Congratulations on your approved paper proposal! Whether you are working with a large or small writing team using cohort or RECVD GIS only data within SAS or R or Python, we hope that these resources help you to efficiently transition from proposal to published manuscript! Below you will find guidance for developing a statistical analysis plan and building accurate, reproducible code. These are not meant to be prescriptive, but instead give a starting place for teams that may want more guidance with next steps!
These documents were put together with the core purpose of facilitating communication between teammates. They are communication tools that help to (1) organize thinking about how to approach the research question, (2) unify the project across multiple analysts (especially if the project will span multiple people or institutions), (3) share your approach with others so that they can replicate and expand on your work!
1. Operations for Approved Study Proposals and Team Roles and Responsibilities
A study team is comprised of several key individuals. It is possible that the same person may fill more than one role depending on the activity.
Resources:
Potential Team Roles and Responsibilities: This document outlines the roles and responsibilities for members of a study team for each approved study proposal. Refer to the work instructions and guidelines in 2-6 for detailed tasks and expectations.
Operations for Approved Study Proposals: This documents outlines the steps that should be completed to support analyses for an approved study proposal.
2. Making a statistical analysis plan
While you have outlined a lot of your analysis in your proposal, creating a statistical analysis plan will allow you to hone in on exactly which variables you want to use, how they should be formatted, which analyses to run, and what output you may want. This can be especially helpful if the writing lead is different than the analyst. If you are working with a Drexel analyst, please contact them for their preferred method of receiving analyses instructions.
Resources:
Guidelines for a Statistical Analysis Plan: This document provides the study population(s), statistical methodology and mock table(s) and figure(s) to support the approved study proposal.
3. Creating clean derived datasets
A detailed description of the name and structure of the analytic datasets, definitions for the variables, will facilitate communication between the study team members.
Resources:
The RECVD Commandments for naming variables: See the RECVD Codebooks for details.
Derived Dataset Specifications: This document provides a detailed description of the dataset structure and variables contained in the analytic dataset that support the tables, figures, and analyses detailed in the SAP.
Data Organization in Spreadsheets: this article offers practical recommendations for organizing spreadsheet data to reduce errors and ease later analyses.
4. Writing clean, reproducible code
Tools have been developed to efficiently store and share data and programming code. These tools are become increasingly important as data and code sharing becomes the norm and a requirement of most funding agencies.
Resources:
Programming Guidelines: This work instruction provides guidelines specifically for the structure and documentation of code, not for its validity.
Validation Protocol: If you want someone to check your code this work instruction provides guidelines for writing a validation protocol that will guide an independent reviewer tasked with confirming the credibility of a program or system of code.
Examples of code from RECVD projects can be obtained by emailing the RECVD Research Coordinator at gslovasiresearch@gmail.com
5. Archiving code
If you want, Drexel will archive code from your project. You may choose to do this to help others across institutions or to make sure that new variables are consistent moving forward. If you would like to do this, we may request that you make edits to your code to ensure it has some of the elements referenced in 4. Writing clean, reproducible Code.
Resources:
For more information on archiving at Drexel contact the RECVD Research Coordinator at gslovasiresearch@gmail.com
6. Preparing good papers and presentations
Here is a collection of resources that are designed to help writing teams prepare good papers and presentations to communicate the results of their analyses.
Resources: