How to critically evaluate the quality of a research article?

When considering a research idea, we are bound to rely on previous findings on the topic. Work done in the field constructs the foundation for our research and determines its course and value. Inaccurate findings may lead to imprecise applications and end in further fallacies in your own new scientific knowledge that you construct.  In order to set a solid basis for research on any topic and to prevent multiplication of misinformation, it is crucial to to critically evaluate existing scientific evidence. It is important to know which information can be regarded as plausible.

So what’s the criteria to determine whether a result can be trusted? As it is taught in the first classes in psychology, errors may emerge from any phase of the research process. Therefore, it all boils down to how the research has been conducted and the results presented.

Meltzoff (2007) emphasizes the key issues that can produce flawed results and interpretations and should therefore be carefully considered when reading articles. Here is a reminder on what to bear in mind when reading a research article:

Research question
The research must be clear in informing the reader of its aims. Terms should be clearly defined, even more so if they’re new or used in specific non-spread ways. You as a reader should pay particular attention should to errors in logic, especially those regarding causation, relationship or association.

Sample
To provide trustworthy conclusions, a sample needs to be representative and adequate. Representativeness depends on the method of selection as well as the assignment.  For example, random assignment has its advantages in front of systematic assignment in establishing group equivalence. The sample can be biased when researchers used volunteers or selective attrition. The adequate sample size can be determined by employing power analysis.

Control of confounding variables
Extraneous variation can influence research findings, therefore methods to control  relevant confounding variables should be applied.

Research designs
The research design should be suitable to answer the research question. Readers should distinguish true experimental designs with random assignment from pre-experimental research designs.

Criteria and criteria measures
The criteria measures must demonstrate reliability and validity for both, the independent and dependent variable.

Data analysis
Appropriate statistical tests should be applied for the type of data obtained, and assumptions for their use met. Post hoc tests should be applied when multiple comparisons are performed. Tables and figures should be clearly labelled. Ideally, effect sizes shou

ld be included throughout giving a clear indication of the variables’ impact.

Discussion and conclusions
Does the study allow generalization? Also, limitations of the study should be mentioned. The discussion and conclusions should be consistent with the study’s results. It’s a common mistake to emphasizing the results that are in accordance with the researcher’s expectations while not focusing on the ones that are not. Do the authors of the article you hold in hand do the same?

Ethics
Last but not least, ere the ethical standards met? For more information, refer to the APA’s Ethical Principles of Psychologists and Code of Conduct (2010).

References
American Psychological Association (2010, June 1). American Psychological Association Ethical Principles of Psychologists and Code of Conduct. Retrieved July 28, 2011 from http://www.apa.org/ethics/code/index.aspx

Meltzoff, J. (2007). Critical Thinking About Research. Washington, DC: American Psychological Association.

 

Edited by: Maris Vainre

About the author

Zorana Zupan

Facebooktwitterrss