Develop your researchPage Menu
Once you have refined your research question, you are ready to select the study design best able to answer that question. This diagram that maps out the main types of research and their characteristics. This YouTube video also walks through different types of research methods.
Statistics How To has an excellent collection of videos and articles to help select the optimal experimental design (type of study) to answer your research question, and statistical tests commonly associated with each design.
Basic science, also commonly referred to as bench or foundational science, generates foundational knowledge that informs applied or translational science. This includes biochemistry, microbiology, and physiology.
Clinical science is alternatively called translational or applied medical science. Generally, it applies basic science research to human subjects, and includes clinical testing on individuals or communities. It often investigates the efficacy (whether an intervention works under ideal circumstances), effectiveness (whether an intervention works under “real world” circumstances), safety or cost of an intervention. Behavioral research, epidemiological studies, and those that investigate the mechanism of disease are also considered to be under the umbrella of the clinical sciences.
Literature reviews synthesize scholarly articles, books and other resources on a specific topic or area of interest. They critically evaluate the quality and quantity of source data, and summarize the findings to help inform the practical application of research. Literature reviews can be descriptive (does not include a formal assessment or analysis), systematic (includes rigorous review of all literature on the topic of interest), or meta-analyses (uses statistical methods to pool data from multiple sources).
- Ten Simple Rules for Writing a Literature Review (Pautasso 2013)
- The Literature Review: A Few Tips On Conducting It (Taylor)
Quantitative research uses objective measurement and statistics to turn numerical variables into generalizable information. It attempts to be conclusive in its findings, using deductive reasoning to answer a question or discover a pattern.
Standardized measurement of patient outcomes is essential to accurately assessing an individual’s baseline status, change over time, and the impact of an intervention for both clinical and research purposes. Outcome assessment instruments should be both valid (measure what you think they are measuring) and reliable (consistent regardless of who administers the test or when). Physiopedia has an excellent guide to selecting outcome measures.
The University of Oxford website contains a list of links to a range of patient self-reported outcome measures.
Orthopaedic Scores has a free calculation service for commonly used musculoskeletal orthopedic assessment instruments, including both clinician and patient self-report.
Desirable properties of outcome measures have been summarized in the table below.
|Thabane, L. et al, Posing the research question: not so simple! Can J Anesth 2009. 56:71-97.|
Creating a survey is a science unto itself. Here are some helpful resources to help insure that a new instrument is valid (measures what you think it measures) and reliable (consistent).
- Survey Design Checklist
- 6 Ways to Prestest Your Survey Before You Send It
- How to pretest and pilot a survey questionnaire
Online Survey Tools
Several secure web applications can be used to collect and analyze survey data.
Statistics is a branch of mathematics that research use to organize, analyze and summarize data. It is also often the most intimidating part of conducting or reading research. Unless the statistical testing needed to analyze your data is basic, we suggest you seek out a statistician for guidance or collaboration. Descriptive statistics or less complicated analysis can be calculated using free online programs. Microsoft Excel can be useful for conducting basic statistics. Statistics How To has an excellent collection of resources to help you understand and interpret basic statistics, calculate a variety of statistics, and generate charts and graphs to represent your data.
Sample Size and Power Calculations
Sample size is the number of participants or observations needed in a study to make the results accurate and reliable. This is generally represented in research as “n”. Power is the ability to find a difference between the items being studied, if a difference actually exists. Power calculations are statistical tests that help determine sample size, and are based on the size of the population of interest, the reliability of the data collection method used, and how much variation there is in your outcome of interest.
- Sample Size and Power Calculations Made Simple
- Power analysis for paired sample t-test | G*Power data analysis examples
- G*Power: Statistical Power Analyses for Windows and Mac
Data Safety and Management
Collecting data for a research study is an exciting phase of the work—and can quickly become overwhelming if you don’t have a good data management plan in place, including measures to safeguard that hard earned information.
- DataONE Best Practices
- Webinar: Best Practices for Research Data Management and Sharing (Johns Hopkins)
Qualitative research is typically exploratory, and utilizes inductive reasoning to understand a phenomenon through narration and/or observation. It can often be helpful to develop hypotheses or ideas for quantitative research.