Collaborative Research: Supporting Whole-Class Science Investigations With Spatial Simulations

Principal Investigator: 
Project Overview
Background & Purpose: 

In the “embedded phenomena” framework, whole classes participate in thematic curriculum units that include a multi-week investigation of simulated scientific phenomena that have been scaled to (co-)occupy the classroom. Classroom computer distributed around the room provide persistent “portals” into local areas of the phenomenon, allowing students to make observations of (and, where appropriate, manipulate) the simulated phenomenon, which “unfold” asynchronously with respect to the regular flow of instruction. The goal of this study is to characterize the impact of the use of this framework in four outcome areas: student learning, acquisition of skill in science practice, participation, and attitudes toward science and inquiry.


This work is set in a public elementary school in a medium-sized, diverse suburban district directly bordering a large Midwestern city.

Research Design: 

The research design for this project is comparative, and is designed to generate evidence which is descriptive (design research) and associative/correlational (quasi-experimental). This project has an intervention consisting of the use of an “embedded phenomenon” activity within two thematic curricular units on earth science (using the seismology simulation “RoomQuake,” in which the classroom becomes an earthquake field, and students read seismograms, trilaterate epicenters, and discover the fault running through the room) and population ecology (using the ecosystem simulation “WallCology,” which situates learners as keepers of a habitat for endangered species nominally occupying the walls of the classroom). The comparative condition consists of use of an alternative activity that removes the features that make the activity spatially and temporally “embedded.” We remove the locality, position-dependence of the computers, persistent representation, and real-time updates, and condense the activity into consecutive science periods. Instructional activities, time on task, and user interfaces (with exception of real-time animation) are all held constant between the two conditions. This project collects original data using diaries/journals/records kept by study subjects; assessments of learning/achievement tests; videoography observation; paper and pencil self-completion questionnaires; face-to-face structured interview-administered questionnaires; and assessments of attitudes toward science and science inquiry.

In our pilot study using embedded WalCology, and in our the first pair of interventions (embedded vs. non-embedded RoomQuake) we employed:

  • A repeated measure of student understandings of scientific phenomena. For Roomquake, concepts include: earth structure, plate movement, faults, ground waves, the use of seismographs, the use of mathematical trilateration, and the spatial, temporal, and intensity distributions of collections of earthquakes. For Wallcology, concepts include: food webs, roles, population dynamics, population estimation and graphing, interdependence of species. Both assessments include multiple choice questions with prompts for explanations.
  • Individual one-on-one interviews with learners to (a) assess skill in data collection, representation and interpretation in each unit, and (b) elicit student attitudes toward "embedding" features.
  • Student "lab notebooks" including activities, data representations, and reflective prompts.
  • Subset of the TOSRA (Test of Science Related Attitudes: Fraser, 1981)
  • Video data of children enacting data collection activities (reading seismographs, observing species), and whole-class discussions.

The main analyses for this study are the comparison of learning outcomes (performance on pre and post tests, responses in interviews, changes in science attitudes) across the embedded and non-embedded versions of each embedded phenomena activity. The effect of embeddedness on these central dependent measures will be tested using HLM to account for the context of the students. In addition, rich descriptive detail about how the embeddedness of scientific phenomena in the classroom may affect the learning process will be explored using data from lab notebooks and student journals and qualitative analysis of videos of small group discussions during data collection and full class discussions.


Only one phase of the counterbalanced design has been completed, using embedded and non-embedded versions of RoomQuake in two classrooms. It is important to note that the embedded condition is confounded with the teacher/classroom at this point in the study; subsequent counterbalancing of classrooms and conditions will ameliorate that confounding.

Significant conceptual learning occurred under both treatments F(1, 64) = 129.9, MSE = 3.54, p <.001, partial eta-squared = .67. There was a significant interaction between conceptual learning gain and embedded condition F(1, 64) = 5.63, MSE = 3.54, p <.05, partial eta-squared = .08, with greater gains in embedded class. There was also a trend (p=.19) toward greater gains for lower knowledge students in the embedded class.

No significant changes were found in science-related attitudes; strongest changes and interactions with embedding condition on “enjoyment” and “science as leisure activity” subscales. Post-test scores on conceptual tests predicted skill acquisition, but in the embedded class all students had better skill acquisition regardless of level of concept understandings. Significant effects were found for treatment (F (1, 62) = 7.33, MSE = .92, p< .01) and post-test score (a significant covariate, F (1, 62) = 5.41, MSE = .92, p< .03). Skill acquisition did not predict concept learning (but may be due to limited range of skill acquisition measure). Analysis of other data sources is in progress.

Publications & Presentations: 

Reports of work related to the project (extensions to the embedded phenomena framework, system development, pilot interventions) include:

Uphoff, B., Bhatt, D., Lopez Silva, B., Frack, M., Malcolm, P., Cain, V., and Moher, T. (2008). WallCology: Studying Ecology using a Distributed, Persistent Virtual Ecosystem in the Classroom. Paper presented at the Annual Conference of the American Educational Research Association (March 2008, New York, NY).

Moher, T., Uphoff, B., Bhatt, D., Lopez Silva, B., and Malcolm, P. (2008). WallCology: Designing Interaction Affordances for Learner Engagement in Authentic Science Inquiry. Proceedings ACM Conference on Human Factors in Computing Systems (CHI 2008) (April, 2008, Florence, Italy), 163-172.

Moher, T. (2008). Learning and participation in a persistent whole-classroom seismology simulation. Proceedings International Conference of the Learning Sciences (ICLS 2008) (June 2008, Utrecht, Netherlands), Vol. 2, 82-90.

Malcolm, P., Moher, T., Bhatt, D., Uphoff, B., Lopez Silva, B. (2008). Embodying Scientific Concepts in the Physical Space of the Classroom. Proceedings 7th International Conference on Interaction Design and Children (IDC 2008) (June 2008, Evanston, IL), 234-241.

Other Products: 

In addition to research, the project is designed to generate a “Phenomenon Server,” an Internet-based facility that will allow teachers to configure and schedule embedded phenomena for “delivery” to their classrooms.