You are here:

Improved attribution recall from diversification of environmental context during computer-based instruction

, Columbia University, United States

Columbia University . Awarded


There is much evidence that spatial context is important to human information processing. For instance, memory for information is improved if it is learned in several rooms or from multiple sources, than in a single room or from a single source. Spatial context can be expected to play an important role in computer-based instruction and distance education using PC terminals, because these types of computer-mediated instruction often occur in a restricted environment such as one's home or office. These settings, while convenient, are often isolated and lack diverse spatial cues. Fundamental research studies suggest this should result in negative effects for memory.

This study investigated whether diversifying the setting in which computers are used as an instructional tool would address this problem by creating more cues for retrieval of information. Participants in the study (n = 71) read three computer-based instructional units, each describing a different city. These city units were presented on a computer in a single site or in three diverse sites. In the diversified condition, participants worked with different city texts in different sites. The three learning sites were made more visually distinct from each other on several dimensions including decorative arrangements. All participants were tested for attribution recall at a new neutral site.

The results confirmed the prediction that diversification of computerized learning site improves recall, i.e., participants in the diversified learning sites performed better on the attribution memory test than students in a single learning site. The main source of error was due to substitution error among the three city names. Furthermore, the type of attribute (unique to each city or shared among all cities) was found to interact with the treatment group (single or diverse). Individual differences in sensitivity to background cues, as measured by Group Embedded Figures Test, was not found to interact with test performance. The results were explained in terms of the diversified context elements providing extra cues for increased discrimination among cities or providing alternate memory paths for retrieving the attributive information.


Yi, J.C. Improved attribution recall from diversification of environmental context during computer-based instruction. Ph.D. thesis, Columbia University. Retrieved May 20, 2019 from .

This record was imported from ProQuest on October 23, 2013. [Original Record]

Citation reproduced with permission of ProQuest LLC.

For copies of dissertations and theses: (800) 521-0600/(734) 761-4700 or