|Abstract or Summary
- The problem studied was to determine if there were any significant
differences (5% level) between control and experimental groups
of students using rapid batch processing and time-sharing computer
facilities in the study of elementary Fortran programming.
Students were randomly assigned to the control (batch processing)
and experimental (time-sharing) groups with all students receiving
the same in-class instruction in Fortran.
Students were measured on four items for each problem solution:
number of minutes to prepare the first run; number of runs required
to obtain a correct run; number of syntax errors per run; and number
of logic errors per run, These measures were reported on student
Following the completion of the solution of three problems, students completed an in-class problem of similar nature which the
instructor submitted to obtain syntax and logic error counts.
Students were then required to use the alternative mode at least
once during the next six weeks of problem solution and tested again in
class using a similar type of problem. Counts of syntax and logic
errors were determined by the instructor as in the previous test,
The study was performed at University of Nevada, Las Vegas,
a largely commuter university of about 6, 000 students, in the Fall
semester 1971 and replicated in the Spring semester.
Facilities used exhibited a high degree of comparability so that
the study might avoid confounding variables. Specifically, a Sigma 7
Xerox Data Systems computer (64K), and the BTM (Batch Time-
Sharing Monitor) system, located on the Reno campus and connected
to the Las Vegas campus by telephone was the facility in use.
Data were tested for mean differences using Mann-Whitney U
(for small groups), Student's t, Analysis of Variance F, and Analysis
of Covariance F, using grade point average and IBM Aptitude Test for
Programming Personnel as covariates.
No differences significant at the 5% level were obtained.
The failure to reject any of the six hypotheses, designed to
detect differences between control (batch processing) and experimental
(time-sharing) groups on measures of time required to prepare the
initial run, number of runs required for successful problem completion, and numbers of syntax and logic errors detected per run,
argues that there is no difference in utilizing time-sharing or fast
batch processing in the teaching of elementary Fortran, when limited
to the kinds of numerical problems utilized to illustrate uses of
particular language features, provided the facilities used are
comparable to the degree maintained in this study.