ISSN 0439-755X
CN 11-1911/B


    30 April 2011, Volume 43 Issue 04 Previous Issue    Next Issue

    For Selected: Toggle Thumbnails
    The Comparative Study on English Words, Chinese Words, Early Words and Pictures
    ZHANG Ji-Jia,WANG Juan,LIU Ming
    2011, 43 (04):  347-363. 
    Abstract ( 1694 )  
    Words’ naming and categorizing can reflect words’ processing course and mechanism. English words, Chinese words, inscription on oracle bones, Dongba pictograph and pictures were different symbolic systems in the history of human characters’ development. It had important theoretical values to investigate the features of these five symbolic systems.
    Two experiments were adopted to investigate five symbolic systems’ characteristics. Perceptive similarity evaluation task was performed in Experiment 1a. Two types of symbols were matched, participants were asked to evaluate the similarity level by 7-point scale. Each kind of materials had 16 stimuli including animal, plant, human organs, natural objects and tools. Twenty-six university students participated in this study and they were required to learn and get acquainted with the materials before the experiment. Semantic consistence evaluation task was performed in Experiment 1b. Twenty-five university students participated in this study. Naming and categorizing tasks were performed in Experiment 2. Two-factor within subject design was used: 5 (symbolic types: English words, Chinese words, inscription on oracle bones, Dongba pictograph and pictures) × 2 (tasks: naming, categorizing). During naming task, participants were asked to read the words loudly or name the symbols presented on the screen as quickly and correctly as possible. Naming time was collected by the computer and naming correct percent was recorded by the examiner. During categorizing task, participants were asked to decide whether the item a word or symbol representing belong to a certain category by pressing “F” or “J” on the keyboards. Stimulus-naming were presented in five blocks of 240 trials. Stimulus-categorizing were presented in five blocks of 480 trials. Twenty-four university students participated in this study.
    The results were as following: (1) Different symbols displayed differently in perceptive similarity judgment task and semantic consistence judgment task. Pictures, inscription on oracle bones, Dongba pictograph and Chinese characters had high similarity in perception, English and other symbols had low similarity; in semantic consistence judgment task, symbols combined with English and Chinese were processed fast. (2) Different kinds of materials were asymmetrical in naming and categorizing: English-reading and Chinese-reading were faster than English -categorizing and Chinese-categorizing, categorizing of inscription on oracle bones, Dongba pictograph-categorizing and picture-categorizing were faster than respective naming. (3) Regardless of naming or categorizing, the reacting of Chinese words and pictures were faster than the reacting of inscription on oracle bones and Dongba pictograph. English-naming was faster than naming of inscription on oracle bones and Dongba pictograph-naming; English-categorizing was slower than Dongba pictograph-categorizing. (4) English words, Chinese words and pictures were asymmetrical in naming and categorizing: English-reading and Chinese-reading were faster than picture-naming but English-categorizing, Chinese-categorizing was slower than picture-categorizing. (5) Dongba pictograph and inscription on oracle bones were asymmetrical in naming and categorizing: the reading of inscription on oracle bones was faster than Dongba pictograph-reading but the categorizing of inscription on oracle bones was slower than Dongba pictograph-categorizing. These phenomena were attributed to the structure characteristics and distinctive features of the five symbols, also the patterns of symbols recording language and participants’ proficiency of symbols. These results may provide psychological evidence for the character’s development as well as evolution rules, and also help to identify the natures of Dongba pictograph and inscription on oracle bones.
    Related Articles | Metrics
    The Modulation Mechanism of Emotional Words on Neutral Stimuli’s Preference
    GUO Jing-Jing,DU Yan-Peng,CHEN Yu-Xia,PENG Dan-Ling
    2011, 43 (04):  364-372. 
    Abstract ( 1231 )  
    As the basic unit of language, words not only conceptualize the objective world, but also convey affective information. Words containing both conceptual meanings and affective information are usually termed emotional words. Although a number of studies have been conducted to reveal the processing mechanism of emotional words, little is known about the modulation mechanism of emotional words on other stimuli, such as neutral stimuli.
    Using an evaluative conditioning paradigm, three experiments were conducted to explore the modulation mechanism of emotional words on neutral stimuli’s preference. In the evaluative conditioning paradigm, a neutral stimulus (unknown Korean words, Conditional Stimuli) was paired with an emotional word (Unconditional Stimuli) and presented repeatedly during the studying phase. In the evaluative phase, after the studying phase, a preference score for the neutral stimulus was obtained to find out whether the evaluation changed towards the direction of paired affective meaning. At the end of the experiment, contingence awareness tests were performed to check the awareness level of participants to relationships between neutral stimuli and emotional words.
    In experiment 1, a cover story was used to conceal the real experimental aim from the participants. The participants were asked to perform either one of two tasks on emotional words during each of two studying phases: word-class judgment versus affective judgment. Experiment 2 replicated experiment 1 except that we disturbed the fixed pairing of the relationship between neutral stimuli and emotional words to expel the potential confusion from the influence of the emotional words’ conceptual meanings on the modulation effects. In experiment 3, we changed the task from word-class judgment to word-form judgment which is distinguished from affective judgment in the word processing depth.
    The results of experiment 1 indicated that regardless of which task the participants performed, their evaluation of new words was always modulated by emotional words without awareness of the contingency relationship between neutral stimuli and affective words. Experiment 2 replicated the findings even after disturbing the fixed-pairing relationship of neutral stimuli and emotional words. To clarify that the results from the two tasks reflected an independent effect of the modulation mechanism rather than an effect of the same depth required by both tasks, experiment 3 adopted new tasks of word incline-vertical judgment and affective judgment. The results of experiment 3 replicated the modulation effects of emotional words on the preferences for the neutral stimuli.
    These results suggest that the modulation mechanism of emotional words on other stimuli is different from the processing of emotional words itself. The modulation process is not only independent of the awareness of the contingent relationship between CS and US, but is also not subjective to task demands. In other words, the modulation mechanism of emotional words on neutral stimuli might be based on automatic association learning, which reflects a bottom-up and automatic process. More importantly, the present study provides important evidence for the interactive view of emotion and cognition, which suggests that human behaviors are influenced by both emotional and cognitive factors.
    Related Articles | Metrics
    Effects of Spatial Frequencies on Recognition of Facial Identity and Facial Expression
    WANG Ya-Min,WANG Zhi-Xian,HUANG Ya-Mei,JIANG Jing,DING Jin-Hong
    2011, 43 (04):  373-383. 
    Abstract ( 1185 )  
    By changing configural or featural/category information, White (2002) revealed that configural changes mainly interfered with facial identity processing while featural alterations largely reduced facial expression processing. With this technique, Goffaux, Hault, Michel, Vuongo, and Rossion (2005) presented that low spatial frequency played in configural changes detection, whereas featural changes detection depended on high spatial frequency. Based on these two studies, we can draw a conclusion that low spatial frequency plays an important role in facial identity recognition while high spatial frequency plays in facial expression recognition. Can this conclusion be really supported by experiments?
    To test this hypothesis, we conducted three Garner experiments in current study. In terms of the hypothesis, high spatial frequency enhances facial expression recognition but not facial identity recognition, while low spatial frequency facilitates facial identity recognition but not facial expression recognition. Dissociation could be found in recognition of facial identity and facial expression.
    Three Garner experiments were performed on 96 participants, with each 32 participants. In experiment 1, we measured Garner effect between facial identity and expression with full spatial face photos. In experiments 2, we complete the measurement on low spatial frequency filtered photos, and in experiment 3 the same measurement were made on high spatial frequency filtered face photos.32 participants were recruited for each experiment.
    Results of experiment 1 indicated that obvious Garner effects were observed in either identity or expression recognition. In experiment 2, high spatial frequency enhanced facial expression recognition by lowering Garner effect, but it had no influence on facial identity recognition. In experiment 3, low spatial frequency exerted no obvious influence on either identity or expression recognition. Analysis of dissociability (measurement of Garner effect) and difficulty (measurement of selective reaction time in base line) showed that high spatial frequency selectively affected expression detection.
    In conclusion, high spatial frequencies were possibly used by us to distinguish facial expression information from identity information in hard-detected recognition task. Low spatial frequencies were easy to interact with both processing level and experimental task.
    Related Articles | Metrics
    An ERP Study of Mix Strategy Selection in Mental Arithmetic
    CHEN Ya-Lin,LIU Chang,ZHANG Xiao-Jiang,XU Xiao-Dong,SHEN Wang-Bing
    2011, 43 (04):  384-395. 
    Abstract ( 1159 )  
    Problem size, split, parity and correctness of answer affect mental arithmetic problem-solving. Behavioral and cognitive neuroscience researches on mental arithmetic showed that there are different strategy selection in problem size effect, split effect, and parity effect. Moreover, these three factors can hardly be distinct from each other. Split and parity even have contradictive effect. These caused confusions in past researches. And also, few researches noticed the mix of the factors. How do people make their choice when these factors are mixed? One hypothesis is that split has priority compare to parity. Moreover, the electrophysiological feathers of parity effect remained unknown. There were also some problems of the presentation mode in past researches. The present study tried to give an answer with a new presentation mode.
    High-density event-related potentials (Neuroscan) of 15 college students were recorded. The stimuli consisted of mathematic equations between 0 and 60. Each trial began with a fixation cross shown for 500 ms. 500ms later, the answer shown for 250ms. Another 500ms later, the equations appeared and remained on the screen for 3000 ms followed by the varied ITI of 800-1200ms. There were 6 blocks, which consisted of 110 trials. All the trials and blocks were randomized on-line. Repeated measures ANOVA was used for data analyze.
    An arithmetic N400 was present for small incorrect problems. The parity congruent problems caused more significant N400 than parity incongruent problems in small split and small problems. A late positive wave was present and manipulated by split and parity for large problems. And the parity congruent problems elicited lower late positive slow wave than parity incongruent problems in small split and large problems. No significant difference emerged between parity congruent and incongruent problems in large split and large problems.
    Problem size, split and parity are the important variables in mental arithmetic research. When processing with these factors, split has priority over parity both in small problems and large problems. Moreover, in small problem processing, the split and parity have effects on retrieval; while in large problem processing, the effect of parity appears in the condition that split information isn’t enough to produce a direct judgment.
    Related Articles | Metrics
    The Effects of Delay Time and Foreknowledge on Task Switching
    HUANG Si-Lin,LIN Chong-De
    2011, 43 (04):  396-402. 
    Abstract ( 1097 )  
    Task switching refer to the process frequently from one task to another task, and the switching process is an important function of executive control, which control switching course between two tasks which compete for the same cognitive resources in working memory. In task switching, switch cost is the performance differences in task-switch and task-repetition trials. Based on the theories of the subtractive method reaction time, switch cost is the index of executive control. Recently, some researches have attempted to specify nature and origins of switch cost to uncover the mechanisms underlying task switching performance. Theoretically, there is disagreement as to whether switch cost actually represents the task-set reconfiguration or the task-set inertia. The aim of the present study was to investigate the effect of delay time and foreknowledge on task switching.
    Explicit task-cueing paradigm was used in the present study. There were 29 participants (12 males and 17 females) volunteered to take part in the experiment. A 3 (delay time: 100ms, 600ms and 1000ms) × 3 (foreknowledge type: no foreknowledge, part foreknowledge and full foreknowledge) ×2 (task type: switch task and repetition task) within-subject design was conducted to examined the mechanisms of effects of foreknowledge on task switching under the conditions of difference delay time. The correct RT and switch cost were measured.
    The results showed: (1) the main effect of delay time was significant, the switch cost with 100ms was larger than with 1000ms, however, no difference significantly among 600ms and 100ms or 1000ms. (2) The amount of switch cost depend on foreknowledge, switch cost was smaller with full foreknowledge than with part and no foreknowledge, but no difference significantly between the later two. (3) The interaction between delay time and foreknowledge did not appear.
    The findings suggested that delay time and foreknowledge acted on task-set inertia and task-set reconfiguration severally, which were two differences cognitive processes underlying switch cost.
    Related Articles | Metrics
    Impact of Trier Social Stress Test (TSST) on Salivary Cortisol Secretion
    YANG Juan,HOU Yan,YANG Yu,ZHANG Qing-Lin
    2011, 43 (04):  403-409. 
    Abstract ( 1219 )  
    Trier Social Stress Test (TSST) is a widely used psychosocial stress paradigm conducted in a laboratory setting which has been shown to reliably elicit a physiological stress response in various study samples. However, no study to date has investigated the responses to the TSST in Chinese sample.
    Therefore, in the present study, healthy male and female volunteers (N=30) performed the standardized TSST protocol, during which several subjective measures (Profile of Mood State and the visual rating scale) and objective measures (heart rate and salivary cortisol) were assessed.
    The results showed that participants exhibited a significant increase in heart rate and salivary cortisol, and reported more anxiety immediately following the TSST compared to baseline. Furthermore, while men and women did not differ with respect to physiological response to TSST, women did report higher levels of anxiety compared to men.
    The findings demonstrated that TSST is a protocol with good applicability in Chinese participants and could be used to conduct psychosocial stress research in China. Sex differences failed to reliably discriminate between male and female in objective measures to stressors which were achievement-oriented.
    Related Articles | Metrics
    Relationship between Adolescents’ Alienation and Pathological Internet Use: Testing the Moderating Effect of Family Functioning and Peer Acceptance
    XU Fu-Zhen,ZHANG Wen-Xin
    2011, 43 (04):  410-419. 
    Abstract ( 1911 )  
    Internet as an important media has become an inevitable part of adolescents’ lives. Although Internet meet various needs of adolescents, such as expressing emotions and seeking sense of belonging, misuse of Internet could result in psychological, social and academic problems during adolescence (Young & Rodgers, 1998; Davis, 2001; Lei & Yang, 2007). In this study, we use the term of pathological Internet use (PIU) to refer to the degree of adolescents’ misuse of Internet and its negative outcomes.
    Alienation is a negative emotional state, reflecting individual incompetence in dealing with others and establishing an effective bond with social groups. Existing researches demonstrated that adolescents with high alienation were inclined to use Internet excessively. Researches on the relationship between pathological Internet use and adolescents’ negative emotions reveal that positive factors of family and peer groups may be of important buffering effects (Thomas & Schare, 2000; Slater, 2003; Gross, 2004; Gao & Chen, 2006; Yeh et al., 2008). Thus it is reasonable to assume that family functioning and peer acceptance could have a moderating effect on the relationship between adolescents’ alienation and their pathological Internet use.
    The present study was to investigate the descriptive characteristics of adolescents’ pathological Internet use and its relation to alienation, and to find whether or not healthy family functioning and peer acceptance could protect adolescents with high alienation from pathological Internet use. A total of 549 adolescents in the 7th (54 boys and 51 girls, Mage=12.84±0.48), 8th (54 boys and 51 girls, Mage=13.92±0.40), 9th (50 boys and 49 girls, Mage=15.06±0.50), 10th (81 boys and 43 girls, Mage=16.29±0.49) and 11th (62 boys and 54 girls, Mage=17.24± 0.43) grades of 5 high schools completed self-report questionnaires and Peer Nomination survey. The self-report questionnaires used in this study included Adolescent Pathological Internet Use Scale (APIUS), General Alienation Scale and Family Assessment Device.
    According to the screening criterion, 8.93% adolescents who had PIU scores higher than 3.15 were identified as specific PIU, indicating that their level of pathological Internet use was rather high. Female adolescents reported a higher level of salience of PIU than male ones, and adolescents of grade seven reported more negative outcomes related to PIU than those of other grades except for grade eight. Adolescents with a higher level of alienation also tended to report more pathological Internet use. Family functioning could moderate significantly the relationship between adolescents’ alienation and their pathological Internet use. Specifically, healthy family functioning played a protective role for adolescents with higher alienation in reducing their pathological Internet use. Despite that existing literatures revealed peers provided important support for adolescents, the protective role of peer acceptance was not observed in this study.
    Overall, our findings confirmed the existence of protective effect of healthy family functioning, which suggested that both characteristics of adolescent’s emotion and the positive effect of family functioning should be considered together in order to promote adolescents to use Internet properly.
    Related Articles | Metrics
    The Working Alliance in Clients’ Eyes: A Qualitative Analysis
    ZHU Xu,JIANG Guang-Rong
    2011, 43 (04):  420-431. 
    Abstract ( 1250 )  
    The working alliance has been a focus of psychotherapy research for several decades in the western literature. Although theorists differ somewhat in their conceptualizations of the alliance, most of them agree that the core of this construct is the collaboration between therapist and client that emphasizes the contributions of both participants. Most of our knowledge in this field derives from theorists’ hypotheses, and clients’ perspective is often neglected. In fact, it was often assumed that clients use the same conceptual dimensions as therapists do to rate the collaborative processes in psychotherapy. This assumption, however, has been challenged by reported low correlation between client-therapist alliance ratings and by results of qualitative studies that show divergence between clients’ and therapists’ perspectives. Moreover, the literature regarding clients’ actual experiences of the alliance is sparse. In an attempt to examine clients’ perspectives in viewing working alliance and the cultural adaptability of western alliance theories to Chinese culture, the present study sought to delineate the alliance from Chinese clients’ perspective via a qualitative approach and addressed the concordance between clients’ perceptions and theoretician-derived views of the alliance.
    Twenty clients at a university counseling center consented to participate in the study. Clients were interviewed by phone about their experiences of being collaborative with their therapists and their opinions about what is important to this collaborative process in counseling. Data were analyzed by a research team using Consensual Qualitative Research (CQR) (Hill et al., 2005; Hill, Thompson, & Williams, 1997). CQR requires that research team members reach consensus about the classification and the meaning of data through three steps, namely, domain coding, core ideas, and cross subject analysis.
    Six domains with regard to the collaborative process in counseling emerged from the data: bond, task, engagement, collaboration pattern, development, and influencing factors. Bond is the emotional and personal aspect of the client-therapist relationship, which serves as the emotional base for collaborative work. Task refers to the activities and things client and therapist need to do during the collaborative process, which is the content of the collaborative work. Engagement implies the degree to which client and therapist devote themselves to the process, lending an impetus to the collaboration. These three domains can be viewed as components of working alliance and the other three as descriptions of major features of the alliance.
    The study results suggest that the working alliance is perceived by Chinese clients to include bond, task, and engagement. This finding is comparable to current theories of alliance in the western literature and advances conceptualizations of the alliance especially when culture differences are concerned.
    Related Articles | Metrics
    Social Value Orientation and Cooperation in Asymmetric Social Dilemmas
    LIU Chang-Jiang,HAO Fang
    2011, 43 (04):  432-441. 
    Abstract ( 1366 )  
    The effect of social value orientation on cooperative behavior in social dilemmas has been widely examined in social psychology research. However, only a small to medium effect size has been found. According to the appropriateness model (Weber, Kopelman, & Messick, 2004), decision making in social dilemmas can be shaped by one’s identity, situational recognition, and the application of rules. Thus, individual characteristics interacting with situational factors affect decision making in social dilemmas. Therefore, the unequal resources with which individuals are endowed in asymmetric social dilemmas set up differences in endowment dominance and serve as contextual cues that may shape cooperative behavior differently for prosocials and proselfs. This study hypothesized that prosocials would be sensitive to difference in endowment dominance whereas proselfs would be immune to the difference (H1) and that, compared to proselfs, prosocials would exhibit a higher level of cooperation when they were endowed with less or when their endowments were worth less (H2).
    The current study examined the above hypotheses via two experiments. In Experiment 1, participants were randomly assigned to an advantaged condition or a disadvantaged condition. Their social value orientation was measured with a nine-item decomposed game instrument. 90 participants interacted in a fictional two-person prisoner’s dilemma situation. The main results showed that 1) prosocials allocated more tokens to their partner when their endowments were worth less than when their endowments were worth more; and 2) prosocials allocated more to their partner than did proselfs when their endowments were worth less.
    Experiment 2 was conducted in a fictional 4-person public good dilemma situation. 191 participants were randomly assigned to an equal condition, an advantaged condition or a disadvantaged condition. Again, their social value orientations were also measured. The main results showed that 1) the amount of their endowments that prosocials contributed in the disadvantaged condition was significantly higher than the average amount of endowments that they contributed in the equal condition and in the advantaged condition; and 2) prosocials contributed more to the group account than did proselfs only when they were endowed with less. This was not the case in either the equal condition or the advantaged condition.
    The results from both experiments supported our hypotheses and indicated that the effect of social value orientation on cooperation was moderated by dominance differences stemming from the initial resources that individuals were endowed. The results enrich our understanding on how prosocials and proselfs differ in their response to different distributions of initial resources. It also suggests that such social value orientation effects are still open for exploration. Future research should further explore how prosocials and proselfs behave differently when they face social dilemmas with different situational characteristics.
    Related Articles | Metrics
    The Effect of Construal Level on Intertemporal Choice and Risky Choice
    CHEN Hai-Xian, HE Gui-Bing
    2011, 43 (04):  442-452. 
    Abstract ( 1637 )  
    It has been proved that there exist a number of similarities between intertemporal choice and risky choice. The similarities imply that some equal processes may be included in these two choices.
    Construal level theory (CLT) proposes a new explanation of these similarities. In CLT, time and probability are both dimensions of psychological distance, and are represented in similar ways (Trope & Liberman, 2003, 2010). Distant or uncertain events are represented as high level construals, whereas near or certain events are represented as low level construals. As events become more distant or less certain, the high-level construals become more influential whereas the low-level construals become less influential in shaping preferences. Meanwhile, delay or probability is represented as low-level construal in valuation of delayed or risky choices, whereas the magnitude of money is represented as high-level construal (Sagristano, Trope, & Liberman, 2002; Trope & Liberman, 2003).
    The purpose of this paper is to shed some additional light on the underlying processes of intertemporal and risky choice by investigating the effects of construal-level mindsets on them. We hypothesize that construal-level mindsets have similar effects on intertemporal and risky choice because time and probability are represented in similar ways. As the magnitude of money (high-level construal) is more influential whereas time and probability (low-level construal) are less influential, participants with high-level mindsets might value delayed and risky money higher than those with low-level mindsets.
    Two experiments were conducted to test this hypothesis. In experiment 1, 140 college students were procedurally primed to high or low level construal mindsets by considering questions related to “why” or “how” they engaged in certain actions, then some participants accomplished intertemporal choices (experiment 1a) and others made risky choices (experiment 1b). In experiment 2, 139 college students were procedurally primed to high or low level construal mindsets by reacting to global letters or local letters, then accomplished intertemporal choices (experiment 2a) or risky choices (experiment 2b). The results indicate that participants with high level construal mindsets value delayed and risky money higher than those with low-level mindsets, both in experiment 1 and 2. These results are consistent with the hypothesis.
    The results may have some important implications. Firstly, time and probability are represented in a similar way, which can explain some parallels between intertemporal and risky choices, especially common difference effect and common ratio effect. Secondly, construal-level mindsets can affect intertemporal and risky choices, which reminds us to pay more attention to the influence of the decision context that may trigger different mindsets. Thirdly, construal-level mindsets can be triggered by both thought and visual perception, which gives additional evidence that the connection between psychological distances and construal levels could be activated automatically and implicitly.
    Related Articles | Metrics
    A Comparison of Three Confidence Intervals of Composite Reliability of A Unidimensional Test
    YE Bao-Juan,WEN Zhong-Lin
    2011, 43 (04):  453-461. 
    Abstract ( 1274 )  
    The widely used coefficient a may underestimate or overestimate reliability when its premise assumption is violated and therefore is not a good index to evaluate reliability. Composite reliability can better estimate reliability by using confirmatory factor analysis (see e.g., Bentler, 2009; Green & Yang, 2009). As is well known, point estimate contains limited information about a population parameter and could not give how far it could be from the population parameter. The confidence interval of the parameter could provide more information. In evaluating the quality of a test, the confidence interval of composite reliability has received more and more attention in recent years.
    There are three approaches to estimate the confidence interval of composite reliability of a unidimensional test: Bootstrap method, Delta method and directly using the standard error in the output of an SEM software (e.g., LISREL). Each of the three approaches produces a standard error of composite reliability. Then the confidence interval can be easily formed based on the standard error. Bootstrap method provides an empirical result of the standard error of composite reliability and is the most credible, but the method needs data simulation technique and is not be easily mastered by general applied researchers. Delta method computes the standard error of composite reliability by approximate calculation, and the method is much simpler than Bootstrap method. LISREL software can directly give the standard error of composite reliability, and this method is the simplest among the three methods.
    To evaluate the standard errors of composite reliability obtained by Delta method and LISREL software, we compared them with that obtained by Bootstrap method, because the latter can be treated as the true value in theory. A simulation study was conducted to the comparison. Four factors were considered in the simulation design: (a) the number of items on each test (k=3, 6, 10, and 15); (b) factor loading (high, medium and low); (c) sample size (N=100, 300, 500, and 1000); (d) the method for calculating the standard error of composite reliability (Bootstrap, Delta, and LISREL). Totally, 48 treatment conditions were generated in terms of the above 4-factor simulation design (i.e., 48=4×3×4×3).
    The simulation results indicated that the difference between the standard errors obtained by Delta method and Bootstrap method was ignorable under each designed condition, except when sample size was small (less than 200)and standardized factor loadings were not high (less than 0.7). However, there was substantial difference between the standard errors obtained by LISREL software directly and Bootstrap method under each designed condition. Noting that the result from Bootstrap method can be treated as the true value, we recommended that Delta method could be adopted to estimate the confidence interval of composite reliability of a unidimensional test. At the same time we revealed that the standard error directly obtained by LISREL software is severely biased.
    We used an example of a unidimensional test to illustrate how to calculate composite reliability and its confidence interval by using Delta method based on LISREL output. We also showed that the same results could be directly obtained by using SEM software Mplus that automatically calculates the confidence interval with Delta method and presents the confidence interval.
    Related Articles | Metrics
    Modeling Language Components in Mathematical Items Using Multiple Random Effects IRT Models
    YANG Xiang-Dong,CHIU Lorraine,WEI Yun
    2011, 43 (04):  462-472. 
    Abstract ( 1015 )  
    Language components in mathematical word problems may have profound impacts on the problem-solving processes engaged by examinees. When a math problem is presented by means of language statements, translation of these language statements into mathematical propositions must be carried out properly before relevant information can be integrated into a coherent representation of the problem. As many studies have shown, the impacts of such language components are unlikely to occur uniformly across examinees. Yet existing approaches that address this issue focus only on (1) examining the first-order correlation between the reading and mathematical abilities of examinees, (2) identifying a general language factor through factor analysis, and (3) examining the average impacts of specific language features on math item properties through linear regression or linear logistic test model (LLTM). While the first two approaches fall short of providing targeted information about the specific aspects of language components that impact examinees’ mathematical problem solving, the third approach is based on an unrealistic assumption of constant impacts of such language components. Therefore, an alternative approach needs to be formulated in order to model individual-specific impacts of particular language features on math item proficiency.
    The current study applies item response theory (IRT) models with multiple random effects to investigate the effects of language components in mathematical items on examinees’ performances. This approach starts by studying relevant literature and searching for a cognitive processing model for mathematical problem solving, and then identifies specific item stimulus features in the problem statements to represent the language components according to the model. By encoding items in a third-grade mathematical test according to the set of identified stimulus features, a set of models that operationalize different cognitive principles are then applied to the data obtained from the test. This approach provides a rigorous method for examining not only the average impacts of specific language features on item properties over the sample, but also variations of such impacts among examinees. Consequently, it allows researchers to investigate the interactions between the characteristics of a mathematical problem and the cognitive abilities of a particular examinee.
    Results from this study show that, among the six item stimulus features that were identified based on Mayer’s cognitive theory of mathematical problem solving and its associated literature, all but one significantly affect the difficulty of mathematical items. Difficult items tend to have more words, feature a higher proposition density, and require the examinee to either translate information from a graph or table or to generate mathematical equations from the problem statements. In addition to the variation of mathematical ability across examinees, proposition density of the mathematical items shows differentiated impacts for different examinees in the testing sample. While the random effects of proposition density inversely relate to the mathematical abilities of examinees in general, proposition density exerts more impacts on the math problem solving for examinees with low abilities, and such impact is diminished as the examinee’s ability increases.
    Related Articles | Metrics