(41C) Probability and Statistics

[SlideRule]

(03-19-2019 01:19 PM)rprosperi Wrote:  Great contribution Kim, thanks for sharing this!
… (a)
If anyone has any comments on the author's books (Peter Zehna) from about 1982, please feel free to share them.
… (b)
Thanks!

(a) you're welcome!

(b) review from The American Statistician (FEB 1986) Vol. 40, No. 1 pgs. 45-46

Probability by Calculator: Solving Probability
Problems With the Programmable Calculator.
Peter W. Zehna. New York: Prentice-Hall, 1982.
x + 182 pp. $9.95.
Statistics by Calculator: Solving Statistics Problems
With the Programmable Calculator.
Peter W. Zehna and Donald R. Barr. New York:
Prentice-Hall, 1982. xi + 308 pp. $13.95.
"When I received these books to review, I was skeptical of the value of a programmable calculator for classroom use. My experience in statistical computing has been with mainframes and personal computers. If the programmable calculator is to gain wide acceptance in our field, however, it must be reasonably accessible to the novice. I therefore decided to borrow a TI-59 calculator and work through the books. This review is an account of this effort, with comments and suggestions.
The preface (it is virtually the same for both books) indicates that the books are based on Zehna's experiences using programmable calculators in his classroom at the Naval Postgraduate School. The calculator of choice is Texas Instruments' TI-59 programmable calculator. All programs are written specifically for the TI-59. This points up a potential problem. Should the calculator be discontinued or supplanted (Texas Instruments has come out with the TI-66 programmable), the programming aspect of the books under review loses much of its value.
The first book, Probability by Calculator, consists of five chapters - About the Book, Probability, Discrete Random Variables, Continuous Random Variables, and Bivariate Distributions - and an appendix containing the programs used for solving the problems in the book.
Chapter 1 is very informative. For instance, readers are told that they should have some background in calculus. Another Prentice-Hall book, Calculus by Calculator by Maurice Weir, is recommended. In addition to the TI-59 programmable calculator, the reader should have the programming manual entitled Personal Programming (PPM) and the master module, with its manual entitled Master Library (MLM). These and other peripherals are included when the TI-59 is purchased (and sometimes when it is borrowed). Before proceeding, readers are asked to acquaint themselves with these manuals. The explicit keystroke instructions given in PPM and MLM are very helpful. It is disappointing to learn that very few keystroke instructions appear in Zehna's books. A check of the appendix reveals programs listed in a printed code.
Section 1.4 is included to help translate the printed code into keystrokes. (I thought it would be more convenient to have the explicit keystrokes, e.g., INV, SBR, rather than the printed code RTN with the explanation of what this code means in Sec. 1.4.) Translation errors (e.g., the code RC* on p. 6 is translated incorrectly) are especially disconcerting. (I had read only through p. 6 and was having serious doubts about whether or not I would make it through the book.)
Chapter 2 is a welcome relief. The material dealing with probability is very well written. In clear and concise terms, the reader is led from the axioms of probability through Bayes's theorem. Although the author may intend this book to be a supplement to a text, it stands well on its own. Reinforced by this clear exposition, the reader is ready to enter the first program into the calculator by using the printed code provided by the author. (It took me five tries to get it right.) Flipping back and forth between the appendix and Section 1.4 is bothersome. In addition, some printed codes are not explained completely. Readers will undoubtedly have a feeling of accomplishment (as I did) if and when the program works. Until that time, entering the program can be a very frustrating experience.
The subsequent chapters follow the pattern of the second. The discussions of discrete, continuous, and bivariate distributions are very well written. The programming can be frustrating. Once the program works, the calculator is valuable for repeatedly solving problems of a particular type. There is a cost involved. In addition to being comfortable with the probability material, the reader must be familiar with the execution of the program. Function keys (e.g., x) are immediately recognizable. The programs are only accessible through the instructions provided at the end of each chapter. The user's task is thus twofold. First, the correct procedure for a given problem must be identified. Then the calculator program must be accessed properly and run. Since the book does not attempt to explain programming, the most that readers can expect to gain is a black box with instructions for its use. The instructions change from program to program. Statistical packages (e.g., SAS and SPSSX) have a general pattern. Once this pattern is understood, the user can move freely among programs. This advantage is not
shared by the TI-59 programs.
The second book, Statistics by Calculator, is similar to Probability by Calculator. It requires the use of a TI-59 equipped with the Applied Statistics Module. The book consists of 10 chapters - About the Book, Sampling, Data Processing, Estimation, Hypothesis Testing, Bivariate Populations, Proportions, Analysis of Variance, Simple Linear Regression, and Multiple Regression - and an appendix with the programs for the book.
The first chapter is almost identical to that of Probability by Calculator. The programs are given in printed code, so the problems of translation still exist. The programs in the book depend on the built-in programs of the Applied Statistics Module. The manual that accompanies this module is required reading and is referenced frequently by the authors. The Applied Statistics Module provides built-in programs that often supplant those provided in the first book.
The strength of this book, as with the first, lies in the explanation of the statistical methods. Special care is taken in pointing out when a procedure is appropriate and when it is inappropriate. Even difficult concepts in hypothesis testing, paired versus independent samples, and u known or unknown are explained well. Again, the printed code format of listing programs is inconvenient.
By comparison, Sourcebook for Programmable Calculators (Texas Instruments Inc. 1979) provides very easy to follow keystroke instructions for methods in probability and statistics. This book is seriously lacking, however, in explanations of the methods.
In conclusion, I come away from this reviewing task with mixed feelings. I am impressed with the treatment of the concepts and methods of probability and statistics. I am disappointed with the choice of a printed code instead of a keystroke listing of the programs. I find some of the programs useful, others overly complex for the problem, and still others redundant because of the preprogrammed Applied Statistics Module.
There may indeed be a place for the programmable calculator in the classroom. Given the ever-changing state of computing power it may not be long before handheld calculators have builtin functions to perform all of the procedures described in these books".
W. R. Stephenson
Dept. of Statistics
Iowa State University
Ames, IA 50011

BEST!
SlideRule