TY - JOUR
AU - Kurtz, Thomas E.
AB - A Role of Time-sharing Computing in Statistical Research THOMAS E. KURTZ Dartmouth College The importance of computing in statistical practice results correctly show that, for the case of comparing is obvious and well-known. Numerous computer pro- two binomials (one set of marginal totals fixed), the grams (not all of high quality [3]) for performing linear well-known Yates correction for continuity is overly regressions, factor analysis, cross-tabulations, analysis conservative with respect to significance level. Similar of variance, etc., are in daily use on practically all results were also shown to hold in most cases when scientific computing systems in this country, if not in neither set of marginal totals is fixed. The purpose of the the world. Less frequent though still important is the present article is to show how these or similar calcula- use of computing in statistical research. For instance, tions can be easily-, quicklj-, and cheaply performed in a tables of critical values are computed, and simulation Time-sharing environment. First of all, it occurred to studies made of the properties of new statistical tech- me that the case of comparing two binomials could be niques. studied with direct (up to round off) calculations of the Not as 
TI - A Role of Time-Sharing Computing in Statistical Research
JF - The American Statistician
DO - 10.1080/00031305.1968.10480512
DA - 1968-12-01
UR - https://www.deepdyve.com/lp/taylor-francis/a-role-of-time-sharing-computing-in-statistical-research-TvHv2vzJ4r
SP - 19
EP - 21
VL - 22
IS - 5
DP - DeepDyve
ER -