Research 101: Understanding Research Studies
One of the secrets of science is to understand the language of science, and science’s primary language is the research study. Research studies allow scientists to communicate with one another and share results of their work. There are many different kinds of research and many varying fields of research. And although journals were designed to help professionals communicate such research findings with one another, many times professionals in one field don’t significantly interact with (or are even aware of) researchers in a different field than themselves (e.g., a neuropsychologist may not keep up on the same research findings as a neurologist). This article reviews the major types of research done in the social, behavioral and brain sciences and provides some guideposts to better evaluate the context in which to place new research.
Types of Research
The basis of a scientific research study follows a common pattern:
- Define the question
- Gather information and resources
- Form hypotheses
- Perform an experiment and collect data
- Analyze the data
- Interpret the data and draw conclusions
- Publish results in a peer-reviewed journal
While there are dozens of types of research, most research done falls into one of five categories: clinical case studies; small, non-randomized studies or surveys; large, randomized clinical studies; literature reviews; and meta-analytic studies. Studies can also occur in widely varying fields, from psychology, pharmacology and sociology (what I’ll call “behavioral and treatment studies”), to genetics and brain scans (what I’ll call “organic studies”) to animal studies. Some fields contribute results that are more instantly relevant, while others’ results may help researchers develop new tests or treatments decades from now.
Clinical Case Studies
A clinical case study involves reporting on a single case (or series of cases) that the researcher or clinician has tracked over a period of some significant time (usually months or even years). Many times, such case studies emphasize a narrative or more subjective approach, but may also include objective measures. For instance, a researcher might publish a case study about the positive effects of cognitive-behavioral psychotherapy for a person with depression. The researcher measured the client’s level of depression with an objective measure such as the Beck Depression Inventory, but also describes in detail the client’s progress with specific cognitive-behavioral techniques, such as doing regular “homework” or keeping a journal of one’s thoughts.
The clinical case study is a very good research design for generating and testing hypotheses that may be used in larger studies. It is also a very good manner for disseminating the effectiveness of specific or novel techniques for individuals, or for those that may have a fairly uncommon set of diagnoses. However, generally a clinical case study’s results are not able to be generalized to a broader population. A case study is therefore of limited value to the general population.
Small Studies and Survey Research
There’s no specific criteria that differentiates a “small study” from a “large study,” but I place any non-randomized study in this category, as well as pretty much all survey research. Small studies are generally conducted on student populations (because students are often required to be a research subject for their university psychology classes), involve less than 80 to 100 participants or subjects, and often lack at least one of the core, important research components most often found in larger studies. This component can be the lack of true randomization of subjects, a lack of heterogeneity (e.g., no diversity in the population being studied), or a lack of a control group (or a relevant control group, e.g. a placebo control).
Most survey research also falls into this category, because it also lacks one of these core research components. For instance, a lot of survey research asks participants to identify themselves as having a particular problem, and if they do, then they fill out the survey. While this will almost guarantee the researchers interesting results, it’s also not very generalizable.
The upshot is that while these studies often provide interesting insights and information that can be used for future research, people shouldn’t read too much into these research findings. They are important data points in our overall understanding of the subject. When you take 10 or 20 of these data points and string them together, they should provide a fairly clear and consistent picture about the topic. If the results don’t provide such a clear picture, then there is likely more work to be done in the subject area before meaningful conclusions can be made. Literature reviews and meta-analyses (discussed below) help professionals and individuals better understand such findings over time.
Large, Randomized Studies
Large, randomized studies that draw from diverse populations and include relevant, appropriate control groups are considered the “gold standard” in research. So why aren’t they done more often? Such large studies, often done at multiple geographic locations, are very expensive to run because they include dozens of researchers, research assistants, statisticians, and other professionals as well as hundreds, and sometimes thousands, of subjects or participants. But the findings from such research are robust and can be generalized to others far more easily, so their value to research is important.
Large studies are not immune to problems found in other kinds of research. It’s just that the problems tend to have a much smaller effect, if there are any, since the number of subjects is so large and mixed (heterogeneous). When properly designed and using accepted statistical analyses, large research studies provide both individuals and professionals with solid findings that they can act upon.
A literature review is pretty much what it describes. Virtually all peer-reviewed, published research includes what might be called a “mini literature review” in its introduction. In this section of a study, the researchers review previous studies to put the current study into some context. “Research X found 123, Research Y found 456, so we hope to find 789.”
Sometimes, however, the number of studies in a particular area of study is so large and covers so many results that it’s difficult to understand exactly what our understanding is at the moment. To help give researchers a better understanding and context for future research, a literature review may be conducted and published as its own “study.” This will basically be a comprehensive, large-scale review of all studies in a particular area published within the past 10 or 20 years. The review will describe the research efforts, expand on specific findings, and may draw some general conclusions that can be gleaned from such a global review. These reviews are usually fairly subjective and are mainly for other professionals. Their use to the general public is limited and they almost never produce new findings of interest.
A meta-analysis is similar to a literature review in that it seeks to examine all previous research in a very specific topic area. However, unlike a literature review, a meta-analytic study takes the review one important step further – it actually pulls together all of the previous study’s data and analyzes it with additional statistics to draw global conclusions about the data. Why bother? Because so much research is published in many fields that it’s virtually impossible for an individual to draw any solid conclusions from the research without such a global review that pulls together all that data and statistically analyzes it for trends and solid findings.
The key to meta-analytic studies is to understand that researchers can alter the results of such a review by being particular (or not very particular) about the kinds of studies they include in their review. If, for instance, the researchers decide to include non-randomized studies in their review, they will often get different findings than if they hadn’t included them. Sometimes researchers will require certain statistical procedures to have been performed in order for the study to be included, or certain data thresholds to be met (e.g., we’ll only examine studies that had more than 50 subjects). Depending upon what criteria researchers choose to include in their meta-analysis, it will effect the results of the meta-analysis.
Meta-analytic studies, when done properly, are important contributions to our scientific knowledge and understanding. When a meta-analysis is published, it generally acts as a new foundation for other studies to build upon. It also synthesizes a great deal of previous knowledge into a more digestable Chunk of Knowledge for everyone.
Three General Categories of Research
While we’ve discussed the five general types of research in behavioral and mental health, there are also three other categories to consider.
Behavioral & Treatment Studies
Behavioral or treatment studies examine specific behaviors, treatments or therapies and see how they work on people. In psychology and sociology, most research conducted is of this nature. Such research provides direct insights into human behavior or therapeutic techniques that may be of value for treating a specific kind of disorder. This kind of research also helps us better understand a specific health or mental health concern, and how it manifests itself in a certain group of people (e.g., teenagers versus adults). This is the most “actionable” type of research – research that professionals and individuals can take action based upon its findings.
Research that examines brain structures, neurochemical reactions via PET or other brain imaging techniques, gene research, or research that examines other organic structures in a human body falls under this category. In most cases, such research helps further our understanding of the human body and how it functions, but doesn’t provide immediate insight or help in dealing with a problem today, or suggest new treatments that will be readily available. For instance, researchers often publish findings about how a particular gene may be correlated with a specific disorder. While such findings may eventually lead to some sort of medical test being developed for the disorder, it may be a decade or two before a finding of this nature translates into an actual test or new treatment method.
While such research is vitally important to our eventual better understanding of how our brains and bodies function, research in this category tends not to have much importance today for people dealing with a mental disorder or mental health problem.
Research is sometimes conducted on an animal to better understand how a specific organ system (such as the brain) reacts to changes, or how an animal’s behavior may be altered by specific social or environmental changes. Animal research, mostly on rats, in the 1950’s and 1960’s focused on studying animal behavior which, in psychology, led to the field of behaviorism and behavior therapy. More recently, the focus of animal studies has been on their biological makeup, to examine certain brain structures and genes related to health or mental health issues.
While certain animals have organ systems that may be very similar to human organ systems, results from animal studies are not automatically generalizable to humans. Animal studies are therefore of limited value to the general population. Research news based upon an animal study generally means any possible significant treatments from such a study are at least a decade or more away from being introduced. In many cases, no specific treatments are developed from animal studies, instead they are used to better understand how a human organ system functions or reacts to a change.
Research in the social sciences and in pharmacology is important because it helps us not only better understand human behavior (both normal and dysfunctional behavior), but also to find more effective and less time-consuming treatments to help with a person is suffering from an emotional or mental health issue.
The best kind of research – large-scale, randomized studies – are also the most rare because of their cost and the amount of resources needed to undertake them. Smaller-scale studies also contribute important data points along the way, inbetween the larger studies, while meta-analyses and literature reviews helps us gain a more global perspective and understanding of our knowledge so far.
While animal research and studies into the brain’s structures and genes are important to contributing to our overall better understanding of how our brains and bodies function, behavioral and treatment research provide concrete data that can generally be used immediately to help people improve their lives.
Grohol, J. (2013). Research 101: Understanding Research Studies. Psych Central. Retrieved on July 3, 2015, from http://psychcentral.com/lib/research-101-understanding-research-studies/