ABSTRACT
In 1987, Glymour and coworkers published a book entitled Discovering Causal Structure: Artificial Intelligence, Philosophy of Science, and Statistical Modeling. It is well known that experiments are not always practicable (or ethical) and that even when they are, they may not answer the questions of interest. Before modern medication was introduced to Asia, generations of Chinese were saved (and perhaps in many cases poisoned) by Sen Nong’s herbal medicines. A major benefit of scientific knowledge is prediction. Since the 1950s, regression and time-series analysis have been used for economic forecasting, which is reportedly a 100–200 million dollar business. An important distinction was proposed by Dempster (1983) to contrast technical statistics and functional statistics. Glymour did the statistical community a good service in giving us a clear picture of how popular statistical causal-inferences are in nonexperimental sciences. It is interesting that Glymour cited the meetings of American Statistical Association to support his advocacy of drawing causal inferences from correlation data.
