Blog 6:  Secondary Data Analysis

Secondary data analysis has several advantages for researchers.  One advantage of secondary data research is that if provides access to social processes that may in some instances be inaccessible to researchers.  One example of this advantage would be for researchers that want to study events that have already happened and will not recur, more specifically events such as assassinations of individuals or presidential elections.  Another advantage of secondary data analysis is that it can save a researcher considerable time and money.  A huge part of considering whether to engage in research is the feasibility of research due to both time and financial constraints.  Using secondary data analysis allows researchers to review findings concerning their research interests, essentially getting the research they need without having to conduct the research.  This saves significant time and money but inevitably causes a potential disadvantage that will be discussed in my writing of disadvantages.  In addition using secondary data analysis allows researchers to compare findings with other samples of data.  The main advantage of being able to compare data to other data is for a standard of comparison and determination if data supports or refutes each other-which can then help a researcher determine the validity of findings and generalizability.  One other very valuable advantage to using secondary data analysis is that it potentially allows for studies to be combined, providing a larger set of data that could be more comprehensive and potentially more generalizable.  The advantages to being able to access and use secondary data analysis have allowed researchers to learn much about past events, and in addition research current phenomena with strong supporting evidence and a significant time and cost savings.

Although there are significant advantages to using secondary data analysis it also has potential for several disadvantages.  The research question is the most important component of research-it determines the focus and scope of research.  When researchers use secondary data they are unable to design the data collection methods which are directly related to the research question.  This disadvantage essentially means that the secondary data that is accessible may not relate to the research question and is therefore worthless to the researcher.  A second disadvantage to using secondary data to research current research questions is that it eliminates the ability to engage in iterative processes.  Being able to make observations, develop ideas, and then make additional observations and changes to concepts allows researchers to be dynamic and evolve their research over time.  Using secondary data disallows for these dynamic processes.  The use of secondary data is very static, it allows for the analysis of past data, but does not allow researchers to “think on the go” and modify their steps throughout the research processes.  Another disadvantage of secondary data analysis for research is that there is no pretest, research, and post-test design that provides valuable data throughout the research process.  The use of multiple levels of testing can provide researchers with data necessary to provide a causal connection and generalizability, the use of secondary data analysis does not allow for this level of research dynamic.  Lastly, a major concern and potential disadvantage of secondary data analysis are the ethical considerations and concern for quality.  The quality of secondary data may be questionable and researchers must go through as many quality control steps as possible to ensure that the data being researched is of high quality- mostly meaning is the data accurate, and going to provide valid and effective data to support the research question.  Ethical considerations when reviewing secondary data are less than current research concerns but still exist.  The central concern whenever data is being analyzed is the confidentiality of personal information.  To best address subject confidentiality, removing all identifying information is the most effective method for preserving confidentiality.  Another ethical concern is to make sure that providing proper credit for previous research is given.  Properly citing previous researchers when releasing research findings ensures adherence to ethical codes.  IRBs may be involved in ethical concerns when the research is at the federal level and past research is being used to help support current research questions.

I reviewed a peer-related journal that concerned the evaluation of cardiovascular risk and mortality in patients using diabetes medication (Gallagher, Smeeth, Seabroke, Leufkens, and Staa, 2011).  The research involved using past research data to show links between drug types and lower or higher rates of mortality.  The past research was termed the “bias analysis”.  The bias analysis was compared to data obtained from current research to determine the connection between mortality rates and diabetic medication users.  The past research was specifically from  the General Practice Research Database (GPRD) in the United Kingdom, GPRD comprises the computerized medical records maintained by general practitioners (Gallagher, Smeeth, Seabroke, Leufkens, and Staa, 2011).  One central concern when using past data is the concern of the translatability of the studies.  Did the secondary data that was analyzed actually support and help answer the current research question?  In this research case the answer is yes because the secondary data analysis provided a specific example of exactly what researchers were looking to answer.  The secondary data analysis was relevant but the findings were significantly different in results from the current research.  The most intriguing finding when the past data was compared to current research findings was that current drug users had an increased risk of death of that of past studied participants.  The discrepancy in data can be used to both provide answers to current questions as well as potentially providing evidence that argues in favor of current hypotheses.   The article provided a valuable example of how secondary data analysis was used to answer current research questions and in this case provide a standard of comparison for the current research results.

Link to PDF: http://journals.plos.org/plosone/article/asset?id=10.1371%2Fjournal.pone.0028157.PDF

Reference:

Gallagher, A., Smeeth, L., Seabroke, S., Leufkens, H., Van Staa, T., & Malaga, G. (2011). Risk of Death and Cardiovascular Outcomes with Thiazolidinediones: A Study with the General Practice Research Database and Secondary Care Data (Thiazolidinediones and Cardiovascular Risk). PLoS ONE, 6(12), E28157.

One comment on “Secondary Data Analysis

  • I don’t agree with one statement you make: “Another disadvantage of secondary data analysis for research is that there is no pretest, research, and post-test design that provides valuable data throughout the research process.” There are secondary data sets that report on results of experiments and contain pretest and posttest data. Just go to the ICPSR website and enter “experiment” as a search term. It’s interesting that you discuss a study that uses computerized medical records from GPs in Great Britain. They must be far ahead of us in developing that kind of data set together.

Leave a Reply

Your email address will not be published. Required fields are marked *