer rejection at age 12 indicating that the two aggressive trajectories were associated with interpersonal maladjustment. (3) Both multinomial Logit modeling and Chi-square analysis indicated, compared with girls, boys were more likely to be classified into trajectories of low-decreasing trajectory and persistently high aggression trajectory. However, this did not mean girls did not experience high risk of aggression, in fact 2% girls were identified to be persistently highly aggressive children. The implications of these findings for aggression and violence intervention and directions for future research were discussed.
    Related Articles | Metrics
    A New Sex-Role Inventory (CSRI-50) Indicates Changes of Sex Role among Chinese College Students
    LIU Dian-Zhi,HUANG Hui-Xin,JIA Feng-Qin,GONG Qian,HUANG Qi1, LI Xia
    . 2011, 43 (06): 639-649.  
    Abstract   PDF (438KB) ( 2777 )
    As a part of acquired social gender, sex role not only plays an important role in the development of individuals, but also influences the evolution of national characteristics. Since the sex-role ideology among Chinese college students is undergoing great changes along with the developing society, existing sex role scales may not be suitable for tapping such changes any longer. In the present study we developed a new sex role inventory for Chinese college students and examined the changes of sex-role ideology among Chinese college students.
    Following standard procedures, a scale was developed by selecting 90 sex role- indicated words (35 positive words on masculinity, 35 positive words on femininity, and 20 neutral words) from 1700 relevant words, which was then tested on 351 college students (179 males, 172 females). After item discrimination analysis and exploratory factor analysis on the initial results, 50 items in the scale were retained to constitute three subscales. The Masculinity Subscale includes 16 items and four factors: leadership, masculinity, rationality, and generosity. The Femininity Subscale includes 16 items and three factors: empathy, femininity, thrift and careful. The Neutral Subscale includes 18 items. Further, a sample of 5008 college students (2280 males, 2728 females) was selected for reexamination of the 50-item scale. 5% random subsample (106 males, 127 females) from the 5008 dataset was analyzed with Confirmatory factor analysis (CFA). The results successfully replicated the factor structure identified above and the fit indices indicated a satisfactory goodness-of-fit. The Cronbach’s alpha coefficients of the Masculinity and the Femininity subscales were 0.89 and 0.84, respectively. The retest reliability were 0.82 (M) and 0.80 (F). Demonstrative validity and expert validity were also satisfactory. The vocabulary in this new 50-item sex role inventory for Chinese college students (CSRI-50) differs substantially from other existing sex role scales: Chinese man is characteristic of the trait of rationality; Chinese women is characteristic of the trait of expression, as well as some traditional Chinese traits such as thrift, industrious and careful. Within the subsample, the proportion of Androgynous, Undifferentiated, and traditional sex role (including both of the Masculinity and Femininity) are about one third respectively. Comparing to previous studies, males showed weaker Masculinity traits, and females showed weaker Femininity traits, while the Androgynous and the Undifferentiated increased proportionally. However, the proportion of Masculinity among males was higher than that among females, and the proportion of Femininity among females was higher than that in males, indicating the continuing influence of the traditional culture.
    Related Articles | Metrics
    Self-Supporting Personality and Psychosomatic Symptoms: The Testing of Trait-Stress-Symptom Congruence and Mediating Model
    XIA Ling-Xiang
    . 2011, 43 (06): 650-660.  
    Abstract   PDF (385KB) ( 1603 )
    Self-supporting personality is viewed as an excellent personality factor in traditional Chinese culture and hypothesized as a health personality. In order to explore the role of self-supporting personality resisting psychosomatic symptoms, the trait-stress-symptom congruence and mediation model was developed, which included stress-mediator model hypothesis, trait–stress congruent hypothesis, and trait-symptoms congruent hypothesis.
    To test this model, the Self-Supporting Personality Scale of Adolescent Students, Life-Events Scale of Adolescence and Symptom Check List90 (SCL-90) were administered to 674 valid subjects.
    The results indicate: (1) The moderate model was not supported, and some self-supporting personality traits could predict psychosomatic symptoms through the mediating effect of stress. (2) The negative predicting effect of interpersonal self-supporting on psychosomatic symptoms was greater than personal self-supporting. Interpersonal self-supporting could negatively predict interpersonal and personal symptoms independently. Personal self-supporting could not independently predict interpersonal symptoms. (3) The negative predicting effect of personal self-supporting on stress was greater than interpersonal self-supporting. Personal self-supporting could negatively predict interpersonal and personal stress independently. Interpersonal self-supporting could not independently predict personal stress. (4) The modified trait-stress-symptom congruence and mediation model of self-supporting personality could be supported.
    In sum, it can be concluded that interpersonal self-supporting negatively predicts psychosomatic symptoms through meditation of interpersonal stress. Personal self-supporting negatively predicts personal psychosomatic symptoms through mediation of stress.
    Related Articles | Metrics
    The Outcome Evaluation in the Altruistic Punishment: An ERP Study
    WU Yan,LUO Yue-Jia
    . 2011, 43 (06): 661-673.  
    Abstract   PDF (553KB) ( 2085 )
    Altruistic punishment means that individuals punish, although the punishment is costly for them and yields no material gain. Recent research found that penalizing rule-breakers activated a brain region called the dorsal striatum, which is involved in experiencing pleasure or satisfaction. And researchers suggested the outcome of altruistic punishment should be equivalent to the money rewards. At this point the researchers had discovered a correlation — that the pleasure-related brain activity occurred along with inflicting the punishment — but a deeper question remained: Did one experience cause the other? Further experiments indicated that inflicting the punishment didn’t cause the players to feel satisfaction. Instead, as they decided to impose the penalty, the players were anticipating feeling satisfied. Besides, these studies concerned the decision making, rather than the feedback outcome itself. Therefore, it is not clear how humans evaluate the outcome of altruistic punishment?
    Study 1 using single trust game to ask the subjects to evaluate their emotions between the “the outcome before the punishment” and “the outcome after the punishment”, “ the outcome of cooperation” and “ the outcome of the altruistic punishment”, the “the outcome of the altruistic punishment” and “the outcome of the non-punishment” have found positive emotions increased and negative emotions weakened after the altruistic punishment compared to the outcome before the punishment. However, the pleasant feelings from the outcome of the altruistic punishment were still far below that from the outcome of the cooperation and higher than that of "the outcome of the non-punishment". And both outcome of altruistic punishment and non-punishment were biased negative emotions, and non-punishment lower ratings. The degree of negative emotions caused by “non-punishment” was greater than “punishment”.
    Study 2 using event related brain potentials studied the evaluative processes in the brain when subjects perceiving the outcome of “punish-other-lose” and “not punish-other-win”, “punish-self-lose” and “not punish-self-not lose” in the repeated trust game, and found the four kind of outcome generated remarkable feedback related negativity which was differentially sensitive to positive and negative feedback. Therefore, the results suggested that subjects evaluate the outcome of altruistic punishment as negative feedback, and the FRN amplitudes of “not punish-other-win” and “not punish-self-not lose” were larger separately than that of “punish-other-win” and “punish-other-lose”. That is, the FRN amplitudes of “not punish” which had a greater degree of negative emotions were larger than those of “punish” which had a smaller degree of negative emotions. These results suggested that the FRN reflected effective motivation process about social outcome.
    Related Articles | Metrics
    The Influence of Self-Construals on the Imagery Advertising Strategy
    YAO Qing,CHEN Rong,ZHAO Ping
    . 2011, 43 (06): 674-683 .  
    Abstract   PDF (368KB) ( 1781 )
    The paper investigates the influence of self-construals on the imagery advertising strategy. Based on recent findings in the narrative transportation theory and the social psychology model of self-concept, the authors demonstrated a moderating effect of self-construals on the imagery effect and this moderating effect did not change with different types of imagery advertisement, i.e., whether or not the imagery advertisement described the consumption scenario. Specifically, for consumers with independent self-construal, imagery advertisement was more effective than advertisement without imagery appeal in enhancing their purchase intention, whereas for consumers with interdependent self-construal, these two strategies had no significant difference. Further, the authors suggested that for independent self-construal consumers, imagery advertisement with either strong or weak argument was effective in enhancing purchase intentions, while for interdependent self-construal consumers, strong argument was a dominant strategy. These findings provided direct evidence for the proposition that independent and interdependent self-construal consumers took different psychological processes when they evaluated imagery advertisements.
    To test the hypotheses, two experiments using student subjects were performed sequentially. As hypothesized, Study 1 tested the moderating effect of self-construals on the imagery effect by means of measuring subjects’ chronically self-construal. The study demonstrated that individuals with independent self-construal were more likely to be transported into an imagery world, therefore, advertisement using imagery appeal was more effective than advertisement without imagery appeal for consumers with independent self-construal. However, for consumers with interdependent self-construal, the two strategies had no significant difference in influencing purchase intention. Importantly, by adding an imagery condition cuing imagination scenarios, we showed that the moderating role of self-construal did not change with different types of imagery advertisement, which meant that consumers could imagine with their own experience or expectations so that the effectiveness of imagery advertisement did not rely on the imagination scenarios described by marketers. Furthermore, the mediating test demonstrated that the effectiveness of imagery advertisement was mediated by transportation. Study 2 manipulated subjects’ temporary self-construals and the argument strength of the imagery advertisement. The results showed that the strong and weak argument imagery advertisement have no difference in enhancing purchase intentions of independent self-construal consumers, whereas for interdependent self-construal consumers, strong argument advertisement was more effective than weak argument. This result further proved that independent self-construal consumers were less likely to consider advertisement information than interdependent self-construal consumer, whose psychological processing might be analytical even when they were exposed to an imagery advertisement.
    The findings systematically demonstrated that the effectiveness of imagery advertisement strategy is moderated by self-construals. The research provides incremental evidence for the transportation account and the persuasion theory. From the perspective of a behavioral researcher, given the imagery appeal in advertisement, consumers’ psychological processing style should be interpreted. From the perspective of a decision theorist, the results reflect a strong boundary condition that contradicts the commonly held belief that persuasion influences purchase intentions through information accessibility. Finally, the findings have important managerial implications in designing imagery advertisement successfully.
    Related Articles | Metrics
    The Familiarity of Influenza A (H1N1), Perception of Vaccine Safety, Vaccination Behaviour and Their Influential Mechanism
    QIN Xin,NIU Cong,HUANG Zhen-Lei,XU Min-Ya
    . 2011, 43 (06): 684-695.  
    Abstract   PDF (438KB) ( 886 )
    Breaking out in Mexico, influenza A (H1N1) spread rapidly throughout the world and caused thousands of deaths. In China, the government took several measures to control the disease’s transmission, such as licensing vaccine production and setting up a vaccination system. Researchers have conducted a number of investigations on risk perceptions of past public health crises and natural disasters, but little systematic empirical research has been done for vaccine or vaccination behaviour. Also, up until now, there seems to be no explicit research conducted on influenza A (H1N1). In this study, we examined the public’s familiarity with influenza A (H1N1), perception of vaccine safety, vaccination behaviour and the mechanism of influence.
    We employed a questionnaire to collect data from several universities located in Beijing. Before conducting the survey, we first performed a pretest with a group of 30 people, and made adjustments based on their opinions and suggestions. The measurements in the questionnaire included: respondents’ demographics, the Big Five Scales, degree of prudence, involvement, degree of familiarity, perception of vaccine safety, vaccination behaviour and two open questions about vaccine safety and why participants had or had not been vaccinated. We delivered 660 copies of the questionnaire with 596 of them collected. Among the collected questionnaires 518 were deemed effective. In addition to descriptive summaries, structure equation models were used to test the hypotheses.
    The results of our main model demonstrated that the degree of prudence had a significant positive effect on the involvement in news and in communication about influenza A (H1N1). However, these two kinds of involvement had different effects on the familiarity with influenza A (H1N1). While the relation between involvement in news and familiarity with influenza A (H1N1) was positive, the involvement in communication showed no significant effect on the latter. Also, both the involvement in news and familiarity with influenza A (H1N1) had a significant positive effect on the perception of vaccine safety; that is, the positive effect of involvement in news was partially mediated by the familiarity with influenza A (H1N1). When people had a high degree of perception of vaccine safety, they were more inclined to get vaccinated. In addition, we evaluated the influences of demographic characteristics and personality on the main model. Of all these characteristics, we found that the agreeableness dimension of the Big Five personality model had a significant positive effect on the perception of vaccine safety, while age and gender played no significant role.
    This study is intended to help understand the mechanisms of people’s social and psychological behaviour in relationship to public health affairs and provide helpful suggestions for public protective measures the government can take in similar circumstances.
    Related Articles | Metrics
    Comparisons of Estimation Methods for Multilevel Random Mediation Effect Model
    LIU Hong-Yun,ZHANG Yue,LUO Fang,LI Mei-Juan,LI Xiao-Shan
    . 2011, 43 (06): 696-709.  
    Abstract   PDF (423KB) ( 1410 )
    The analysis of mediation effects is important in education, psychology, and other social sciences research. The approaches used in regression and path analysis for investigating such effects are widely known. These methods, however, are inappropriate if the data are clustered in nature, due to the violation of the assumption of independence of observations and biased standard errors. Therefore, a method for analyzing the mediation effects within multilevel models has been developed and proposed. Several procedures have been recommended and implemented in existing commercial software for testing of mediation effects in multilevel models. But most of these methods assumed that the effects are fixed, even for random indirect model. As a result, it is highly needed to examine the indirect effects under different conditions. There are few studies on this topic in Mainland till now.
    Following Bauer, Preacher, and Gil’s (2006) study, the purpose of the present article focused on the multilevel random mediation effect model (1-1-1) and examined various analytical procedures for random multilevel meditation analysis. The performances of these procedures under different conditions were compared using Monte Carlo simulations method. First, in order to address why multilevel random mediation model is necessary, the improvement in using the random multilevel mediation model compared to two compact models, the multilevel fixed mediation model and the single-level traditional mediation model is examined. Second, three different estimation methods, restricted maximum likelihood estimate (REML), maximum likelihood estimate (MLE), and minimum variance quadratic unbiased estimate (MIVQUE) are compared in different conditions. The results indicate that we can obtain unbiased estimation of the mediation effect, correct standard error, and proper result of hypothesis test through using the multilevel random mediation model, comparing with using the other two compact models. Moreover, the differences of multilevel fixed mediation model and single-level traditional mediation model are trivial. For the estimation random mediation effects in multilevel random mediation model, it is better to use restricted maximum likelihood estimate (REML) and maximum likelihood estimate (MLE), comparing with minimum variance quadratic unbiased estimate (MIVQUE). Only when the model has problem on converging, can one use MIVQUE instead, but researches should pay attention to the reliability of MIVQUE under different conditions.
    This paper consider the use of multilevel modeling to estimate mediation models in which there is lower level mediation, and all terms are random. It could be concluded that tests of random multilevel mediation can be problematic when more fixed effects models are used. For testing random indirect effects, different estimation methods might reach similar results. The REML method of SAS MIXED procedure is better than that of the MIVQUE method in the studied conditions. Recommendations are provided for testing multilevel mediation.
    Related Articles | Metrics
    Developing On-line Calibration Methods for Cognitive Diagnostic Computerized Adaptive Testing
    CHEN Ping,XIN Tao
    . 2011, 43 (06): 710-724.  
    Abstract   PDF (416KB) ( 881 )
    Like all computerized adaptive testing (CAT) applications, some items in the item bank maybe flawed or obsolete or overexposed and they should be replaced by new items (Wainer & Mislevy, 1990), item replenishing is essential for item bank maintenance and development in cognitive diagnostic CAT (CD-CAT). In regular CAT, on-line calibration method is commonly used to calibrate the item parameters of new items. However, until now no reference is publicly available about on-line calibration for CD-CAT. Thus, this study investigated the possibility to extend some current methods used in CAT to CD-CAT situation. Three representative on-line calibration methods in regular CAT were under investigation: Method A (Stocking, 1988), marginal maximum likelihood estimate with one EM cycle (OEM) method (Wainer & Mislevy, 1990) and marginal maximum likelihood estimate with multiple EM cycles (MEM) method (Ban, Hanson, Wang, Yi, & Harris, 2001). Under certain theoretical justifications based on the Deterministic Inputs, Noisy “and” Gate (DINA) model, these methods were generalized to CD-CAT situation, denoted as CD-Method A, CD-OEM and CD-MEM, respectively.
    Two simulation studies were conducted to compare the performance of the three CD-CAT on-line calibration methods in terms of item-parameter recovery. In the first study, the new items were randomly assigned to the examinees and then were calibrated accordingly. 2000 examinees were generated assuming that each examinee has 50% probability of mastering each attribute, 360 operational items were simulated and their guessing and slipping parameters were all randomly drawn from U (0.05, 0.25). 20 new items were simulated and the Q matrix corresponding to the new items was constructed by randomly selecting 20 rows from the Q matrix corresponding to the operational items, and the item parameters of new items were also randomly drawn from U (0.05, 0.25). The Shannon Entropy method was employed to select the next item and the Maximum A Posterior method was used to update the knowledge state (KS) estimates of examinees. As to the second study, the new items were first administered to a sub-group of the examinees and then were pre-calibrated; then for the remaining examinees, the new items were selected adaptively according to their initial parameter estimates to fit the examinee’s current KS estimates; finally, the item parameters of the new items were re-calibrated sequentially. Note that all the simulation conditions in the second study remained the same as those in the first study except the new items were adaptively selected.
    The results of Study 1 indicated that CD-Method A outperformed the other two methods in that it yielded the smallest estimation errors, and the simulated CD-CAT test was able to provide relatively accurate KS estimates for the examinees. The results of Study 2 showed that the adaptive calibration design could improve the item-parameter recovery compared with the random calibration design for CD-Method A, CD-Method A and CD-OEM.
    Though the results from the two studies are very encouraging, further studies are proposed for the future investigations such as different sample sizes, different cognitive diagnostic models and different attribute hierarchical structures.
    Related Articles | Metrics
Copyright © Acta Psychologica Sinica
Support by Beijing Magtech