Copy and Pasting vs. Customization: Using Qualitative Analysis of NIH Grantees’ Data Management Plans to Shape Future DMP Support

Abstract

In 2019, our school began requiring all of its NIH grantees to submit data management plans designed to prepare researchers for future requirements, such as the NIH Final Policy on Data Management and Sharing. In 2020-2021, the second year of this mandate, we achieved near 100% compliance with 600+ project-level DMP submitted. This presentation discusses analyzing the free text responses for the 2020-2021 set of DMPs. These responses describe practices related to data collection procedures, QA/QC procedures, data retention periods, data sharing, and access permissions when team members leave. Additionally, we asked respondents to identify relevant funder data management or sharing requirements, citations for data reuse, and applicable standards for data and metadata so that we could gauge awareness of these topics.

Our analytical process involves iteratively creating a codebook based on a randomly selected subset of DMP responses. Once the codebook has been developed, we will independently code a second subset of DMP responses to establish inter-rater reliability, then code the entire data set of DMP responses. We will summarize these results using descriptive statistics and visualizations of the themes identified.

We will also explain how we plan to use these findings to create guidance for future support and development of DMPs for NIH grantees. Of particular interest will be any evidence for substantive customization of DMPs when the research warrants the sharing of data, versus evidence of “box ticking” behaviors and other priorities that take precedence over sharing. By identifying existing practices, this analysis may also help us surface successful data management and sharing strategies that can be propagated across communities of practice.