Nebraska Academy of Sciences
Date of this Version
1973
Document Type
Article
Citation
Published in Transactions of the Nebraska Academy of Sciences, Volume 2 (1973).
Abstract
A computer program developing simple statistics (N, MIN, MAX, RANGE, MEAN, MEDN, SDEV, CVAR, and standard errors of the latter four) is available at the University of Nebraska, Lincoln.
Data are read into an array of 40 measurements (six columns; maximum 9999.9mm) for a sample maximum of 50 specimens. The "shape" of the array may be adjusted within the maximum "area" of 2000 T'-field nodes in each data-set. The data-card input field provides IDENTIFICATION (cc 1-8), CARD (an A, 13, C or Din c 9 indicates measurements 1-10, 11-20, 21-30 and 31-40 respectively within the array), SEX (M or T' in c 10), SITE (code in cc 11-12 indicates stratigraphic or geographic location), POSITION (L or R in c 13 shows the skeletal geography), AGE-CODE (e.g., "JUV" in cc 15-17 indicates an immature specimen), BONE-CODE (consecutive numbers for all mammalian skeletal elements in cc 18-20), and ten MEASUREMENTS of cc 6 each. Verification subroutines with data-in-error messages arc provided.
Output may be on cards, tape or printout. Preliminary statements (sample identification, measurement descriptions), data and statistical results are associated for each data-sequence. The program is open-ended.
[n 1969-70, I took measurements on a large number of skull and post cranial elements of Holocene bison in the collections of the University of Nebraska State Museum, which had been obtained from several High Plains sites. These measurements, described and illustrated elsewhere (Hillerud, 1972b), were recorded directly on computer coding-forms and were punched as a series of card data-decks. See Hillerud (1972a, pp 39-40) for a discussion of the repeatability and reproduceability of these measurements. The computer program described below was used to generate simple univariate statistics on these data.
Although every computing center has an equivalent simple-statistics program in its files, most programs are not specifically designed to accommodate biological data. PRϴGRAM BIϴSTAT is presented here because:
(a) its range of measurements, from 0.1 to 9999.9mm, covers a major part of the macro biological size spectrum,
(b) measurement definitions, other pertinent descriptive information and a list of the raw data are printed in proximity to the statistical printout for each data-set, facilitating its interpretation,
(c) the raw measurements can be recorded directly upon computer coding-forms, and are printed in separated columns, providing easier checking for errors in data-cards,
(d) the verification subroutine provides machine identification of mixed data-sets, helping to prevent the generation of improper statistics. If cards with errors are encountered, the subroutine stores the contents of the cards for later printout as DATA IN ERROR messages. But it allows the operator to control grouping of selected data-sets, in order to obtain a larger sample, by defining a few control-cards as "comment-cards." The verification of raw data is almost a necessary element in the generation of bio-statistics, but it is not commonly provided in standard programs.
Comments
Copyright 1973 by the author(s).