| Summary: | "Practically all students in a course in statistics keep a notebook, but even a casual inspection reveals that often undue emphasis is given to a trivial unit, while a major topic receives superficial treatment. This outline has developed slowly over a period of time in an attempt to meet this difficulty and to develop in the student a functional and integrated knowledge of the subject. The present material is the result of a series of revisions of earlier forms used in the writer's classes and in those of Dr. H. E. Garrett at Columbia University and Dr. E. E. Cureton at Alabama Polytechnic Institute. A student should have some experience in computing the various statistics (1) in order to understand better the assumptions and to know when and where they may be justifiably used, (2) so that he can compute the measures for his own use, (3) because actual manipulation of the data helps him to remember these factors, and finally, (4) to gain an estimate of the time and cost of statistical work. Occasionally a graduate student who has had a course in statistics "to develop the ability to read and understand psychological and educational literature" comes with a problem involving forty variables and proposes to work out a multiple regression equation with a criterion and thirty-nine independent factors, or some other equally impractical problem. In order to reduce computation to a minimum, a single set of basic data is presented. Computations of the various formulas are based on these data, so that the results are cumulative; that is, the frequency distributions for the means are used for the medians, standard deviations for checking correlation plots, and so forth. A fundamental obligation of the instructor is to provide the student, as a matter of course, with the simplest forms of computation and whenever possible with check methods. Training should also be given in the planning of an experiment and the selection of the statistics to be employed before a single measurement is made. The next and perhaps the most vital step in the training of students is the interpretation of the statistics after they are computed. In a course, time does not permit one to carry out an actual experiment. This difficulty has been partially remedied by selecting from the literature experiments that emphasize a particular technique. These are presented to the student with a brief abstract and the basic table or tables. The student is then required to draw the proper conclusions. In order that he may get training in expressing the results clearly and concisely, the student is requested to prepare a brief digest suitable for publication in a technical journal. Then the complete reference is given, and he is requested to verify his conclusions by checking with the author. Note that these problems increase in complexity as the student's knowledge of statistics and his experience in preparing such reports increase. Another device to get the student to think through the material presented in class or in the text used is the completion sentences that appear throughout the work. Further, at the end of each section true-false questions are presented that require a minimum of computation, and that are, for the most part 'thought' rather than factual questions"--Preface. (PsycINFO Database Record (c) 2010 APA, all rights reserved).
|
|---|