Retro-cue effect (RCE) refers to the phenomenon that individuals can use retro-cues to improve their visual working memory (VWM) performance of target items after the disappearance of memory targets. To explain the mechanism of RCE in VWM, five different hypotheses have been proposed by previous studies: the hypothesis of enhancing target representations, the hypothesis of forgetting non-target representations, the hypothesis of preventing memory degradation, the hypothesis of preventing interference from probe array and the hypothesis of cognitive phase separation. Although RCE has been repeatedly observed in previous studies, the mechanism of RCE remains unclear. In this study, we conducted three experiments to test these hypotheses.
In Experiment 1, participants were asked to memorize four colors in a recall task. They needed to recall the color of the target item when the probe array presented. There are three experimental conditions, the normal cue condition, the short interval no-cue condition, and the long interval no-cue condition. In the normal cue condition, a memory array (four colored squares) presented for 200 ms. Then, the memory array disappeared for 450 ms and a retro-cue presented, followed by a 900 ms blank. Then the probe array presented. In the short interval no-cue condition and long interval no-cue condition, no retro-cue presented after the memory array, but the probe array would present after the memory array disappeared for 450 ms (short interval no-cue condition) or 1400 ms (long interval no-cue condition). The design and procedure of Experiment 2 were similar to those of Experiment 1, except we used a grey-wheel cue condition and a colored-wheel cue condition to replace the long interval no-cue condition. These two new conditions were similar to the normal cue condition, except the retro-cue would appear with a distractor of a gray wheel (grey-wheel cue condition) or with a distractor of a colored wheel (colored-wheel cue condition). The design and procedure of Experiment 3 were similar to those of Experiment 2, except a long-grey-wheel short-cue condition and a long-grey-wheel long-cue condition were used to replace the normal cue condition and colored cue condition. In the long-grey-wheel short-cue condition, the retro-cue presented for 100 ms, but the grey wheel presented for 1000 ms. In the long-grey-wheel long-cue condition, the retro-cue and grey wheel presented for 1000 ms.
The results of Experiment 1 showed that there was no significant difference in memory performance between the short interval no-cue condition and long interval no-cue condition, while the performance of the normal cue condition was better than that of short and long interval no-cue conditions. The results of Experiment 2 showed that participants obtained the RCE under the normal cue, the grey-wheel cue, and the colored-wheel cue conditions. However, the degrees of RCE obtained by these three conditions were different (normal cue > grey-wheel cue > colored-wheel cue). The results of Experiment 3 showed that participants obtained the RCE under the grey-wheel cue condition, the long-grey-wheel short-cue condition, and long-grey-wheel long-cue condition. The degree of RCE obtained by these three conditions was the same.
The results of the present study supported the hypothesis of cognitive phase separation, which suggested that the retro-cue can separate the internal attention reallocation process and decision-making process, avoiding cognitive interference from the probe display. In addition, when the retro-cue was accompanied by the distractors, the type of distractors (rather than the presentation time) affected the degree of RCE. This study provides further evidence for the hypothesis of cognitive phase separation, which is crucial to solve the debate on the mechanism of RCE and to understand the relationship between attention and VWM.
Languages may differ in terms of the proximate units of phonological encoding in spoken word production. It is widely accepted that phonemes are the primary units used in Indo-European languages. However, it is controversial what the functional units (syllables or phonemes) are in Chinese spoken word production. In the framework of proximate unit principle hypothesis for Mandarin Chinese spoken word production, phonological content is activated in a syllable while its syllabic frame is retrieved. Both syllabic content and frame are linked sequentially and tone is also specified at this point. Crucially, syllables are retrieved at an early stage of word form encoding, and then the phonemes of the syllable are retrieved in parallel and linked to their position in the syllable frame. The present study investigated the temporal courses of syllables and phonemes in word production using an implicit priming paradigm, and electrophysiological signals were measured concurrently. Specifically, participants were young native Chinese speakers with low level of English ability for excluding the influence of the second language.
In an implicit priming paradigm, four sets of word pairs served as experimental stimuli. Each set consisted of four pairs, and the first word of a pair was the cue word, while the second was the response word. In the homogeneous conditions, four response words in a set with common phonological properties. For instance, the four response words in a set were危机(/wei1ji1/, crisis), 围墙 (/wei2qiang2/, bounding wall), 尾巴 (/wei3ba/, tail), and 卫星(/wei4xing1/, satellite), which shared the first syllable but not tone. The heterogeneous conditions used the same word pairs, but assigned them to sets in which they didn’t share any properties. Two homogeneous conditions (syllable shared and initial phoneme shared) were used in the study. Participants were asked to memorize the sets of four pairs of associate words, after which they were presented with the cue words, and were required to produce the response words.
Behavioral results didn’t show any syllable or phoneme effects. EEG results revealed that syllable overlap condition elicited significant effects in the time window of 100~400 ms, and the effect firstly showed up in the left and right anterior regions, following in the middle anterior region, and finally in the left and right anterior regions, right and middle posterior regions. While the phoneme effects only appeared between 500~600 ms after the cue words were presented, and were distributed in the right anterior and posterior regions. Both the syllabically and phonemically homogeneous conditions were characterized by more positive ERP waveforms when compared to the heterogeneous conditions. We suggest that syllable overlap effect arises at an early stage of phonological encoding, whereas the phoneme overlap effect possibly arises at a late stage of phonological encoding, or phonetic encoding in Chinese spoken word production. Our findings provide support for the assumption of the proximate unit principle.
The conceptual mapping perspective places metaphor within the framework of embodied cognition. Conceptual Metaphor Theory holds that individuals cannot directly understand the meaning of abstract concepts which are constructed by perceptions of bodily experiences. Thus, moral concepts are acquired through physical experience. Previous studies on the orientational metaphor of moral concepts have found that there is a psychological reality for the metaphorical connection between moral concepts and vertical orientation. However, there are few studies on the metaphorical connection between moral concepts and left-right horizontal orientation. In this study, undergraduate and postgraduate students were selected to examine the metaphorical connection between moral concepts and horizontal orientation and its influence on the depth of conceptual processing through six experiments.
Experiments were performed in E-prime 2.0. In experiment 1, participants used a forced-choice task to place Chinese moral words on the left or right side of a cartoon person, in order to explore the metaphorical connection between moral concepts and horizontal orientation in the mind. During experiment 2, the Stroop paradigm was adopted and participants read words aloud and made moral judgments about words appearing on the left or right side of the screen. The aim was to explore the metaphorical connection between Chinese moral concepts and the horizontal orientation at high levels of conceptual processing. Experiment 3 utilized the same paradigm as experiment 2, but participants did not read words aloud in order to explore the metaphorical connection at medium levels of conceptual processing. Experiment 4 used the same paradigm as experiment 2. Participants judged the truth or falsehood of words rather than moral judgments in order to explore the metaphorical connection at low levels of conceptual processing. In experiment 5, a priming paradigm was adopted to explore the mapping of the target domain to the source domain at low levels of conceptual processing. A visual cue was shown on the left or right side of the screen. Then a word appeared in the position of the visual cue and participants judged whether it was true or false. In experiment 6, a priming paradigm was adopted to explore the mapping from the source domain to the target domain at low levels of conceptual processing. A word was presented on the screen and participants judged its truth and falsehood. Then participants classified and judged the letters “p” or “q, ” which were presented on the left or right side of the screen.
In experiment 1, we registered participants who placed moral and immoral words on the left or right side of the figure. Chi-square test was used to analyze data. Results showed that the number of moral words placed on the left side of the figure was significantly higher than that of words placed on the right side. Regarding immoral words, results were the opposite. In experiments 2, 3, and 4, we collected the reaction time of participants regarding lexical judgment, and repeated ANOVAs were used to analyze data. Results of experiments 2 and 3 revealed a significant Stroop effect. That is, when moral words were presented on the left side, reaction time significantly decreased. However, experiment 4 did not show this effect. In experiment 5, reaction time of participants in the judgment task was recorded. Experiment 5 utilized the same method to analyze data as experiment 4. The effect of spatial priming on moral words was not found in experiment 5. In experiment 6, reaction time of participants in the judgment of letter classification was recorded. Data from experiment 6 was also processed in the same way as experiment 4. Effects of moral priming on the categorization of letters were not found.
In conclusion, results from the six studies suggest that the psychological reality of morality is oriented to the left and immorality to the right. This metaphorical connection is affected by the depth of conceptual processing.
Metaphors are among cognitive styles, which refer to the construction of psychological ideas using other concepts. The core of our conceptual system is derived directly from our experience of perception, physical movement, and social characteristics. Embodied philosophy is the philosophical foundation of cognitive linguistics and closely related to the study of metaphors. In terms of spatial metaphors, “up” is generally believed to represent morality, enthusiasm, and high status, while “down” represents immorality, negativity, and low status. With regard to weight metaphors, “heavy” generally denotes high authority and more respect, while “light” depicts low authority and less importance. Kinship is one of the social relationships formed on the basis of blood relations and marriage. In every language, numerous words represent kingship, which are called kinship words. The purpose of this study is to explore whether kinship words can be represented by concepts of “up-down” and “heavy-light” and whether gender culture has an impact on the spatial and weight metaphors of kinship words.
Three experiments are conducted in this study, with over 100 volunteer participants for each experiment, involving 30 individuals from three nationality groups. All the participants are from Yunnan Province, with the same academic qualifications but different cultural backgrounds. The kinship words in the materials are paired with gender from the same generation, such as “father-mother”, “brother-sister”, and “son-daughter”. In Experiment 1, a pair of kinship words is presented on the screen vertically, while the participants are asked to judge whether both words are kinship words. In Experiment 2, a pair of kinship words is presented on the screen horizontally before a picture of a balance is shown, and the participants are asked to judge the tilt direction of the balance. In Experiment 3, a picture of a balance appears before a pair of kinship words is presented, and the participants are asked to judge the position of the male/female words.
Studies have shown that (1) the Bai, Yi, and Mosuo people are influenced by the concept of seniority when processing kinship words, which is closely related to the tradition of showing respect and love to elders. (2) Kinship words could be represented by “up-down” and “heavy-light” concepts, with “up/heavy” representing high power and more respect and “down/light” representing low power and less importance. (3) Cultural schema affects the metaphorical representation of kinship concepts. Moreover, gender culture has an influence on the spatial space and metaphor of kinship words. Metaphorical consistency between male and “up/heavy” among the Yi people and between female and “up/heavy” among the Mosuo people is observed, whereas male/female concepts are not significantly related to “up/down” and “heavy/light” concepts among the Bai people.
Most previous studies on kinship words have studied spatial metaphors and mainly focused on the structure of kinship words. The present study considers spatial and weight metaphors and is innovative in terms of gender culture perspectives. We find that the gender concepts of the three nationalities have different connections with spatial and weight concepts, which demonstrates the impact of gender culture on cognition.
Music and speech share many acoustic commonalities and cognitive mechanisms. Previous studies have found that music training can improve categorical perception (CP) of Mandarin tones in adult musicians. However, it remains to be established whether music training can enhance the categorical perception of Mandarin tones in young children and whether the training effects can be influenced by the training duration.
The present study used a 2 (group: music training vs no-training) × 3 (test time: pre vs 6-month post vs 12-month post) between-and-within-subjects design to investigate the effects of music training on 4- to 5-year-old children’s CP of a Mandarin lexical tone continuum (from Tone 1 to Tone 2). The music training consisted of 110 sessions, 30 minutes per session, and three sessions per week for 12 months involving 20 preschoolers. The children were assigned to two groups, music training group (n = 20, age range from 49.69 months to 51.42 months, SD = 2.91 months) and control group (n = 20, age range from 51.69 to 52.56 months, SD = 3.0 months). In the music training group, the instructor guided children in activities leading to playing the small carillon, while children in the no-training group were given routine class activities. Each session of music training consisted four parts: Part 1 was “listen and sing songs” in which children learned to master notes and focus attention on subtle pitch changes; Part 2 was “listen and discriminate musical notes”, children learned to play a single note accurately according to the background music; Part 3 was “listen and play the carillon”, children listened to pitch changes in the background music, sang the notes and played the whole song melody; Part 4 was “play the carillon along with actions”, children listened to the background music and learned to play the carillon along with simple dancing actions. Children’s CP of tone continuum was measured before the learning began, after 6- month and after 12- month training using two tasks (identification test and discrimination test).
This study investigated if music training can enhance children’s boundary position, boundary width, within-category and between-category discrimination accuracy in CP of Mandarin Tone 1 and Tone 2 through 2 (group: music training vs no-training) ×3 (test time: pre vs 6-month post vs 12-month post) repeated measures ANOVA. The results revealed that although the perceptual boundary positions and ability to discriminate between-category tone pairs were unaffected by training, the boundary width values and within-category discrimination accuracies differed significantly between the experimental and control groups. The analysis of boundary width values and within-category discrimination accuracy revealed a significant interaction between group and test time. An analysis of simple effects further indicated that in the pretest and 6-month posttest, there was no significant effect between music training group and no-training group. In the 12-month posttest, the boundary width decreased significantly and the within-category discrimination accuracies increased significantly in the music training group, while no significant differences were found on boundary width and within-category discrimination accuracy in the control group. These results suggest that long-term music training can enhance children’s CP of Mandarin tonal contrasts.
In conclusion, our results supported the OPERA theory that music training can raise the steepness of boundary widths and enhance children’s sensitivity to subtle pitch differences between within-category sounds in the presence of robust mental representation in the service of CP of lexical tonal contrasts.
Reading comprehension is one of the most complex behaviors that we engage in on a regular basis. Decoding and language comprehension are two important components of reading comprehension. While up to now, little research has been devoted to directly explore the effects of phonological memory and central executive function on diverse reading comprehension components (decoding and language comprehension). In addition, the primary school stage is a critical period of development of children's reading ability and cognitive ability. However, few research has focused on the developmental changes in the relationship of phonological memory, central executive function and reading comprehension among children in different grades. Therefore, the purpose of this study is to examine the contribution of phonological memory and central executive function on decoding and language comprehension of children in different grade levels.
Two hundred and fifty-six children completed Chinese Characters Reading, Chinese Word Reading, and Pseudo-Word Reading to investigate decoding ability, as well as a Listening Comprehension Test to exam language comprehension. Additionally, we used Digit Span and Non-word Span to measure children's phonological memory, and GO/NOGO task, Stroop task, Digit updating task and Digit shifting task to investigate children's central executive function. Correlation analysis and structural equation models (SEM) were used to investigate the effects of phonological memory and central executive function on decoding, language comprehension of children in different grades.
Results indicated that, the effect of phonological memory and updating on decoding were significant. Also, the effect of updating and shifting on language comprehension were significant. On further analysis, we divided the four grades into two levels based on the reading stage theory, that third grade and fourth grade as one level, fifth grade and sixth grade as another level. In the third grade and fourth grade level, phonological memory and updating accuracy contributed to decoding. Moreover, updating accuracy and shifting cost predicted language comprehension. In the fifth grade and sixth grade level, only the effect of updating on decoding was significant.The results indicated that, the prediction of phonological memory on decoding disappeared as children progress through school. More importantly, we found that updating had a relatively stable contribution to decoding across grade levels.
It revealed that different functions of working memory played different roles in reading, which seem that not all cognitive abilities are equally important in the reading process. Furthermore, phonological memory and central executive function produced different predictive effects in decoding and language comprehension as the grades grew.
The majority of research on eating behaviors has been limited to an almost exclusive focus on pathology and is centered on the female group. Recently, one form of adaptive eating that has gained recognition is “intuitive eating”, which is defined as eating according to internal physiological cues of hunger and satiety rather than external or emotional signals. That is, individuals who eat intuitively are not preoccupied by food and dieting. They often choose food that helps their bodies function well and is pleasing to their palate. They do not ignore hunger cues or classify food into acceptable or unacceptable categories. Therefore, intuitive eating was found to be linked with greater unconditional self-regard and body satisfaction, as well as lower levels of both depression and disordered eating. Adolescence, in particular, acts as a critical period in the development of eating attitudes and behaviors. Adolescents devote a great deal of attention to physical appearance, and are inclined to engage in unhealthy eating behaviors to reduce an unsatisfactory body image; this has a negative impact on their physical and mental health. Due to these circumstances, it is essential to identify the key factors influencing adolescents’ healthy eating behaviors. Using the perspectives of attachment theory and the acceptance model of intuitive eating, this study aimed to investigate the effect of friendship quality on intuitive eating among Chinese adolescents, and the potential mediating role of self-compassion and positive body image on this association, as well as explore a possible gender difference. This research not only provides suggestions for parents and educators to increase adolescents’ intuitive eating, but it also identifies significant factors that influence intuitive eating in order to foster relevant practical prevention strategies and interventions.
Participants were 2438 students (M = 13.14 years, SD = 1.08) recruited from three middle schools (Grades 7 to 9) in Wuhan, Hubei Province, China. There were 1162 girls (47.7%) and 1276 boys (52.3%). They completed a set of self-report measures on friendship quality, self-compassion, positive body image, and intuitive eating. All the measures have acceptable reliability and validity.
The results indicated that after controlling for age, gender, and body mass index, friendship quality was positively associated with intuitive eating. Self-compassion and positive body image mediated this association, which contained three significant mediating pathways: the separate mediating effects of (a) self-compassion and (b) positive body image, and the serial mediating effect of (c) self-compassion and positive body image. In addition, significant differences in mediating effects per gender were only found in the relationship between friendship quality and the dimensions of intuitive eating (“unconditional permission to eat” and “eating for physical rather than emotional reasons”).
Our findings highlight the relevance of friendship quality, self-compassion, and positive body image in the understanding of adolescents’ intuitive eating. This study suggests that parents should create a warm and friendly family atmosphere which will contribute to adolescents’ peer relationships and friendships. Educators should direct students to be compassionate toward their own shortcomings and failures to improve levels of body appreciation and body satisfaction. These factors will play important roles in promoting intuitive eating. Moreover, future interventions for intuitive eating should be designed to increase adolescents’ self-compassion and positive body image through effective intervention approaches and measures.
Following the PRISMA-Protocol and using meta-analytic techniques, this study examines the effects of interventions for mental health literacy among Chinese people, with knowledge of mental health stigma and help-seeking as outcome variables. The study also explores the influences of five moderator variables (types of intervention, types of contact, types of intervention staff, types of region, and types of experiment) on the intervention effects in order to identify the active ingredients of the intervention. Another purpose of this study is to clarify the relationships among knowledge, stigma, and help-seeking in intervention, in order to provide some theoretical references for more effective improvement of mental health literacy.
Studies were identified by searching six foreign databases (PubMed, PsycARTICLES, PsycINFO, Elsevier Web of Science and Cochrane Controlled Trials Register) and three Chinese databases (CNKI, Wanfang, VIP). The search terms consisted of subject headings such as mental illness, mental health, mental disorder; outcome measures such as attitude, knowledge, help-seeking; intervention methods such as health education, health promotion, curriculum; regions such as Mainland China, Hongkong, Taiwan, Macao; and these groups of search terms were paired and combined. Selection criteria included: (1) Participants: Chinese people living in Mainland China, Hongkong, Taiwan and Macao; (2) Study design: the study provided a control group, including randomized and non-randomized trials, examining the mental health literacy at pre and post intervention; (3) Outcome measure: using knowledge, stigma, or help-seeking to measure mental health literacy; (4) Intervention: education and contact to improve mental health literacy; (5) The included studies are peer-reviewed papers and include master's and doctoral thesis. Excluding criteria were (1) participants with psychological disorders; (2) non-Chinese and non-English literature.
A total of 38 studies were included in the meta-analysis. Results: (1) The immediate effect of the intervention on knowledge, stigma and help-seeking was significant and medium-to-large (knowledge: g = 0.70, stigma: g = -0.52, help-seeking: g = 1.18). At a 6-month follow-up, only the effect of intervention on knowledge was significant (g = 0.67); at longer follow-ups (> 6 month), the effect of intervention on knowledge was still significant (g = 0.74), and the effect of intervention on stigma was significant (g = -0.34). (2) Results of the subgroup analysis: In terms of knowledge, the subgroup analysis indicated the types of region, the types of participants, the types of interaction and contact in intervention, and the types of experiment were significant. With respect to the help-seeking, only the types of region and the types of interaction and contact in intervention were significant. (3) Results of regression analysis: In terms of knowledge, the moderating effect of regions was significant after controlling other variables. In terms of stigma, the moderating effect of the types of experiment was not significant. Studies have shown that mental health literacy interventions have obvious cultural differences. Intervention effects on knowledge are stable in time, while the intervention effects on stigma are unstable. The intervention of interaction on knowledge is effective, but not conducive to stigma. The role of contact in intervention on knowledge and stigma is unclear.
Implications for future studies: (1) Improve the conceptual framework of mental health literacy. (2) Further explore the moderating variables of mental health literacy and develop individualized intervention programs. (3) Cultural factors should be considered when developing an intervention plan. (4) Explore new methods to improve mental health literacy.
The acceleration of the globalization and the occurrence of trade wars have caused people to be in an insecure state of ontology, that is, people’s ontological security is threatened. Ontological security refers to people’s confidence in the continuity of self-identity and the stability of the surrounding social and physical environment. In the marketing field, some scholars have explored how consumers cope with disruptions when their routine behaviors are disrupted. However, to date, few studies have explored the effects of ontological security threats on consumers’ brand attitudes and the underlying mechanisms. In the current study, we proposed that consumers would prefer their hometown brands when ontological security is threatened. Hometowns can provide people with a sense of security by providing routine daily life and building a personal group identity, thereby recovering their ontological security and increasing the preference for hometown brands. Thus, when people’s ontological security is threatened, they tend to increase their attachment to their hometowns. In addition, we proposed that individuals whose ontological security are threatened can be recovered from natural habitat exposure, and the differences in their hometown brand preferences would be reduced.
One pretest and three formal experiments were conducted to test our hypotheses. In the pretest, we used two scenarios (i.e., economic and cultural globalization) as stimuli to examine whether the scenarios could manipulate the participants’ threat of ontological security. In Experiment 1, we tested the effect by which the threat of ontological security influences consumers’ preferences for hometown brands through a 2 (ontological security: threatened vs. not threatened) × 2 (brand: hometown vs. non-hometown) between-subject design. In Experiment 2, we examined the robustness of the effect found in Experiment 1 and tested the mediating role of hometown attachment through a 3 (threat type: ontological security threat vs. life security threat vs. control) × 2 (brand: hometown vs. non-hometown) between-subject design. In Experiment 3, we investigated the moderating role of the natural habitat exposure by a 2 (ontological security: threatened vs. not threatened) × 3 (exposure: natural habitat vs. non-habitat vs. non-nature) between-subject design.
Results of the pretest and three experiments supported our predictions. Specifically, the results of the pretest showed that the scenarios we selected could successfully manipulate the participants’ threat of ontological security. Therefore, we used the scenarios in Experiments 1 and 3 for manipulations. Results of Experiment 1 indicated that when ontological security was threatened, consumers were inclined to choose their hometown brand (the willingness to visit the hometown tourist destination was higher). In Experiment 2, we found that the influence of the threat of ontological security on consumers’ preference for hometown brands was replicated in another product category (storage box). This effect was found to be mediated by hometown attachment. We also distinguished between ontological and life securities in Experiment 2. Finally, in Experiment 3, results showed that the influence of the threat of ontological security on consumers’ preference for hometown brands was moderated by natural habitat exposure. When ontological security is threatened, in comparison with non-habitat and non-nature exposure, natural habitat exposure helped people build emotional connections with nature, which could be utilized as a resource to cope with ontological security threats. Thus, people no longer need to seek a sense of security from home attachments. Therefore, no significant difference was observed in the preferences between hometown and non-hometown brands.
This study provides some theoretical and practical contributions. First, this study enriches the literature on ontological security by exploring the relationship among ontological security, consumer brand preferences, and decision making. Second, this study expands the literature on brand preferences by focusing on the psychological factors that influence the preferences of hometown brands. Third, we indicate the mediating role of hometown attachment and the moderating role of natural habitat exposure. Moreover, our findings have important practical implications. When the target consumers are individuals whose ontological security is threatened (e.g., immigrants and the elderly), marketers can emphasize the hometown information of the brand, leading to a higher brand preference. Non-hometown brands can use natural habitat contextual cues (e.g., displaying natural habitat-related products and brand images in advertisements) to help consumers build a wider range of place attachments from contextual stimuli, thereby reducing their focus and dependence on hometown brands.
Problem-solving competence is defined as the capacity to engage in cognitive processing to understand and resolve problem scenarios where a solution is not obvious. Computer-based assessments usually provide an interactive environment in which students can solve a problem by choosing among a set of available actions and taking one or more steps to complete a task. All students’ actions are automatically recorded in system logs as coded and time-stamped strings. These strings are called process data. The process data have multi-level structures in which the actions are nested within a single individual and therefore they are logically interconnected. Recently, researches have focused on characterizing process data and analyzing the response strategies to solve the problem.
This study proposed an extended MMixIRT model which incorporated the multilevel structure into a mixture IRT model. It can classify latent groups at process level that have different problem solving strategies, and estimate the students’ abilities at the student level simultaneously. This model takes the accumulated response information as the specific steps at the process level and defines a more free matrix to determine the weight information used for ability estimation at the student level. Specifically, in the standard MMixIRT model, the student-level latent variables are generally obtained from the measurement results made by the process-level response variables, while students’ final responses are used to estimate their problem-solving abilities in the extended MMixIRT model.
This research applied process data recorded in one of the items (Traffic CP007Q02) of problem solving in PISA 2012. The samples were 3196 students from Canada, Hongkong-China, Shanghai-China, Singapore, and America. Based on the log file of the process record, there were 139,990 records in the final data file. It was found that (1) The model can capture different problem-solving strategies used by students at the process level, as well as provide ability estimates at the student level. (2) The model can also analyze the typical characteristics of students’ strategy in problem-solving across different countries for targeted instructional interventions.
It is concluded that the extended MMixIRT model can analyze response data at process and student levels. These analyses not only play an important role in the scoring, but also provide valuable information to psychometricians and test developers, help them to better understand what distinguishes well performing students from the ones that are not, and eventually lead to better test design.