When economic research has potentially important implications for public policy it behoves researchers to take as much care as possible in their treatment of data and in their statistical analyses. No empirical econometric study can be without its limitations and often those limitations may be hidden in the detail of quite sophisticated econometric technique. By contrast, some work uses only relatively simple statistical analysis and, in that sense, appears to be more comprehensible and more persuasive to the non-specialist reader.
Chapple (2000a) has contributed to the "Closing the Gaps" public policy debate with a paper that questions whether gaps exist, and whether, if they exist, they are growing. We see three major problems with this paper. First, the evidence presented is very poorly documented. Sources of information are not quoted (effectively hearsay is used) and data are not freely available. Second, the paper ignores several well-known caveats concerning the validity of inferences from statistical analyses. Third, even if one accepts the results as tabulated and plotted in the paper, it is quite possible to draw radically different substantive conclusions as to their implications than those drawn by the author.
