Projects

Task Switching & Executive Functions

Traditionally, a switch cost is defined as an increased response time after switching between two tasks compared to repeating the first task. Asymmetrical switch costs occur when the cost of switching is smaller for the more complex of two tasks. We have found that young children exhibit asymmetrical switch costs at both general and specific levels, i.e., block and trial statistics, respectively. However, the patterns of these asymmetries are remarkably different across task domains. To date, we have identified unique developmental trajectories for three domains: figure matching, arithmetic, and reading. The results from these projects indicated that the developmental patterns of asymmetrical switch costs vary greatly depending on the domain of the tasks that are being switched. It seems that development, experience, and knowledge acquisition drive these age-related changes in asymmetrical switch costs.

The developmental trajectories that we observe in switch costs may reflect developmental changes in executive functions. We have been identifying how the development of executive functions, experience, and knowledge acquisition interact with each other to drive these developmental changes. Arguably, efficient switching relies on two executive function processes: inhibition and switching. Using various modeling techniques like structural equation modeling, we have identified the relative contributions of inhibition and switching across development for switch costs. Our results indicate switching and inhibition contribute differentially to these dvelopmental changes. Moreover, the ability to switch between tasks is important for producing a correct response and the ability to inhibit is essential for a quick accurate response. For more details see: Family Thinking Skills and Mind Match Chess and Think-Art! 

Causal Reasoning of Scientific Phenomena

We have been investigating how the development of causal reasoning might interact with learning scientific phenomena and whether these results can lead to similar educational interventions. Grasping scientific phenomena includes more than the mere memorization of scientific facts. It requires the creation of an appropriate model of the important aspects of the phenomena and how these aspects are related. Many, but not all, of these relations are causal. A growing number of studies suggest that children (and adults) have remarkable causal reasoning and learning abilities in laboratory settings, claiming that they can reason like intuitive scientists. Unfortunately, abilities demonstrated in laboratory studies do not seem to transfer to achievement in the science classroom. we have been investigating how causal learning and reasoning influences learning and understanding in the science classroom.

One major difference between the laboratory and the classroom is the complexity of the causal structures. Causal structures in the science classroom contain many more potential causes that interact with each other in complex ways. Young children seem particularly sensitive to the frequency of the occurrence of the causal events associated with more than one potential cause (Ellefson, Johnstone, & Chater, 2005). This frequency can bias their ability to create accurate causal structures of events, especially when this bias is accompanied by inconsistent outcomes. There are many examples of inconsistency of results in the science classroom and inappropriate use of the frequency of events that may hinder scientific reasoning in educational settings. The nature and complexity of the scientific materials themselves may hinder scientific understanding. In collaboration with Professor Nick Chater, Dr. Anne Schlottmann, and Dr. David Sobel, we have developed a series of studies to investigate how students’ naïve views of physical and biological sciences constrain their understanding of causal structures across scientific domains. Furthermore, in work with Dr. Kelly Goedert, we are investigating using studies of expertise in these areas to document a trajectory of causal reasoning across levels of scientific knowledge to inform better practice in science education.

We have been investigating the nature of students’ causal models in high school science with Dr. Christian Schunn. The results indicate that the complexity of student models of scientific phenomena is related to their knowledge of the domain. Students who have more knowledge of the domain produce causal structures that are more complex, include multiple potential causes, and multiple effects. However, the causal relations that they describe are simple and rarely include interactions. A separate study indicates that students use reasoning structures that are too simple to effectively understand these interactions, especially under uncertain conditions (Ellefson & Young, 2006). Science experiences that require students to think about and investigate complex interactions may provide opportunities for students learn science more effectively. We have been developing and implementing such experiences for high school chemistry and biology classrooms. For example, through the process of genetic engineering and manipulating the environment to turn genes on or off, biology students are learning about the complex interactions between the genes and environments while learning about the process of how DNA becomes a trait. For more details see: Children's Chemistry Reasoning 

Design-Based Learning for Science Education

In collaboration with our colleagues at the University of Pittsburgh's LRDC, we have been developing high school science curricula in chemistry and biology that incorporate design-based learning. This educational approach incorporates fundamental principles of engineering design to the science classroom. The practice of engaging in an engineering design process allows students to investigate scientific phenomena more meaningfully than traditional classroom practice. We plan to enhance our current research program by utilizing electrophysiological measures to understand further how educational instruction and cognitive development influence the learning of complex information.

Literacy & Language Development

In collaboration with Rebecca Treiman and Brett Kessler, we have been investigating the influence of educational practices in literacy on cognitive development. Learning about letters is an important foundation for literacy development. Should children be taught to label letters by conventional names such as /bi/ for b or sounds such as /b/? We studied parents and teachers in the U.S. and England, finding that the former stress letter names with young children whereas the latter begin with sounds. Looking at 5- to 7-year-old children in the two countries, we found that U.S. children produced more correct responses in letter name tasks than English children. English children outperformed U.S. children on letter sound tasks, and the country differences declined with age. We found that children use the first-learned set of labels to inform the learning of the second set. English and U.S. children thus made different types of errors in letter name and sound tasks. The children’s invented spellings also differed in some ways, reflecting the labels they used for letters.

In collaboration with Dr. Janet Vousden and Professor Nick Chater, we have been investigating the application of the Simplicity Principle to early literacy education. The basic premise of the Simplicty Principle is that there is a cognitive advantage to learning under simpler rather than complex forms of instruction. Using computational models, we are exploring the utility of different literacy educational approaches (i.e., whole words, onset-rimes, graphemes, etc.) as appropriate decoding schemes for young readers of English texts. Our early results suggest that the simpler form of instruction (i.e., graphemes) have a higher utility than more complex forms (i.e., whole words or onset-rimes).

Learning Complex Hierarchical Structure

We are investigating how complex information should be presented to enhance learnability. The difficulty in learning complex academic subjects like mathematics and science might not necessarily be dependent on the actual content, but rather the organization of the to-be-learned information. Hierarchical problem solving is required in a number of content areas. In order to understand specific properties, students must break a chemical into its molecules, a sentence into its phrases, or an object’s trajectory into the various forces that will determine its movement. The surface structures of mathematical, linguistic, and scientific materials are complex and may interfere with students’ ability to understand them. We have explored how middle school and university students learn hierarchical structure using a sequential learning task (Ellefson, Young, Christiansen & Espy, 2004). In previous research (Ellefson & Christiansen, 2000-b), we have found that the sequential learning task successfully assesses the relative difficulty of complex linguistic structures. In more recent projects, we have been investigating how the use of cues and presentation order facilitates the implicit acquisition of the underlying structure when students remembered a series of letter strings, constructed using structurally complex rules. More recently, in collaboration with Dr. Morten Christiansen, Dr. Fenna Poletiek, and Dr. Christopher Conway, we have begun to investigate how sequencing of increasingly complex information might be critical to improving learnability.