The Advantages of Microcomputers for On-site Collection and Analysis of Behavioral Data (1987)
©1987, 2013 by Dallas Denny
Source: Denny, Dallas. (1987). The advantages of microcomputers for on-site collection and analysis of behavioral data. Paper for Department of Special Education, George Peabody College of Special Education.
The Advantages of Microcomputers for On-site Collection and Analysis of Behavioral Data
Dallas Denny
11 March, 1987
Submitted to
Dr. Floyd Dennis
Dr. James Fox and
Dr. Richard Porter
in Partial Fulfillment of the Requirements for the Doctor of Education Degree
George Peabody College of Vanderbilt University
Abstract
Abstract
Portable microcomputers share most of the advantages and disadvantages of dedicated electronic data collection devices (event counters). However, because portable computers are programmable, they have a number of advantages which event counters do not have. For example, a program can be created to fit the specific needs of an observational situation, and modified as the need arises. Additionally, portable computers can be used for on-site analysis of data after it is collected, to compute inter-observer agreement, and even to generate a report.of findings. As researchers come to value continuous as opposed to discontinuous collection of behavioral data, the economy and programmability of portable computers should lead to their wide-spread use in field settings.
Within observational sessions, discontinuous collection of data (time sampling): introduces sampling error (Johnston & Pennypacker, 1980; Martin & Bateson, 1986; .Sackett, 1978). Johnston and Pennypacker have noted, “Such sampling error is a major contributor to measure-induced variability, which then becomes thoroughly confused with treatment-imposed variability, severely compromising the accuracy and generality of experimental interpretations″ (1980, p,149). Although it is considered methodologically superior to discontinuous collection, continuous collection of data makes greater demands on observer energy and concentration, and may require expensive and cumbersome equipment (Sackett, 1978).
Portable electronic devices for the collection of data (event recorders) have been available since the mid 1970’s. Their introduction freed researchers from dependency upon large computers and mechanical tabulating devices, and made continuous collection of data in field settings more feasible (Martin & Bateson, 1986).
Bakeman and Gottman (1986) and Martin and Bateman (1986) have discussed the pros and cons of using electronic event recorders. Both sets of authors agree that event recorders should not be used when paper-and-pencil methods would work just as well. “…. there is a satisfying physicality about pencil marks on paper. The whole record can be seen easily and parts of it modified with nothing more complicated than an eraser″ (Bakeman & Gottman, 1986, p. 66). “There is no point in using a complex and expensive. piece of machinery when paper-and-pencil would do just as well” (Martin & Bateson, 1986, p. 78). Bakeman & Gottman (1986) note there is a start up cost associated with the use of event recorders; an investment, of both money and time is needed to purchase and learn to use the equipment. Martin & Bateson (1986) note that event recorders can serve to distance the observer and the data; for instance, it might be necessary to transfer data to a desktop computer-before it can be visually inspected. They also comment that the ease of collecting data with event recorders can lead to collection of trivial or irrelevant data; that a malfunctioning event recorder can lead to permanent loss of data; and that event recorders may lack the versatility t0 meet the needs of every researcher.
The primary advantage of using electronic event recorders. is that they tend to ensure more accurate recording of data. The trained observer can pay less attention to the mechanics of recording data and more attention to the subject (Bakeman & Gottman, 1986). Also, the observer can record more rapid streams of behavior and observe more categories of behavior than is possible with pencil-and-paper (Martin & Bateson, 1986). Because of the built-in clock in the recorder, the duration of behaviors is recorded with more precision than with paper-and-pencil methods (Martin & Bateson, 1986). And finally, because time-consuming and laborious transcription of data into machine-readable form for analysis isn’t necessary transcription errors are eliminated (Bakeman & Gottman, 1986; Martin & Bateson, 1986).
Two machines which are representative of state-of-the-art electronic event recorders are the OS-3 by Observational Systems, Inc. and the Datamyte by Datageneral, Inc. These handheld units are battery-operated and microprocessor-based. Each has a keypad and a small LCD viewing screen. Both have a large amount of random access memory for storage of data. The OS-3 and Datamyte are operated in a similar manner. As events are observed, they are recorded by presses of the keys. An internal clock times each key press and preserves the time and the code number of the event in memory. After the observational session is over, data can be retrieved and displayed on the viewing screen or transferred to a desktop computer for analysis.
Despite the advantages they provide, event recorders are dedicated data collection devices, and thus have a number of limitations which microcomputers do not have. Because their operating system is hard-wired, the ability to revise data-collection strategies on site is limited. Major and even minor revisions of the operating system are generally not possible. The units are expensive and are not readily available in most cities. If a unit fails and a replacement is not at hand, here may be a considerable wait while a new unit is sent from the factory.
Although portable microcomputers share the same disadvantages as electronic event recorders, they offer most of the advantages of event recorder. Additionally, they provide a number of advantages of their own. They are no more expensive than event recorders, and some models are considerably less expensive. They are readily available in most cities and towns. The viewing screens are generally larger than those of event recorders. Portable computers offer telecommunication abilities. Most importantly, portable computers include a programming language in firmware, or else programming languages can be easily loaded into memory. This allows on-site revision of software and ensures transportability of software across machines.
Because of the ability of portable computers to be reprogrammed, new developments in the technology of data collection and analysis can be incorporated in their software. With proper programming, their versatility is virtually unlimited. They can be programmed to allow on-site analysis of data, or to check inter-observer reliability. They can be interfaced with mechanical counters, freeing the human observer for other tasks. In the field, they can serve to augment or replace memo pads, typewriters, word processors, electronic calculators, alarm clocks, timers, file cards, appointment books, and address books.
As they have become more sophisticated and powerful, portable computers: have approached desktop computers in their ability to process large amounts of data. At the same time, their size and price have dwindled. As more sophisticated software for these machines emerges, and as more researchers come to realize the advantages of continuous as opposed to discontinuous collection of data, portable computers should come into common use in observational settings, and may even largely supplant event recorders.
References
Bakeman, R.,& Gottman, J.M. (1986). Observing interaction: introduction to sequential analysis. New York: Cambridge University Press.
Deni, P., Szijarto, K., Eisler, A., & Fantauzzo, C. (1983). BASIC programs for observational research using the TRS-80 Model 100 portable and Model 4 computers. Behavior Research Methods & Instrumentation, (6), 616-617.
Denny, D., Fox, J.,, & McEvoy, M. (1986). Using the TRS-80 Model 100 portable computer as a data collection device. Presented at the 11th annual meeting of the Association for Behavior Analysis, Columbus Ohio, May 24-27.
Denny, D., & Fox, J. (1987) Collecting and analyzing sequential-data with a portable computer. Twentieth AnnuaI Gatlinburg Conference on Research and Theory in Mental Retardation and Developmental Deficiencies, Gatlinburg, TN, March 25-28.
Johnston, J. M., & Pemnypacker, H.S. (1980). Strategies and tactics of human behavioral research. Hillsdale, NJ: Lawrence Eribum Associates.
Martin, P., & Bateson, P. (1986). Measuring behaviour. New York: Cambridge University Press.
Ormsby, D.E., Deitz, D.E.D., & Cole, K.B. (1986). A microcomputer system for recording and analyzing social interactions in the natural environment. Nineteenth Annual Gatlinburg Conference on Research and Theory in Mental Retardation and Developmental Deficiencies, Gatlinburg, TN, March 13-15.
Sackett, G.P. (1978). Measurement in observational research. In Sackett, G.P. (Ed.) Observing behavior, Volume II: Data collection and analysis methods. Baltimore, MD: University Park Press.