Marginal Science Strategies

Although multiple regression has become common in the socieal sciences, the author suggests that social sciences have become suspect in terms of the use of it. In a previous article, the author suggest that the social science may lose it’s standing because, not a lot of studies have ratio data and ratio analysis.

Not all data collection indicate that there is a “O” baseline, equidistance between integer levels, and related. Indeed some data analysis, one may find that ordinal and interval numbers are not that far apart.

However, the author is suggesting that the fields (political science, psychology, sociology, and economics (humans are non- rational) are less amenable to multi-variate analysis. Thus, the name and tradition will probably grow, something must be given up.

The social sciences are not hard science. On the other hand the field may change the name from social sciences to behavioral methodologies. Not only may it win the hearts and minds of chemists, physics but others. Many studies may include both hard and behavorial together.

A physics finding may be ranked compared to others of the same area. The context and measures may be not be flexible for the natural sciences. Why was the physics study used in the first place? Is there any way the two (science and behavorial) be mixed together? Yes, it is called Chi-Squared.

There are other issues that must be addressed. One must be especially uncertain when the study is analyzed by meta-analysis. Rather than give a lengthy diquistion , look for critisicms of meta-analysis on the internet.

Due to financial or other stimuli, can a researcher expand the margin of error without increasing the sample size? Well it may be tried. The sciences may want to look the other way.

What about convenience samples? Journals may be filled with them, but will an even large number increase the accuracies? Small purposive and explorative samples may be used because the area studies are so hard to find respondents.

Please see the criticisms of dummy variables. What do we find? Isn’t it really a nominal variable rather than a ratio one? So what is a happy ending? A caveat may be placed at the end of study that indicates that all the above have been used, because the study is exploratory. It is also followed by the name of a “working paper.” The reviewers send the copy for a second analysis by a senior researcher who is given time and effort to do a second analysis and the person receives credit.