- Methodist Research Institute
- BASIC SCIENCE RESEARCH
- CLINICAL RESEARCH GROUP
- MEDICAL RESEARCH LABORATORY
- BIOSTATISTICS & DATA SERVICES
- BIOMEDICAL COMMUNICATION
- Medical Library
There are many common challenges in designing and performing research experiments and questions surrounding statistical issues are usually plentiful. MRI offers statistical expertise to investigators at IU Health who are involved in a wide variety of projects.
From laboratory experiments to clinical trials, the use of statistics in medical research is ever present. Whether being used in a simple descriptive manner or to evaluate complicated mathematical models, statistics are often needed to help answer scientific questions. In general, most research studies consist of collecting data from a sample of the target population and making inferences about the population based on the findings in the sample. Although there are many specific uses for statistical theory in these research projects, statistics play two main roles in the experimental process:
Planning stage – The focus is on the design of the experiments, which is essential to produce any type of meaningful result. Many experiments or research projects take a turn for the worse before any data are even collected. A poor study design can doom a research project from the onset. Ensuring that necessary data are collected using proper methods is a key step in building a quality study. Too often data are collected without regard to the research question, only to find that the information gathered is insufficient for drawing meaningful conclusions. This insufficiency may be in the type of information collected or in the amount collected. Attention must be given to the amount of data needed in order to have a reasonable chance of detecting the hypothesized differences.
Data analysis – Appropriate model-building techniques, computation, and interpretation of results are of paramount importance when answering the research question. Once an experiment is completed and all information has been gathered, statistical analysis is often required. With the advancement of software packages, it is possible for almost anybody to perform statistical tests. Unfortunately, this also affords the opportunity for misuse and abuse of statistics. Understanding the underlying assumptions and statistical theory behind these models and tests is of great importance when making inferences based on the results. Many commonly used concepts, such as confidence intervals, power and P-values are often misunderstood, leading to misstated conclusions.
Collaboration or consultation with a statistician at all stages of the research process will be beneficial to the project and will aid in avoiding common pitfalls, helping the project to move along more efficiently.
So, are you wondering which of your outcome variables might best capture the desired information? How your data should be analyzed? How many subjects will be needed for your study to have adequate power? What statistical information is needed for your IRB submission? These are just a few examples of issues that MRI’s statistical consulting can help answer.
To discuss your project, please contact Colin Terry (317.962.2045).