Loading...
ISSN 0439-755X
CN 11-1911/B
主办:中国心理学会
   中国科学院心理研究所
出版:科学出版社

Next Issue

    25 November 2021, Volume 53 Issue 11 Previous Issue   

    Reports of Empirical Studies
    For Selected: Toggle Thumbnails
    Reports of Empirical Studies
    Effects of endogenous spatial attention and exogenous spatial attention on multisensory integration
    TANG Xiaoyu, TONG Jiageng, YU Hong, WANG Aijun
    2021, 53 (11):  1173-1188.  doi: 10.3724/SP.J.1041.2021.01173
    Abstract ( 320 )   PDF (853KB) ( 985 )   Peer Review Comments
    Attention effect has been greatly explored in multisensory integration. Previous studies found that endogenous spatial attention enhanced multisensory integration, while exogenous spatial attention decreased multisensory integration. However, there is no research integrates two kinds of spatial attention into the multisensory integration. In present study, we used endogenous-exogenous spatial cue target paradigm to investigate the effect of endogenous spatial attention and exogenous spatial attention on multisensory integration.
    The present study consisted of 2 experiments. In these experiments, we mainly manipulated the endogenous cue validities (including cued, uncued), exogenous cue validities (including cued, uncued) and target modalities (including visual, auditory, and audiovisual modalities). Thirty-six students in Liaoning Normal University were recruited in Exp. 1. The visual (V) target was a white pentagonal asterisk block (2°×2°). The auditory (A) target was a 1 600 Hz sinusoidal tone presented by speakers. The audiovisual (AV) target was composed by the simultaneous presentation of both the visual and the auditory stimuli. At the beginning of each trial, the fixation stimulus was presented for 600~800 ms in the center of the monitor. Following the fixation stimulus, the endogenous cue was presented for 200 ms, which could predict (80%) the location of the target. Then a visual white square served as a exogenous cue was presented for 200 ms at the left or right location randomly. Before the target that appeared for 100 ms, the inter stimulus interval (ISI) lasted for 150 ms. The target (A, V, or AV) randomly appeared for 100 ms in the left or right locations. At last, the fixation stimuli appeared for 1000 ms to wait for the correspondence responses to targets. During the experiment, the participants were asked to locate targets by pressing buttons (F/J) as quickly and accurately as possible. Thirty-six college students were recruited in Exp. 2. In Exp. 2, visual stimuli included checkerboard squares (3.2°×3.2°) with gray (RGB: 147, 149, 152) in the middle and black (RGB: 0, 0, 0) in the middle; auditory stimuli included high-frequency sinusoidaltone (1400)and low-frequency sinusoidal tone (714). The visual target was one of the above two visual stimuli, and the auditory target was one of the above two auditory stimuli. The pairings of two single channel stimuli constituted audiovisual target. The Exp. 2 consisted of four audiovisual targets. Each set of audiovisual targets were balanced among the subjects.
    The results showed that the responses to AV targets were faster than V or A targets, indicating the appearance of the bimodal advancement effect in both experiments. In addition, we found exogenous spatial attention’s race model (probability difference) showed significant smaller at cued compared to uncued condition in both experiments, while the endogenous spatial attention reduced the extent to which the exogenous spatial attention weakened the effect of multisensory integration in Exp. 2, but not in Exp. 1. From the results of the relative amount of multisensory response enhancement (rMRE), In Exp. 1, the effects of endogenous spatial attention and exogenous spatial attention on multisensory integration were marginal significant. In Exp. 2, the two kinds of spatial attention had significant interaction with multisensory integration.
    In summary, exogenous spatial attention decreased multisensory integration in both experiments. Endogenous spatial attention is involved in multisensory integration and has an impact on exogenous spatial attention in Exp. 2. The effects of endogenous spatial attention and exogenous spatial attention on multisensory integration in an interactive manner.
    References | Related Articles | Metrics
    Same-category advantage on the capacity of visual working memory
    SUN Yanliang, SONG Jiaru, XIN Xiaowen, DING Xiaowei, LI Shouxin
    2021, 53 (11):  1189-1202.  doi: 10.3724/SP.J.1041.2021.01189
    Abstract ( 301 )   PDF (707KB) ( 795 )   Peer Review Comments
    Visual working memory (VWM) is a limited-capacity cognitive system that is responsible for temporarily storing up to three or four items, maintaining their availability for current cognitive processing. Although VWM capacity is limited, the limit is not fixed. Factors such as the complexity, statistical regularity, real-world spatial regularity, and perceptual grouping principles of memory objects can modulate this capacity. However, the potential influence of higher-order conceptual regularities, such as categorical relationships among memory objects, remains an open and controversial issue. The effect of object category on VWM capacity, if any, has two possibilities: a mixed-category advantage and a same-category advantage. The first is consistent with the neural resource theory, by which the ability to simultaneously process multiple items is limited by the extent to which those items are represented by nonoverlapping neural representations. The same-category advantage is consistent with a prediction yielded by an ideal-observer analysis of VWM, based on the rate-distortion theory. Although the mixed-category advantage is predicted by the neural resource theory, almost all current supporting evidence solely involves faces. On the other hand, although the same-category advantage is consistent with the ideal-observer prediction, there is still lack of direct evidence to support the generalization of this prediction from low-level to high-level features. Hence, in the present study, we used behavioral and electrophysiological methods to explore this issue.
    Here, we report two behavioral experiments and one event-related-potentials (ERPs) experiment that assess whether category knowledge affects VWM capacity. The experiments were carried out on 60 undergraduate students.
    A 2 (memory load: two or four) × 2 (category: same or different) × 2 (posture: high similar or low similar) within-subject design was used in Experiment 1. The results showed that category knowledge modulates the capacity of the VWM and leads to a same-category advantage.
    In Experiment 2, we changed the presentation of MIs from simultaneous to successive and replicated the findings from Experiment 1, demonstrating that category knowledge leads to larger memory capacity in the same-category rather than in the different-category condition, even if the MIs are sequentially presented.
    In Experiment 3, in addition to Cowan’s K, the contralateral delay activity (CDA, an electrophysiological index of VWM capacity) was measured to further explore the processing mechanism underlying the same-category advantage. The CDA results show for the first time that same-category objects can induce a smaller contralateral delay activity (an index of VWM capacity) than different-category objects. The CDA results combined with behavioral results indicated that category knowledge can help compact the representations of same-category objects and therefore enlarge the total information capacity of VWM.
    In conclusion, our data clearly demonstrate an advantageous same-category effect on the capacity of VWM, which indicates that categorical relationships among objects play an important role in expanding the capacity of VWM by enabling the grouping of same-category objects. This suggests that VWM capacity is not fixed but can be flexible depending on the type of information to be remembered. Moreover, our data also suggest that the ideal-observer prediction can be extended from low-level to high-level features.
    References | Related Articles | Metrics
    The influence of dispositional mindfulness on the academic affective forecasting biases of middle school students
    SUN Lin, DUAN Tao, LIU Wei, CHEN Ning
    2021, 53 (11):  1203-1214.  doi: 10.3724/SP.J.1041.2021.01203
    Abstract ( 438 )   PDF (623KB) ( 1325 )   Peer Review Comments
    Human beings have affective forecasting biases. Dispositional mindfulness may weaken the narrow focus of attention by promoting the expansion of the scope of attention and the association of emotional experience, to better simulate future emotions. For adolescents, schoolwork is an important developmental task as well as an arena for their emotion generation. In the present study, we propose the following hypotheses: Middle school students have academic affective forecasting biases (H1); dispositional mindfulness affects the forecasting bias (H2); and dispositional mindfulness reduces affective forecasting biases by weakening attentional focus (H3). To test these hypotheses, three experiments including field and laboratory situational experiments were conducted.
    In the first experiment, 267 middle school students completed the Five Factor Mindfulness Questionnaire (FFMQ), and were invited to predict their feelings on achieving or not achieving their goals in the forthcoming midterm exam. After the midterm exam, they were asked to report their actual emotional experience at the moment they were informed of their final scores. In the second experiment, 70 middle school students were divided into two groups—high/low dispositional mindfulness groups—and invited to predict their emotions if they were to succeed or fail in the Word Combination Test (WCT). The test involved randomly giving participants positive or negative feedback. After the test, the participants were asked to report their real emotional experience of the success or failure. In the third experiment, 136 middle school students participated in a field experiment of achievement score feedback. Participants completed the FFMQ and attention focus questionnaires first, and then predicted their emotions regarding the result of the WCT. After completing the WCT, the participants were asked to answer the questionnaire about their emotional experiences.
    In sum, the results of the three experiments congruently prove the existence of academic affective forecasting biases in adolescents. In addition, the influence of dispositional mindfulness on affective forecasting biases was as follows: whether it was a positive or negative situation, adolescents with higher dispositional mindfulness had relatively smaller affective forecasting biases, while adolescents with lower dispositional mindfulness had relatively larger forecasting biases, but the tendency of overestimating and underestimating the emotional experience was not significant. To confirm this, we conducted a meta-analysis combining the three experimental data. The results showed that the mean effect size of dispositional mindfulness in positive situations (achievement of test goals, success feedback) was d = 0.56, Z = 4.31, p < 0.001; and the mean effect size of dispositional mindfulness in negative situations (unfulfilled test goal, failure feedback) was d = 0.42, Z = 3.47, p < 0.001. Further, dispositional mindfulness reduces the cross-context robustness of academic affective forecasting biases. In addition, in the third experiment, we found that attentional focus played a mediating role in the effect of dispositional mindfulness on affective forecasting biases.
    This study has two theoretical implications. First, it confirms the existence of affective forecasting biases in middle school students regarding their academics. Second, it reveals the effect of dispositional mindfulness on reducing biases of affective forecasting and its possible mechanism, which could be the amplification-bonding mechanism of mindfulness. In terms of practical implications, the real examination and simulated achievement feedback situations presented in this study are a reflection of real school educational activities and, therefore, have stronger ecological validity.
    References | Related Articles | Metrics
    Effects of family affective involvement on aging self-stereotypes: An analysis based on latent growth model
    XU Ran, ZHANG Baoshan, LIN Yao
    2021, 53 (11):  1215-1227.  doi: 10.3724/SP.J.1041.2021.01215
    Abstract ( 337 )   PDF (659KB) ( 1007 )   Peer Review Comments
    With the aging situation in China accelerating continuously, learning how to maintain physical and mental health in old age and promote successful aging has become the focus of researchers. As the study on this topic progresses, more and more researchers are realizing that aging is not only a biological decline but also the result of an altered psycho-social structure in the form of aging self-stereotypes. As a prerequisite central process by which aging stereotypes influence individual psychological behaviors and physiological constructions, aging self-stereotypes increase the likelihood that individuals will activate stereotypes of the group to which they belong or identify and engage in stereotype-related behaviors, thereby directly influencing their expectations, goals, and behaviors regarding self-development. Previous research has shown that the “possible future selves” facilitate processes related to the stereotypical self-conceptualization of old age. The “possible future selves” often derive from various areas of the older person’s experiences. As the family gradually becomes the most important place for older adults to socialize, the experiences of emotional support and interaction from family members are likely to influence the expectations of older adults on their “possible future selves” and the internalization of stereotypes as they grow older. However, few studies have focused systematically on the effects of older adults’ affective involvement from family members on aging self-stereotypes. There is also a lack of exploration regarding the dynamic trajectory of the relationship between family affective involvement and aging self-stereotypes in a longitudinal framework. Thus, the present study sought to examine the developmental trajectories of family affective involvement and aging self-stereotypes in older adults and the causal relationship between the two.
    In this study, a sample of 257 older adults aged over 55 was followed up for one year. The participants were asked to anonymously complete questionnaires. They were assured by the researchers that the study would be conducted purely for research purposes and that participation was voluntary. Latent growth modeling and cross-lagged analysis were used to examine older adults’ initial level of and growth in family affective involvement and aging self-stereotypes, respectively, as well as predict family affective involvement on aging self-stereotypes.
    Latent growth modeling showed that older adults’ perceptions of affective involvement from family members decreased linearly during the follow-up period while aging self-stereotypes increased linearly. In addition, the initial level of family affective involvement significantly predicted the initial level of and growth in aging self-stereotypes among older adults. Specifically, the lower the initial level of perceived affective involvement of family members, the higher the initial level of aging self-stereotypes in old age. Older adults’ lower initial levels of affective involvement from family members also predicted more rapid growth in aging self-stereotyping. Finally, the rate of decrease in family affective involvement predicted the increases of aging self-stereotypes over time, that is, a steeper decline in the family affective involvement of older adults predicted a steeper increase of aging self-stereotypes. Similarly, cross-lagged analysis showed that family affective involvement negatively predicted aging self-stereotypes after six months.
    Our findings suggest that developmental trajectories of aging self-stereotypes in older adults are influenced by changes in family affective involvement. This study expands the existing research on family affective involvement and aging self-stereotypes. A better understanding of the causal effects of older adults’ affective involvement from family members on aging self-stereotypes can also help ameliorate intervention programs designed to reduce the internalization of aging stereotypes and improve negative aging self-stereotypes.
    References | Related Articles | Metrics
    Relationship between narcissism and aggression: A meta-analysis
    ZHANG Lihua, ZHU He
    2021, 53 (11):  1228-1243.  doi: 10.3724/SP.J.1041.2021.01228
    Abstract ( 523 )   PDF (508KB) ( 1567 )   Peer Review Comments
    Aggression and violence are prevalent public health problems, tremendously harming individuals, families and society. Supposedly, low self-esteem is an important cause of aggression. However, some researchers have suggested that aggression may be attributable to threatened egoism, that is, the inflated and narcissistic view of self that is threatened, rather than low self-esteem itself. Numerous studies have explored the relationship between narcissism and aggression. However, these results appear somewhat inconsistent in different studies. Therefore, this meta-analysis was conducted to explore the strength and moderators of the relationship between narcissism and aggression.
    We included Chinese and English literature from 1965 to 2021. A total of 177 independent effect sizes (121 studies, 73687 participants) were found within the criteria of the meta-analysis. On the basis of the characteristics of studies, we selected the random-effects model. After coding the data, independent effect sizes were analyzed using Comprehensive Meta-Analysis Version 3.3 program.
    The results of the funnel plot and Egger test showed no publication bias. Results showed a significant positive correlation (r = 0.27, 95% CI [0.25, 0.29]) between narcissism and aggression. Additionally, the moderation analyses revealed that the strength of the relationship was moderated by gender and report modality of narcissism, but not by report modality of aggression and culture. Meanwhile, different types of narcissism related differently to aggression, in that covert narcissism was more positively correlated with aggression compared with overt narcissism, and maladaptive narcissism was more positively correlated with aggression compared with adaptive narcissism.
    Based on the meta-analysis, narcissism and aggression were closely related. The mechanisms of aggression must be identified to develop effective prevention and intervention strategies to alleviate the public health problems caused by aggression. Future research could: (1) The present study found that report modality of narcissism plays a moderating role in the relationship between narcissism and aggression. Therefore, to gain insights into the reporters’ bias and obtain accurate and complete information regarding narcissism, the data of multiple reporters can be employed. (2) Overt narcissism and covert narcissism are distinct structures, and the existing studies on the relationship between narcissism and aggression have paid less attention to covert narcissism. The present study found that covert narcissism is more likely to be a risk factor for aggression than overt narcissism. Therefore, future research could strengthen the exploration of covert narcissism.
    References | Related Articles | Metrics
    The influence of culture on wise reasoning in the context of self-friend conflict and its mechanism
    WEI Xindong, WANG Fengyan
    2021, 53 (11):  1244-1259.  doi: 10.3724/SP.J.1041.2021.01244
    Abstract ( 728 )   PDF (560KB) ( 1497 )   Peer Review Comments
    Wisdom involves certain types of pragmatic reasoning to navigate challenges in social life. Scholars presented aspects of wise reasoning that include perspective taking, consideration of change and alternatives, intellectual humility, search for compromise, and adopting an outsider’s vantage point. Researchers have found that most WEIRD (western, educated, industrialized, rich, and democratic) people can reason more wisely about friends’ social problems than their own (i.e., Solomon’s paradox). However, it is not clear whether Solomon’s paradox will exhibit different characteristics and mechanisms among Chinese people who are mainly interdependent selves. We hypothesized that 1) Americans endorsed greater wise-reasoning strategies in the friend-conflict condition than in the self-conflict condition, while among Chinese, endorsement of wise-reasoning strategies did not differ between the two conditions; 2) People with independent self exhibited greater wise reasoning in the friend-conflict condition than in the self-conflict condition, whereas people with interdependent self didn’t have significant differences between the two conditions.
    In study 1, we recruited 594 American participants from MTurk and 610 Chinese participants from a Chinese survey platform to participate in “a survey of daily life”. Participants who responded less attentiveness to the study was excluded. The final sample consisted of 282 American participants (125 females, 155 males 2 with unreported gender; M = 36.87 y, SD = 11.04; 218 Whites, 21 African Americans, 27 Asian, 13 Latino, and 3 other) and 295 Chinese participants (151 females, 144 males; M = 23.22 y, SD = 4.34). Participants were randomly assigned to either self-conflict or friend-conflict conditions. They were asked to think about a close relationship that was currently not going very well, and then responded to the wise-reasoning scale and self-construal scale. The results using the alignment and ANOVA both showed that Americans reason more wisely about friend’s interpersonal conflict than about their own, while the Chinese didn’t have significant differences between the two conditions. Moreover, the interaction effect of independent self and conflict type on wise reasoning was significantly positive in Chinese culture. Among participants high in independent self, endorsement of wise-reasoning strategies was greater in the other-conflict condition than in the self-conflict condition (β = 0.21, t(287) = 2.48, p = 0.014), whereas among participants low in independent self, endorsement of wise-reasoning strategies did not differ between the two conditions (β = -0.07, t(287) = -0.81, p = 0.419). In study 2 (n = 710) and 3 (n = 537), we created a 2 (priming independent vs. interdependent self-construal) × 2 (self-conflict vs. friend-conflict) between-subjects design online and in study 4 (n = 200), we created same design in lab. The results all showed that independent participants reason more wisely about friend’s conflict than about their own, while the differences between the conditions were not significant among interdependent participants.
    The results from the two studies indicate that Solomon’s paradox is not universal and may only exist in individuals with high independent self. This study also suggests that we should not only focus on the WEIRD samples but also pay attention to the WEIRD researchers who use these samples to make inferences about humans in general.
    References | Related Articles | Metrics
    Mechanisms of different purposes of social media usage on employees’ job performance
    LI Qiaoling, ZHAO Junzhe, QIAO Shiqi, GUO Tengfei, WANG Minghui, ZHAO Guoxiang
    2021, 53 (11):  1260-1270.  doi: 10.3724/SP.J.1041.2021.01260
    Abstract ( 942 )   PDF (465KB) ( 2109 )   Peer Review Comments
    Most previous studies have explored the effect of employees’ usage of social media from a single perspective and resulted in inconsistent findings. In addition, some studies ignored the possible interactions among different characteristics of social media usage, leading to an incomplete understanding of social media. Therefore, this study attempts to examine two characteristics of social media usage, integrate the two paths of resource buffer and resource gain, and explore the influence mechanism of social media usage on employees’ job performance.
    Specifically, based on the job demands-resources theory, social media has an impact on an organization as a job resource. From the perspective of resource gain path, employees can obtain information about their work through social media to improve their working ability; In addition, the increase in interactive communication through social media deepens mutual understanding and enhances trust. From the perspective of resource buffer path, social media provides employees with a wealth of professional and personal resources. These resources help employees cope with the demands of the workplace, such as distractions and conflicts. Therefore, from the two paths, we assume that work state and interpersonal relationships can act as parallel multiple mediations between social media usage and job performance, respectively.
    Using a “leader-subordinate” pairing survey, 519 questionnaires were distributed in enterprises in Shanghai and Henan, and 369 valid questionnaires were collected. We found that work-based and social-based social media usage both affect job performance through work engagement. Additionally, the social-based social media usage also affects job performance through relationship conflict mediation. Interaction effect tests further suggest that work-based and social-based social media usage are mutual substitutions for influencing work performance through work engagement.
    This study has three main contributions to the literature. First, this study investigates the influencing mechanisms of different types of social media usage. Second, the study draws on the job demands-resources theory to explain the impact of social media usage on job performance, enhancing our understanding of reasons we use social media. Finally, the study highlights the substitution between different types of social media usage, offering clarification on the internal mechanism between social media usage and job performance.
    References | Related Articles | Metrics
    The polarization effect of project presentation in ranking list on project evaluation
    DUAN Shen, MENG Lu
    2021, 53 (11):  1271-1285.  doi: 10.3724/SP.J.1041.2021.01271
    Abstract ( 115 )   PDF (629KB) ( 209 )   Peer Review Comments
    As a means of communication to help consumers quickly identify the advantages and disadvantages of brands or products, the ranking list has been paid more and more attention by merchants and consumers. However, most of the existing researches on the ranking effect focuses on comparing the influence mechanism of different ranking list items or positions on consumers, ignoring the influence of different display modes of the same list on consumer evaluation. For this purpose, based on the spatial metaphor theory, the paper studied different display modes (vertical or horizontal) of the same ranking list to explore their impacts on the difference in consumer evaluation of items ranking list.
    This study intends to explore the above relationship and its internal mechanism through five experiments. Firstly, two online experiments (Experiment 1a [N = 200] and Experiment 1b [N = 332], scenario of mobile phone brand evaluation) were conducted to verify that the evaluation polarization effect existed in both the positive ranking list and the negative ranking list. Secondly, Experiment 2 (N = 228) was a laboratory experiment to expand the partial ranking list to the whole item ranking list form (gift evaluation scenario). The results show that in the full-ranking list form, consumers’ evaluation difference of items in the vertical ranking list is significantly higher than that in the horizontal ranking list, with the increase of the number of items in the ranking list, evaluation difference becomes more obvious.
    Experiment 3 (N = 187) adopted common school ranking scenarios to verify the mediating role of item perception difference in the above relationship, the higher evaluation polarization effect of the vertical ranking list than the horizontal ranking list is due to Consumers have greater perceived item differences for items in vertical ranking lists. Finally, Experiment 4 (N = 279) applied the form of brand ranking list to verify the main effect, meanwhile, the moderating effect of attribute evaluability was also examined. When the evaluability of attributes is low, the evaluation polarization effect of different lists will appear. On the contrary, when the item attribute of the ranking list is highly evaluable, this polarization effect of evaluation disappeared.
    This study expands the relevant theories of ranking effect and spatial metaphor, and provides practical guidance for merchants to flexibly use ranking lists to effectively communicate with consumers. In particular, the contributions of the study include the following: (1) we explore the influence of the change of ranking item spatial position on consumer item evaluation, and thus expand the theoretical framework of ranking effect research; (2) we effectively combine spatial and numerical elements in the context of the ranking list, further expend the theory of numerical cognitive bias under the same dimension of number representation forms; (3) we expand the theoretical application scenarios of “spatial metaphor”. In this study, the two dimensions of spatial metaphor theory (location metaphor + distance metaphor) were employed to systematically illustrate the internal mechanism of the influence of spatial representation of ranking list on evaluating polarization effect. Our study effectively brings the two sub-dimensions of spatial metaphor theory into the overall consideration of the same theoretical framework; (4) we expand the precursor psychological factors of evaluation polarization from a cognitive perspective. The emergence of the study increases and expands the expression forms and application scenarios of evaluating polarization effect, deeply exploring the psychological cognitive process of evaluating polarization effect in theoretical scenarios.
    References | Related Articles | Metrics
    A high-efficiency and new online calibration method in CD-CAT based on information gain of entropy and EM algorithm
    TAN Qingrong, WANG Daxun, LUO Fen, CAI Yan, TU Dongbo
    2021, 53 (11):  1286-1298.  doi: 10.3724/SP.J.1041.2021.01286
    Abstract ( 114 )   PDF (662KB) ( 190 )   Peer Review Comments
    Cognitive diagnostic computerized adaptive testing (CD-CAT) includes the advantages of both cognitive diagnosis (CD) and computerized adaptive testing (CAT), which can offer detailed diagnosis feedback for each examinee by applying fewer test items and time. It has been a promising field. An item bank is a prerequisite for the implementation of CD-CAT. However, its maintenance is a very challenging task. One of the effective ways to maintain the item bank is online calibration. Till now, there are only a few online calibration methods in the CD-CAT context that can calibrate Q-matrix and item parameters simultaneously. Moreover, the computational efficiency of these methods needs to be further improved. Therefore, it is crucial to find more online calibration methods that jointly calibrate the Q-matrix and item parameters.
    Inspired by the SIE (Single-Item Estimation) method proposed by Chen et al. (2015) and information gain feature selection criteria in feature selection, an information gain of entropy-based online calibration method (IGEOCM) was proposed in this study. The proposed method can jointly calibrate Q-matrix and item parameters in a sequential manner. The calibration process of the new items was described as follows: First, for the new item j, the q-vector can be calibrated by maximizing the information gain of entropy-based on the basis of the attribute patterns of examinees and the examinees’ responses to item j. Second, the item parameters of the new item j are estimated by the EM algorithm based on the posterior distribution of examinees’ attribute pattern, the examinees’ responses to item j, and the q-vector estimated in the first step. The first and second step are repeated for all other new items to obtain their estimated Q-matrix and item parameters item by item. Two simulation studies were conducted to examine whether the IGEOCM could accurately and efficiently calibrate the Q-matrix and item parameters of the new items under different calibration sample sizes (40, 80, 120, 160, and 200), different attribute pattern distributions (uniform distribution, higher-order distribution, and multivariate normal distribution), the different number of new items answered by examinee (4, 6, and 8), and different item selection algorithms (posterior-weighted Kullback-Leibler, PWKL; the modified PWKL, MPWKL; the generalized deterministic inputs, noisy and gate model discrimination index, GDI; and Shannon entropy, SHE). Furthermore, the performance of the proposed method was compared with the SIE, SIE-R-BIC, and RMSEA-N methods.
    The results indicated that (1) The IGEOCM worked well in terms of the calibration accuracy and estimation efficiency under all conditions, and outperformed the SIE, SIE-R-BIC, and RMSEA-N methods overall. (2) The accuracy of the item calibration increases as the sample size increases for all calibration methods under all conditions. (3) The SIE, SIE-R-BIC, RMSEA-N, and IGEOCM performed better under the uniform distribution and higher-order distribution than under the multivariate normal distribution. (4) The number of new items answered by the examinee had a negligible impact on the calibration accuracy and computation efficiency of the SIE, SIE-R-BIC, RMSEA-N, and IGEOCM. (5) The item selection algorithm in CD-CAT affects the Q-matrix calibration accuracy of the SIE and IGEOCM methods. Under the higher-order distribution and multivariate normal distribution, the SIE method and IGEOCM had higher Q-matrix calibration accuracy when the item selection algorithms were MPWKL and GDI.
    On the whole, although the proposed IGEOCM is competitive and outperforms the conventional method irrespective of the calibration precision or computational efficiency, the studies on the online calibration method in CD-CAT still need to be further deepened and expanded.
    References | Related Articles | Metrics