The authors have over the years tried to revise meta-analysis because it's basic premise is to add apples and oranges together and analyze. In other words, various data on the same subject are chosen using different samples, research strategies, and number properties. The findings are then homogenized and a statistical analysis is used (Snell, J. and Marsh, M. 2003)
Our first article, we tried to reduce all studies to a nominal level. Thus the variety of data analysis would be similar and although some robust attributes in the statistical analysis were lost, we tried to add some caution, reliability, and validity. This reduction of heterogeneity hopefully might improve meta-analysis (Snell, J. and Marsh, M. 2003)
Years later, we discovered that meta-analysis had become a cottage industry. That an army of statisticians have revised the process with new homogenizing modification that gave new credence to meta-analysis. Further, this research design had meta-sized throughout the social sciences. Thus, we suggested that a study by Ioannidis indicated that a meta-analytic study when compared with a classic study in which both analyzed the same information, that there was a 35% difference in outcomes.(Ioannidis,1998) Further, the meta-analytic design was the culprit. Thus, the authors suggested that it is hard to compete with a method that is inexpensive and simple to complete to publish a study. Further, that these studies have also been on national news, as well as in books relative various social science subjects. Further research methods books also include meta-analysis. (Snell, J. & Marsh, M. 2009)
We then suggested that meta-analysis would improved if all studies had the same research strategy and same analysis. Thus although the number of studies had convenience samples, all else was the same.
Further, if there were 10 studies of which 5 used exactly the same procedure. One should use those 5 in stretching toward greater validity.(Snell, J. & Marsh, M. 2009)
We now suggest another strategy that may be as simple and inexpensive as the last strategy. The problem with searching for similar studies using the same procedures is that a literature review does not uncover enough studies.
Thus, we suggest that the author work with others. They should promise of attribution or authorship for gathering data using the exact questionnaire. Then distribute it in large introductory classes throughout the country. …