Confounding Variables, Reliability, and Validity - Essay Example

A confounder will be related to both the independent and the outcome variables. A fairly classic, everyday example of this is a fictional study that found a correlation between ice cream consumption and murder rates.  In studying the two variables in a specific area, e.g., New York City, it would become clear that there is a relationship between ice cream consumption and murder because, over the year, they both rise and fall in a seemingly related pattern.  Does homicidal proclivity derive from consuming too much ice cream?  No, it doesn’t.  The confounding variable here is the heat.  As summer becomes hotter, people eat more ice cream; research has shown they also tend to have more violent outbursts.  In this case, there is no valid relationship between the two variables even though, from a limited perspective, there appears to be one. 

Reliability and Validity

     For a research study to be effective or believable, it has to have strong conclusions (validity) that are based on consistent measurements (reliability); thus, the research design needs to consider these two elements and factor in methods that ensure a high quantity of each (Cohen, Manion, & Morrison, 2000, p. 129).  This approach provides the reader or researcher with confidence that the data was measured in a scientific and repeatable way, and that the conclusions are accurate and truthful as we can best understand them.  A classic area for the explanation of these principles can be seen in any research study that employs the use of questionnaires with a sample population of individuals; the sample size can make the difference between a valid, reliable study and one that cannot be depended upon at all.  In most studies with small sample sizes, the authors will list that fact as a limitation or use it to call for additional research.  Similarly, those studies that have a large sample size consider that fact as an “asset” due to the strength or validity it lends to the conclusions (Kane, 2005, p. 59).  Regardless of the question, a study group of two individuals will not yield reliable results.  Conversely, a study population of hundreds of thousands of individuals will.  In sum, there is a high reliability which promotes the ability to draw truthful conclusions giving the study validity.