ISSN 0439-755X
CN 11-1911/B


    25 September 2015, Volume 47 Issue 9 Previous Issue    Next Issue

    For Selected: Toggle Thumbnails
    Attentional Guidance from Activated and Inhibitory States of  Working Memory Representations
    ZHANG Bao, SHAO Jiaying, HU Cenlou, Huang Sai
    2015, 47 (9):  1089-1100.  doi: 10.3724/SP.J.1041.2015.01089
    Abstract ( 1808 )  

    Attention and working memory are two of the core cognitive processes in the human’s information processing system. Working memory and visual attention are intimately related, and the contents of working memory can be referred as “activated” representations severed as ongoing cognition and action. In the meantime, working memory representations currently within the focus of attention can guide attentional selection and behavioral execution. In recent decades, a mount of studies have shown that the activated representations in working memory could top-down capture attention, even if the representations were irrelevant to the task goals, which displayed a robust working memory-driven guidance effect. However, whether the inhibitory representations in working memory can also guide attention, is still a controversial issue? Here, in 4 experiments, the authors manipulated the states of working memory representations with the directed forgetting task, and attempted to explore the effect of activated and inhibitory states of working memory representations on working memory based attentional guidance respectively in task-irrelevant and task-relevant experimental situations. The participants in present study were asked to firstly perform a directed forgetting task, then to search for a target of circle among the distractors of squares, and finally to perform a memory test. In the directed forgetting task, the remember cue in experiment 1 and 3, and the forget cue in experiment 2 and 4 were used to respectively indicate the participants to remember or forget one of two colored squares already stored in working memory. The to-be-remembered (TBR) item and to-be-forgotten (TBF) item in working memory would reappear in visual search task and might match color with one of distractors only in experiment 1 and 2 (i.e., task-irrelevant situation), and might match color with either the target or the distractor in experiment 3 and 4 (i.e., task-relevant situation). In experiment 1 and 2, the results suggested that no matter what types of cues used in the directed forgetting task, when the distractor in visual search task matched color with TBR item, the TBR-matched distractor could capture more attention and slow down the visual search, displaying the attentional guidance effect. However, when the distractor matched the TBF item, neither attentional guidance effect nor attentional inhibition effect was observed. In experiment 3 and 4, the results for the TBR items showed that the visual search was accelerated under TBR-target matching condition and slowed down under TBR-distractor matching condition both in experiment 3 and 4, suggesting that the TBR item in working memory could guide attention biased to the TBR-matched items in visual search task. The results for the TBF items showed that when TBF item matched with distractor, there was no attentional guidance effect found both in experiment 3 and 4. When TBF item matched with target, the guidance effect was still not found in experiment 3. However, an attentional inhibitory effect which was opposed to the attentional guidance effect was observed in experiment 4, suggesting that the inhibitory state of the TBF item could postpone the response to the TBF-matched target. In conclusion, these results indicated that (1) the activated working memory representation could effectively guide attention bias both to the search target and to the distractor in visual search which matched features with such representation, and these attentional guidance effect could not be eliminated or reversed by the inhibitory motivation; (2) the inhibitory working memory representation could transfer the inhibitory state to the visual search task and postpone the response to the search target which matched the features with such representation.

    Related Articles | Metrics
    Neural Processing of Recollection, Familiarity and Priming at Encoding
    YE Xiaohong, CHEN Youzhen, MENG Yingfang
    2015, 47 (9):  1101-1110.  doi: 10.3724/SP.J.1041.2015.01101
    Abstract ( 1146 )   PDF (678KB) ( 64 )  


    The distinction between neural mechanisms of explicit and implicit expressions of memory has been well studied at the retrieval stage, but less at encoding. Several studies employed a novel paradigm to measure explicit memory and priming-without-explicit memory in one test, and contrasted the neural signals of these two processes at the encoding stage via a Dm analysis. However, dissociations obtained in these studies are complicated because of the contamination from familiarity, a fast automatic process in which memory judgments can be driven by the increased fluency of reprocessing studied information. Familiarity is also a form of explicit memory but different from recollection. Growing evidence has indicated that familiarity-based recognition judgments might rely on the same process that supports implicit memory or priming. Therefore, it is necessary to concurrently acquire and compare the encoding processes yielding later recollection, familiarity and priming in a single test.
    In this study, a two-stage forced-choice recognition test was adopted within the subsequent memory (Dm) paradigm, so as to simultaneously acquire neural correlates of recollection, familiarity and priming in a single test. During the study phase, participants were instructed to judge the color of each word. There were two stages during the forced-choice recognition test phase. In the first stage, two words (one from the study list, and one new) appeared concurrently, and participants were instructed to indicate the studied word. If they could not recognize the studied word, a guess was permitted to make a choice. In the second stage, a cue to make a confidence judgment appeared directly following the recognition response, and participants indicated whether the foregoing studied-selection was based on remembering, knowing or guessing. Here, “remembering” refers to the retrieval of specific details from the study phase supporting the recognition decision, “knowing” refers to the recognition supported by a weak feeling of familiarity with few details retrieved from the study phase, and “guessing” refers to “absolutely no feeling of memory” such that the stimulus in no way felt “old”. The study ERP data were then classified into four categories as “subsequent remembered” (later retrieved with detailed information), “subsequent known” (later retrieved with a feeling of familiarity)”, “subsequent primed” (later retrieved without conscious awareness) and “subsequent forgotten” (not retrieved). Differences in subsequent memory effects (Dm effects) were measured by comparing ERP waveform associated with later memory based on recollection, familiarity or priming with that associated with later forgotten items. In addition, interference during encoding was introduced in Experiment 2 to determine whether three Dm effects were different from Experiment 1. The interference task was to judge the orientation of arrow which appeared with word at the same time.
    The results showed that, in Experiment 1, the recollection Dm effect involved a robustly sustained (onset at 400 ms) prefrontal positive-going Dm effect which was right-lateralized, and a later (onset at 800 ms) occipital negative-going Dm effect. Familiarity involved an earlier (300~400 ms) prefrontal positive-going Dm effect and a later (500~600 ms) parietal positive-going Dm effect. Priming involved a negative-going Dm effect which onset at 700 ms, mainly distributed over anterior brain sites. In Experiment 2 with an interference task during encoding, a similar priming Dm effect that was negative-going during 600~800 ms at central site, and a similar familiarity Dm effect that was positive-going during 600~800 ms at frontal site were still observed. However, there was no evidence of Dm effect associated with recollection.

    Taken the two ERP results together, we inferred that there would be a sequence of components that represented cognitive processes underlying the encoding of verbal information into episodic memory, and separately supported later remembering, knowing and priming.

    Related Articles | Metrics
    Emotional Memory Enhancement Effect in Dual-processing Recognition Retrieval
    MAO Xinrui, XU Huifang, GUO Chunyan
    2015, 47 (9):  1111-1123.  doi: 10.3724/SP.J.1041.2015.01111
    Abstract ( 1814 )   PDF (733KB) ( 121 )  

    In studies of recognition retrieval, emotional memory enhancement effect was described as better memory performance for emotional stimuli than neutral ones. Based on dual-processing theory, recognition retrieval can be divided into two different processes: familiarity and recollection. Two important event-related potential correlates, the FN400 (a negative shift in frontal regions at 300~500ms time window), and the late positive complex (LPC; a positive peak over posterior regions at 500~800ms time window) was associated with familiarity and recollection, respectively. Some researchers considered that emotional memory enhancement effect occurred in recollection but not in familiarity. However, some indirect evidences showed that emotion could enhance memory strength in familiarity-based retrieval. Our research focused on two controversies: 1) whether emotion can enhance familiarity, and 2) how arouse and valence of emotion affect memory enhancement effect. In the current experiment, we used modified “remember/know” paradigm with ERPs recorded, to investigate how emotion influences familiarity and recollection in long-term study-test duration. Subjects were instructed to learn the pictures (including neutral, negative and positive pictures). And after one week, they made “remember/know/guess/new” recognition judgments towards stimuli intermixed with learnt and new pictures. Finally, the valence and arouse of experimental pictures were evaluated by the subjects participated in the experiment. Behaviorally, for studied pictures endorsed as “know”, the memory performances of emotional pictures were better than neutral ones, and there was no significant difference between emotional valences. As for the “remember” judgments, the memory performances of negative pictures were better than positive and neutral ones. ERP results show that, for the pictures judged as “know”, the FN400 old/new effects were significant in emotion condition but not in neutral condition, suggesting that emotion arousing enhanced familiarity-based retrieval. For pictures judged as “remember”, the LPC potentials of negative pictures were more positive than positive and neutral ones, showing that emotion valence modulated recollection-based retrieval. Our findings suggest that: in long-term study-test duration, emotional pictures could result in memory enhancement effect in the familiarity-based as well as the recollection-based retrieval. The emotional memory enhancement effect was not only modulated by the arousing as well as the valence of emotion. Emotional arousing played a predominant role in enhancing memory strength for familiarity-based retrieval. Emotional valence exerts influence exclusively on recollection-based retrieval in a way that only negative pictures could make enhancement effect on recollection.

    Related Articles | Metrics
    When does Inhibition of Affect Labeling Occur: ERP Study
    YUE Pengfei, DU Wanwan, BAI Xuejun, XU Yuanli
    2015, 47 (9):  1124-1132.  doi: 10.3724/SP.J.1041.2015.01124
    Abstract ( 1014 )   PDF (799KB) ( 55 )  


    Affect labeling refers to the processing of emotional stimuli, with words describing the stimuli. Many studies using fMRI have found that affect reactions are weakened when they are being tested in an affect labeling condition. However, studies using ERP technology have found dissimilar results. Given this contradiction, this study uses ERPs technology to study the change of affect labeling at two stages of the EEG to test this issue.
    The present study used positive and negative emotional faces as experimental materials, with happy emotional face images as positive stimuli and angry emotional face images as negative stimuli. Half of these faces were male, and half were female. The research used a 2 (labeling types) × 2 (emotion types) experimental design, in which there were two levels of labeling types: affect labeling and gender labeling. And emotion types had two levels: anger and happiness. In the Experiment, the affect labeling condition referred to observing the facial emotion expressed in the description given below the face, and then selecting the appropriate word (angry or happy). Gender labeling refers to observing the face pictures, and then choosing the right words for the face's gender (Li Na or Zhang Tao). If the subject believed the face images are men, they should choose the marked word “Zhang Tao”, or “Li Na” if they believe the images are women. The dependent variables were accuracy rate, reaction time, and the subjects to perform tasks during EEG.
    The results show that: (1) before the end of affect labeling: in the negative emotion condition, the amplitudes have no differences between affect labeling and gender labeling in the late positive potential (LPP), however the amplitudes of affect labeling were higher under the positive emotion condition in the LPP; (2) after the affect labeling: the amplitudes of affect labeling in the LPP were lower, compared with gender labeling, regardless of positive emotion or negative emotion.

    The above results reveal that the inhibition of the affect labeling on emotions occurs when subjects label emotional stimuli. In the process of emotional labeling the positive emotion was enhanced, and then was weakened.

    Related Articles | Metrics
    The Assessment of Marital Attachment and Its Relationship with General Attachment among Older Adults
    WANG Dahua; YANG Xiaoyang; WANG Yan; Richard B. Miller
    2015, 47 (9):  1133-1142.  doi: 10.3724/SP.J.1041.2015.01133
    Abstract ( 2338 )  


    According to attachment theory, during interactions with their caregivers infants build their “internal working models”, which shape their perception and reflection of social information in their future life. Collins and Read (1994) propose a hierarchical model to illustrate adults’ multiple attachment models and the differences between the general attachment model and the relationship-specific attachment model. Because marital relationships play an important role in later life, an investigation of older adults’ marital attachment and its relationship with the general attachment formed during childhood would increase the understanding of life-span development of attachment. The Older Adults’ Marital Attachment Scale (OAMAS) was recently developed targeting both the specific relationship, as well as the specific age period, and the developers have demonstrated its applicability. However, more evidence is needed using different samples to prove its reliability and validity. As a result, the current study aimed, firstly, at validating the scale, and, secondly, at investigating the association between marital attachment and general attachment among older adults.
    A total of 697 older adults, over 60 years old, dwelling in communities in Beijing participated in the current study. The participants completed the Older Adults’ Marital Attachment Scale (Zhai, et al., 2010), Relationship Questionnaire (Bartholomew &Horowitz, 1991), Marital Satisfaction subscale (Olson, Fournier, & Druckman, 1983), 15-item Geriatric Depression Scale (Burke, Roccaforte, & Wengel, 1991) and the Clock Drawing Test. After 86 of them with either dementia or depression were excluded, the final valid sample size was 611 participants. The cases were divided randomly into two subsets, one for exploratory factor analysis, and the other for confirmatory factor analysis. Cluster analysis and cross-tab analysis were then conducted on the total data. The SPSS 17.0 and Mplus 7.0 were used for analysis.
    The main findings were as follows: (1) the final 15-item revised OAMAS showed a three-dimensional construct of attachment, namely anxiety, avoidance, and security; (2) the revised OAMAS exhibited acceptable reliability and good criterion-related validity; (3) the older adults’ marital attachment could be clustered into four types, including secure, preoccupied, dismissing, and fearful. The secure and the preoccupied comprised the major marital attachment style with 43.9% and 22.9%, respectively; (4) with regard to general attachment, the majority included the secure and dismissing types; and (5) about 39.9% of the total sample presented identical attachment types between marital relationship and general relationship.

    These results indicated that the three dimensioned Older Adults’ Marital Attachment Scale could be a desired measure in studies concerning marital attachment among older adults. In addition, general attachment shaped in early life might not robustly predict the specified attachment in such as marital relationship. From a developmental viewpoint, the moderate consistency between marital and general attachment styles in later life also suggested that the attachment is a contextual property, instead of a stable trait.

    Related Articles | Metrics
    Friend-reference Effect in Older Adults
    ZHOU AiBao;LIU PeiRu;ZHANG YanChi;YIN YuLong
    2015, 47 (9):  1143-1151.  doi: 10.3724/SP.J.1041.2015.01143
    Abstract ( 1400 )  

     A great deal of research adopting the self-reference effect (SRE) paradigm has investigated whether an individual’s self-construal includes intimate others, such as Mother or Friend. Most of the studies used undergraduates as subjects, but no research has investigated the friend-reference effect in older adults. So two experiments were conducted to investigate whether older adults can show friend-reference effect and to examine the feature of this effect.

    Thirty-six urban and 36 rural older adults participated in Experiment 1. A 3 (Reference Condition: Self, Friend vs. Lu Xun) × 2 (Living Area: City vs. Countryside) mixed design was used. A repeated measure ANOVA for the rate of free recall revealed that both urban and rural older adults showed a friend-reference effect and a self-reference effect, F(2, 140) = 13.46, p< 0.001, η2 = 0.16. In order to investigate the difference between the friend-reference effect and the SRE, we computed the reference effect index (the friend-reference effect index = free recall rate of friend-related items – free recall rate of other-related items; the SRE index = free recall rate of self-related items – free recall rate of other-related items). Then a 2 (Reference Effect Index: Self vs. Friend) × 2 (Living Area: City vs. Countryside) mixed design was used. A significant interaction between Reference Effect Index and Living Area was significant, F(1, 70) = 8.03, p = 0.006, η2 = 0.10. Simple effects analysis revealed that compared to rural elderly people, the friend-reference effect index was much bigger than the SRE index in urban elderly people, p = 0.042.
    Thirty-six older urban adults with low-levels of education and well-educated rural elderly adults volunteered to participate in Experiment 2. A 3 (Reference Condition: Self, Friend vs. Lu Xun) × 2 (Level of education: High vs. Low) mixed design was used. A repeated measure ANOVA on the free recall rate revealed that both older urban adults with low-levels of education and well-educated rural elderly showed the friend-reference effect and the SRE, F(2, 140) = 10.65, p< 0.001, η2 = 0.13. A 2 (Reference Effect Index: Self vs. Friend) × 2 (Level of education: High vs. Low) mixed design was used. A significant interaction between Reference Effect Index and Level of education was significant, F(1, 70) = 7.72, p = 0.007, η2 = 0.10. Simple effects analysis revealed that compared to urban elderly people with low-levels of education, the friend- reference effect index was much bigger than the SRE index in well-educated urban elderly.

    These results suggested that Friend was included in one’s self-construal for elder people. And the level of education plays an important role in the friend-reference effect of elderly people.

    Related Articles | Metrics
    The Effect of Leader Emotional Intelligence on Group Performance and #br# Employee Attitude: Mediating Effect of Justice Climate and Moderating Effect of Group Power Distance
    RONG Yan; SUI Yang; YANG Baiyin
    2015, 47 (9):  1152-1161.  doi: 10.3724/SP.J.1041.2015.01152
    Abstract ( 1631 )  

    Recently, increasing numbers of researchers have shown interests in leader emotional intelligence. Even though some of them are optimistic on the relation between leader emotional intelligence and team performance, critics still remain and mixed results are found. What’s more, the mechanisms and constrains that explain the effectiveness of leader emotional intelligence, especially at group level, are relatively unexplored. We address those research gaps by exploring the effect of leader emotional intelligence on leader-member interaction in the perspective of justice climate and group power distance. We also synthesize the performance and attitude at the group level to further examine the influence of leader emotional intelligence in working groups. Justice climate is a suitable reflection of leader-member interaction while emotional intelligence underlines leaders’ social skills. So that, exploring the effectiveness of leader emotional intelligence in the perspective of justice climate grasps the essential characteristic of emotional intelligence. On the other hand, group power distance is a contextual factor that affects the influence of leaders on group outcomes, thus it constrains the extent to which leader emotional intelligence could be effective. The sample for this study consisted of 74 working groups. We used the ability model to define emotional intelligence because it showed better psychometric properties. The questionnaires for leaders included emotional intelligence, group performance, and organizational citizenship behavior, while the questionnaires for group members included procedural justice, interactional justice, satisfaction and organizational commitment. Justice, power distance and four dependent variables were aggregated to group level. Multiple mediator model, moderated mediation and bootstrapping were used in the analysis. The result showed significantly positive relations between leader emotional intelligence and group level outcomes such as task performance, organizational citizenship behavior, satisfaction and organizational commitment. Those relations were mediated by procedural justice climate or interactional justice climate. Furthermore, group power distance moderated the relation between leader emotional intelligence and interactional justice climate: the relation is stronger in groups with higher, rather than lower, group power distance. Group power distance worked indirectly through interactional justice climate to affect group performance and attitude. The current study makes significant contributions both theoretically and practically. Theoretically, we join the debate of emotional intelligence’s effectiveness by showing promising evidences supporting the positive effects of leader emotional intelligence on both performance and attitude at the group level. We also demonstrate the mechanism and contingency in the perspective of leader-member interaction by examining the roles of justice climate and group power distance. Practically, the result of this study suggests that emotional intelligence is an important ability we should consider when selecting the leaders. Leaders can make use of their emotional intelligence to create a justice climate in the organizations. Leaders should also pay attention to group power distance because it will affect the effectiveness of their emotional intelligence.

    Related Articles | Metrics
    Mr. Right & Superman: Effect of Implicit Followership on Employee’s Behaviors
    KONG Ming, QIAN Xiaojun
    2015, 47 (9):  1162-1171.  doi: 10.3724/SP.J.1041.2015.01162
    Abstract ( 1210 )  

    Leaders’ Implicit Followership Theories (LIFTs) are defined as “leaders’ personal assumptions about the traits and behaviors that characterize followers”. It is an extension of implicit leadership theories (ILTs) supported by the social cognitive theory and cognitive information processing theory. From the “followers-centered” perspective, this paper analyzed the leadership process of “how leaders and followers perceive, decide, and act”.

    This paper mainly discussed the influence factors on employee’s behaviors from the perspective of implicit followership. The social exchange theory provides a theoretical framework for understanding how employee’s beneficial social exchange with their organizations lead to their positive attitude and behaviors. The study also tests the mediating effect of leader-member exchange (LMX) and the moderating effect of psychological empowerment. It aims to address the following three research questions: (1) Can positive LIFTs affect employee’s behaviors, especially for the In-Role Behavior (IRB) and Organizational Citizenship Behavior (OCB), and how? (2) Can “Mr. Right” be “Superman”? Why and how? (3) Can and how LMX influence the relationship between “Mr. Right” and “Superman”?
    Questionnaire surveys of both employees and their leaders in 19 large and medium-sized enterprises located in Beijing, Shenzhen, and Hangzhou were conducted. A total of 480 questionnaires were collected and among which 278 were valid (the efficiency is 57.9%). In order to reduce the common method variance, 2 questionnaire surveys with an interval of two months were conducted. After the CFA, we examined the reliability of the scale and tested the correlation between the variables. Multiple linear regression analyses were carried out based on the data from 278 employees in 19 companies.
    The conclusion is that positive LIFTs are positively correlated with employee’s IRB and OCB. Moreover, the results of HRM show that LMX mediated the relationships between positive LIFTs and IRB/OCB. And the moderating effect of psychological empowerment is significant. The results of this study not only extends the application of implicit theory and cognitive theory in the studies on leadership, but also enriched the content of social exchange theory.
    Related Articles | Metrics
    A Meta-Analysis of the Relationship between Team Demographic Diversity and Team Performance
    WEI Xuhua; LIU Yongmei; CHEN Sixuan
    2015, 47 (9):  1172-1187.  doi: 10.3724/SP.J.1041.2015.01172
    Abstract ( 997 )  

    Over the past decades, team demographic diversity has become a topic of considerable interest to industrial and organizational psychology scholars and organizational managers. However, there is little consistent evidence regarding the relations between team demographic diversity and team performance. There are at least two potential reasons to explain these inconsistencies. First, there are different forms of team demographic diversity and the specific type of diversity should have different effects on team performance. For example, team demographic diversity can be categorized as separation, variety and disparity based on the statistical distribution of team members' characteristics. Second, past researchers suggest considering contextual issues in team demographic diversity research. Rather than test the direct relationship between team demographic diversity and team performance, they have pointed out that contextual factors (e.g., cultural context) should play an important moderating role in the relationship between team demographic diversity and team performance.

    In order to explain the inconsistencies in past research examining the link between team demographic diversity and team performance, we conducted a meta-analysis to examine the effects of different types of team demographic diversity on team performance. Our meta-analysis was based on 345 effect sizes from 137 Eastern and Western empirical studies with 79,639 teams. Each author independently coded the data and resolved discrepancies through discussion. In our coding system, we coded diversity as separation, variety, or disparity based on the measures of diversity used in each empirical paper. Further, we collected contextual data to examine the potential moderating effects of contextual factors, such as performance types, cultural context and team types.
    Results of main effects showed that team demographic variety had significantly positive effects on team performance, whereas team demographic separation and disparity were unrelated to team performance. Further, moderation analyses showed that the relations between team separation, variety, disparity and team performance were moderated by performance types, cultural context and team types. Specifically, considering performance type as a moderator, variety and disparity were more positively correlated with innovation performance compared to general task performance. With respect to cultural context, team demographic variety in eastern countries was more positively correlated with team performance compared to variety in western countries, whereas team demographic disparity in western countries was more negatively correlated with team performance compared to disparity in eastern countries. Regarding team types, team demographic variety was more positively correlated with performance in top management teams (TMTs) and research and development (R&D) teams compared to general work teams.

    Our results showed that different demographic diversity had distinct effects on team performance, depending on the specific diversity type and context (e.g., performance types, culture and team types). However, many researchers rarely distinguish between different types of demographic diversity. Thus, we suggest that future studies should pay more attention on this issue by specifying the demographic diversity types. Further, teams in Eastern countries should increase diversity as variety to improve their performance, whereas teams in Western countries should not only pay attention to team demographic variety, but also need to decrease team demographic disparity to avoid its negative effects on team performance. Overall, our findings have specific implications for companies to improve their performance through team demographic diversity management.

    Related Articles | Metrics
    The Block Item Pocket Method to Allow Item Review in CAT
    LIN Zhe; CHEN Pin; XIN Tao
    2015, 47 (9):  1188-1198.  doi: 10.3724/SP.J.1041.2015.01188
    Abstract ( 592 )  

    Most computerized adaptive testing (CAT) do not allow examinees to review items because it will drastically decrease measurement precision and bring about extra cheating strategies (Wainer, 1993; Wise, 1996). Allowing item review is essential to make CAT comparable with traditional tests. It also matters in application. Item review enables examinees to correct mistakes due to carelessness, which can further improve the precision of ability estimation. No such option may cause some negative consequences for their overall performance especially in high-stake examinations, such as tension or anxiety (Vispoel, Henderickson, & Bleiler, 2000). Therefore, it is worth trying if allowing item review could alleviate problems mentioned at the beginning (Wise, 1996; Vispoel, 2000, 2005).

    Several methods have been proposed, including the successive block method (Stocking, 1997) and the item pocket (IP) method (Han, 2013). However, both methods are limited in some ways. Stocking’s method does not allow examinees to skip items and requires a large number of blocks which may bring about some extra adverse effects because of frequent decision to go to next block. Han’s method can avoid limitations of Stocking’s. But it requires an appropriate IP size and may result in high bias in large IP size situation. The present study proposed the block item pocket (BIP) method which sets fewer but larger blocks with a proper total IP size. This method keeps advantages of Stocking’s and Han’s and overcomes their disadvantages.
    Two simulation studies of two response strategies were conducted to evaluate validity of the BIP method. Item parameters were randomly drawn from uniform distribution (b ~ U (-3, 3)) and (α ~ U (0, 2)). Each examinee was administered a fixed-length CAT with 30 items. The initial item for each examinee was randomly drawn from θ ~ U (-0.5, 0.5). For the CAT administration, the Maximum Fisher Information method was adopted to select items. The interim and final scores were estimated using MLE method in most conditions. When responses were less than 5 or when all answers were correct or wrong, EAP method was adopted. Each study contained five conditions: non-review, 1 blocks IP method, 2 blocks, 3 blocks and 6 blocks BIP method. Statistics like BIAS, MAE, and RMSE were used as evaluation criteria.

    Results indicated that: (1) BIP method had better estimate precision than IP method at low ability level under normal strategy; (2) When dealing with Wainer-like strategy, BIP method was far more precise than item pocket method at all ability levels; (3) As the number of blocks increased, estimate precision got closer to non-review condition. Advantages of this new method and future directions were discussed.

    Related Articles | Metrics