Social interdependence theory and applications

Health Care

Evaluating the research endeavors of someone other than oneself is a task that requires honesty, integrity, and fairness. The one performing the evaluation must be sure that all evaluation opinions are properly evidenced and supported, as there exists an enormous difference between having an opinion about that which has been researched as opposed to being opinionated. In addition the research being evaluated, albeit experimental, descriptive, clinical or a newspaper research report, must exhibit a clear statement of the problem and hypothesis, a fully described method of data collection and analysis, and thorough evidentiary support as to need. Without meeting these conditions the research being conducted is of little value. The remainder of this paper will focus on one particular form of research, namely the newspaper research report, with respect to a health care issue; i.e., teenage health related matters and educational issues and the statistical methods employed to draw conclusions about the data.

The newspaper research report chosen is one published in the Commercial Dispatch authored by Rebecca Ennis (June, 2004) and entitled "State Ranks Last In Kids Count Survey." The overall focus of the research article was a survey assessment of the rate of teen pregnancy 15-17 years of age, percentage of single-parent families, percent of children in poverty, percent of children 16-19 not in school or working, child birth weight mortality rate in the State of Mississippi. Data presented visa via various surveys framed the conclusion that although teen pregnancies in the State of Mississippi from 1998-2002 dropped by 24 per cent, the State continues to remain last in a national decline with respect to the number of teen pregnancies reported. Reported as well was an increase of 2.1% in births to unmarried teens; 15% of all babies born were to single teenagers; 9% drop in births to teens who are married; an18% increase in the number of teens between the ages of 16 and 19 not working; 8% increase in the number of low birth weight babies; a 5% increase in birth baby mortality; and an increase in the percentage of children eligible for free lunches.

Having all survey information in hand the next task is to evaluate the soundness and efficacy of the resulting survey data. Knowing that statistics is a branch of scientific methodology, as it deals with the collection, classification, description, and interpretation of measurement data, it is essential that the statistical tool employed be robust and able to deliver the necessary value in order to drawn conclusions about the research being conducted and reported. The soundness of the statistical tool is expected for those statistics, and equally applied, that are basic (percentages) as well as for those that are complex (Baye's coefficients). In all research the essential purpose of the statistic employed if to describe and draw inferences about the numerical properties of the chosen sample or population as well as to compare data gathering procedures. Further, the chosen statistical tool is directly aligned, and responsible, for answering the posed research question and testable hypothesis. Without a felt research need, proposed research question and testable hypothesis, the statistical tool has little meaning nor does the resulting data.

The information gathered by Ennis (2004) and reported in her newspaper research report is based upon surveys conducted in Mississippi with respect to certain assessed area-addressing teenagers (i.e., pregnancy, birth child's birth weight, mortality rate of birth child, school lunch usage, poverty, employment of teenagers). All survey information gathered was presented in percentage form and comparisons were made between Mississippi teenagers and those on a national basis. However, in the hands of a research novice percentage reporting can be extremely erroneous and the inferences garnered from percentage values very misleading. When employing percentages various errors occur if the following reporting principles are not made known, namely those reported below.

Effective use of percentages must include an accounting of the population upon which the percentages were reported; i.e., thousands, hundreds, or very few. Reporting that a certain percent of a particular survey group falls into one category or another is almost uninterruptible when the survey population is extremely small. Survey accountability increases as the numbers surveyed increases. 10% of a sample survey of 10 has far less meaningfulness then 10% of a sample survey of 100,000. Unfortunately Ennis does not report the size of the sample used in the Mississippi data collection survey. Therefore, the conclusions she draws are suspect.

Survey percentages are not to be used to make comparisons between variables being surveyed. For example: and 2.5% increase in one variable cannot be affiliated with an increase of decrease in another variable. There can be absolutely no cause and effect relationships established as a result of an increase or decrease between or amongst survey variables. Any comparison must be accomplished through use of more appropriate statistical tools that are specifically designed to show whether or not cause and effect relationships exist. Therefore, Ennis' statement that teen pregnancy has increased in Mississippi as well as has poverty amongst teenagers a causal relationship or effect between the two cannot be reported. Unfortunately the author uses the two increases to establish a relationship between poverty and teen pregnancy.

Survey percentages cannot be used to infer future trends, as this constitutes the necessity of using pre-selected inferential statistics. Here again Ennis fails in prudent research reporting by stating that the survey percentages depict a future increase, or minimally no increase, in teen pregnancies due to poverty. Percentage comparison simply cannot provide as such inference as they are not raw data.