Getting Started Guide
The Biostats Core has devised a series of items to consider when submitting a request, designing a research study, sending a protocol to the UND IRB or IACUC, and/or submitting a manuscript for publication. We believe that these items will strengthen the quality of your research and make it more likely to be accepted for publication.
Required Trainings
All key personnel on research studies are required to have current ethics training on file with the UND IRB office. The links below will take you to the appropriate training pages through UND.
Human Studies
https://und.edu/research/resources/human-subjects/human-subject-education.html
Animal Studies
Research Questions
- Please refer to the following UND TRANSCEND video when consulting the Core on this topic: Idea Generation for Hypotheses and Experiments.
- A review of the existing literature in your field is the most effective way to develop a testable research question. You may even be able to get a systematic review or meta-analysis published in the process! See the UND SMHS Literature Review guide for more details.
Proposal and/or Protocol Preparation
- Provide an overview of your research project and include any relevant references for similar studies. This can be helpful for determining appropriate effect size and power analysis.
Software
- The software the Core uses and supports for sample size calculations, randomization schemes, and data analysis includes SAS, R, SPSS, and Excel. Exceptions may be made on a case-by-case basis. Keep this in mind when submitting a request to the Core.
Study Design
- Please refer to the following UND TRANSCEND videos when consulting the Core on this topic: Designing an Epidemiological Study and Communicating Your Data to Statisticians.
- Consider your budget when designing your study ensure that your funding is adequate for what you are proposing to do.
- Have an idea of what data elements, or variables, you are planning to collect. This will include what you collect, where you collect it, how you collect it, and where it will be stored.
- A final study design is not a requirement of working with the Core, staff can assist you in creating a final study design for implementation.
Sample Size
- Please refer to the following UND TRANSCEND video when consulting the Core on this topic: Power Analysis Basics.
- Many factors go into determining sample size for a specific outcome in a study, so the Core is unable to provide a rule-of-thumb number (e.g. 30 subjects).
- We recommend using 90% power for outcomes – especially when the study has multiple primary and/or secondary outcomes.
Survey Support
- For publicly available data, NHANES has some useful tutorials on survey sampling, combining survey cycles, and appropriately
analyzing survey data in SAS, SUDAAN, STATA, and R.
- Keep some of these details in mind when designing and sharing your surveys with the Core.
- The Core can assist in developing survey questions, creating instruments in Qualtrics or REDCap, and managing survey workflows to ensure accurate and timely data collection.
Data Management
- Please refer to the UND SMHS Library Data Management and Research guide and UND Chester Fritz Library Data Management guide when consulting the Core on this topic.
- It would also be prudent for the client to familiarize themselves with the Data Management & Sharing Plan from UND SMHS Grants Management.
- De-identify data sets prior to sharing with the Core. For file sharing, we recommend using LiquidFiles. It is a secure file transfer system that has an Outlook plugin that can be used.
Data Analysis
- Please refer to the following UND TRANSCEND video when consulting the Core on this topic: Stats to Data: Matching Statistical Tests to Datasets.
- If the Core and client(s) have misalignments on the analysis method, the Core may perform a sensitivity analysis (e.g. intention-to-treat analysis vs per-protocol analysis vs completers analysis). Ideally, we would prefer that any sensitivity analysis is prespecified in the protocol.
- It may be helpful to read through a paper by Andrew Gelman and Eric Loken on the garden of forking paths in frequentist statistics when considering any analysis choices based on the data.
Data Interpretation
- The Biostatistics Core staff offer assistance with interpretation of results. We will work with PIs to provide optimal analytic procedures and results reporting.