Bill Notz and Dennis Pearl

Advances in the teaching of statistics, in statistics software, in computer technology, and in multimedia technology now offer the opportunity to rethink the way in which we teach statistics. Specifically, we now have available:

* well-written texts that emphasize data analysis and statistical concepts (Introduction to the Practice of Statistics by D. S. Moore and G. P. McCabe and Statistics by Freedman, Pisani, Purves, and Adhikari are examples.)

* accessible statistics software that encourages data exploration and is easy for students to learn (Data Desk is exemplary in this regard.)

* videotape material that gives broad coverage to statistical concepts and applications (The Annenburg/CPB tapes Against All Odds)

* multimedia technology that is affordable (Apple's Quicktime CD technology is an example.)

Our goal is to combine and augment these resources to develop a unified curriculum and supporting materials for teaching introductory statistics. We will deliver:

* laboratory manuals designed as a bridge between standard texts and computer software that will guide students through experiments that generate data for analysis

* video stories that take the student "into the field" to see how data were collected and to help them understand the context of the data analysis

* "start-up" materials to help students, especially disadvantaged students, get started using computers for statistics

* a machine-readable encyclopedia of examples and problems that will include data, descriptions of experiments, and computer-based retrieval

* a computer archive of edited "bites" from existing video material keyed to specific statistical concepts

* an experimental test of the effect of early enrollment in statistics by minority students on their subsequent academic performance and attitude towards science and mathematics

When this project is complete, it will represent the first time that such a complete package of multimedia materials will be used in introductory statistics courses. The time frame for the grant will allow us to anticipate the wide-availability of inexpensive technology and have materials pretested on a broad audience ready for immediate dissemination.

In order to improve and coordinate instruction in the large, multisection lectures that we are currently teaching under the GEC at OSU, we would like to develop lab manuals and introductory tutorials for our statistics software as quickly as possible. Year 1 of the grant, therefore, will see the development of these materials.

Lab manuals would be written in book or manual form as well as on the computer in a Hypercard format. The manuals will be designed to integrate the material in the data analysis textbooks with the computer software available in the labs. For example, important concepts such as sampling distributions, the meaning of confidence intervals, and proper interpretation of alpha-level testing are often difficult to convey in a traditional manner and are not easily grasped by simply reading about them. These concepts rely on the behavior of statistics in repeated sampling and are best understood through experimentation, both physical and computer-based by means of simulations. Statistical software currently on the market allows for interactive data analysis but not interactive teaching. Instructions on how to select random samples, carry out simulations, or conduct special analyses, such as constructing a median trace, are rather complicated on standard packages and are not necessarily provided in the manuals that come with the software. Lab manuals providing clear instructions and examples need to be prepared. Manuals will also contain exercises and examples that will help illustrate and reinforce the concepts for the students. Two sets of manuals, one developed by Professors Notz and Velleman which emphasizes statistical methods and working with data following Introduction to the Practice of Statistics by Moore and McCabe, and one developed by Professors Pearl and Stasny which emphasizes experimentation and statistical concepts following Statistics by Freedman, Pisani, and Purves, will be written. We hope to prepare the manuals in such a way that it is possible to use them, with relatively little modification, in courses based on other data analysis texts.

To make it as easy as possible for students to learn to use the Data Desk software package with a minimum recurring input of human resources, we shall prepare user-specific courseware which will take students through a Tour of Data Desk (opening the program, creating variables, entering and editing data, creating a few simple graphics, and obtaining a printed copy of their results). This will be developed by Prof. Velleman. Throughout the period of the grant Prof. Velleman will also be creating new features for Data Desk which will enhance its use in the classroom. These developments, however, will not be supported by the funds requested in this proposal.

Additionally, we plan to develop a computer-based encyclopedia of interesting data sets and examples, including misuses or deceptive uses of statistical arguments, from a wide variety of disciplines to supplement those in the textbooks. Everyday examples can be selected from USA Today, Newsweek, Time, and other papers and magazines. Several sources for materials from many disciplines exist, for example, The New England Journal of Medicine, Science, and Lancet. Statistical journals such as The American Statistician, Biometrics, Technometrics, Journal of the American Statistical Association, Journal of Educational Statistics, Journal of Quality Technology, and Journal of Business and Economic Statistics are all sources of real data sets and illustrative examples. Faculty and textbook authors at other universities have developed data bases which they may be willing to make available to us. In addition, through OSU Statistical Consulting Service, which has been in operation for many years, we have available a large number of real data sets.

Data sets for the encyclopedia must be organized and edited for teaching purposes and stored in both ASCII and Data Desk formats. Working groups in the Section on Statistical Computing of the American Statistical Association and a group hosted by Apple Computer have been working to define a portable file format for data. If such a format becomes available, we will use it so that the data can have maximum portability to a variety of software products on a variety of hardware platforms. Professor Velleman is working with both these groups and will be responsible for keeping the project up-to-date on developments in this area. Documentation explaining the background of the data set, what variables were measured, and the objective of the original study must be included. Suggestions for how to use these data sets in class also need to be prepared. Finally, data sets will need to be catalogued according to the subject area represented and an access program developed. The development of such an extensive data base will require human resources (release time for all the PIs in this proposal, support for two statistics graduate students, and a computer specialist), a Macintosh computer, and a large capacity storage media. Our goal is to complete the core of the encyclopedia in the first two years of the grant, with additions and modifications continuing through years 1-3.

This task will be supervised by Professor Notz who has served as Co- Director of the Statistical Consulting Service and has been recognized for his excellence in the classroom by being awarded an Alumni Distinguished Teaching Award at OSU. This prestigious award is given annually to only eight individuals from a faculty size of 4500. In addition he has twice received the Thomas E. and Jean D. Powers Award for Excellence in the Teaching of Statistics.

The immediate impact of such courses on the quality of undergraduate education at the Ohio State University (OSU) will be significant since there are approximately 30,000 undergraduate students who will be required to take statistics under the newly approved General Education Curriculum (GEC). At Cornell University (CU) these materials will be used in their largest undergraduate introductory statistics sequence. CU has a history of pioneering efforts for innovative statistics teaching in this sequence. CU was one of the first universities to use computers in an introductory statistics course. This course provided the environment for which Data Desk was originally developed and in which preliminary versions of the program design were tested. As prominent institutions, successful efforts at OSU and CU are likely to be noticed and emulated at colleges and universities nationwide. For example, at OSU we have already had formal inquiries about the content of our GEC courses from Colby College, Michigan State University, the University of California (Davis), Virginia Commonwealth University, and Western Kentucky University. Dissemination of materials developed under this grant will improve undergraduate teaching of statistics, make effective and integrated use of new materials and technology being developed, increase enthusiasm for statistics among students, and make statistics more enjoyable to teach nationwide.

One question that has arisen along with the development of the GEC at OSU is when students ought to take a data analysis course. An early start may eliminate some computing and math anxiety and better prepare the student for further courses in the sciences. Alternatively, students at this stage may lack the motivation to study a conceptually difficult subject and may lack an understanding of the relevance of the material to their lives. In this later case, taking the course early would be a waste of time. The materials we will develop should eliminate these difficulties by clarifying concepts and demonstrating relevance.

To verify this, along with developing the materials for teaching introductory statistics courses, we plan to conduct a designed experiment in which some minority students and a control group of similar students will be followed over four years to determine if an early start in statistics improves students' academic performance and attitude towards the sciences. If the early start hypothesis is supported by our data, this may provide a simple means for improving the academic experience of minorities, as well as generating interest in science and mathematics.

Return to Cobb Paper | Return to Table of Contents | Return to the JSE Home Page