Effects of the COVID-19 pandemic on university-based research conduct and training: Exploring university-based researchers’ perspectives and experiences

Ethical approval and consent

The study was reviewed and approved by the Texas A&M Institutional Review Board (IRB2020-0622M). All included study participants provided electronic informed consent by reviewing the study information and checking “I agree to participate in this study.”.

Study design and study respondents

A cross-sectional online survey was collected (June 2020 – July 2020) from 872 individuals in a large Research 1 university system focused on public state-wide educational, research, and service in the United States designated as a member of the American Association of Universities. Faculty and staff of the university system were sent recruitment emails through the university system email lists. Since multiple email lists were used, some individuals received multiple recruitment emails. The recruitment email contained a brief description of the study and a link to the online survey, and the survey began with detailed study information and consent section. 872 opened the online survey, and 863 (99.0%) provided informed consent to participate in the study. Among those who consented, 709 (82.2%) were eligible to participate in the study. A person was eligible to complete the survey if he/she was a current faculty or staff in the university system and engaged in any research activities. The final sample included in this study was 642 respondents who reported whether they engaged in HSR or not. Of 642 respondents included in this study, 238 self-reported engaging in HSR. Given the limited data availability, approximate response rate was calculated for a sub-group (principal investigators at the main campus location). This sub-group (n = 248) constituted over 82% of the respondents from the main campus location and 48% of the overall respondents who reported their principal investigator status. The sub-group had about 11% response rates (248 out of ~2300). As a recruitment incentive, at the end of the survey, respondents were invited to participate in a drawing for a chance to win $100 gift cards, and five of eligible researchers were randomly selected and awarded from the drawing. A copy of the survey can be found in the Extended data (Ory, 2022).

Measures

This study used an online survey that was developed by the researchers and was sent to a number of university system researchers from various disciplines via email to evaluate its face validity. The questions used in this study was presented in Table 1. Main outcomes were specified to capture several possible aspects of COVID-19 impact on research: (1) overall impact on their research; (2) impact on specific research activities; (3) impact on student engagement; and (4) direct COVID-19 related changes (example: social distancing and wearing of personal protective equipment). Aspects of overall impact included: perceived severity of the impact, directionality of the impact, expected severity of the impact, and changes in research focus. Specific research activities included: data collection, data analyses, and dissemination of the study findings. Particularly for HSR, this study also examined participant recruitment, participant retention, and participant interaction. Because engaging students in research is considered an important aspect of research activities in large university systems, three different areas of student engagement in research were included: student recruitment, student financial support, and research-related learning opportunities for students. The direct changes in research due to COVID-19 included specific concerns and safety measures (e.g., social distancing and wearing of protective equipment) in relation to impacts on research activities. The questions that used 5-Likert point scale ranging from ‘very negative’ to ‘very positive,’ were dichotomized into ‘negative’ and ‘neither or positive.’ Dichotomization was performed because there were too few respondents reporting positive impacts. For example, for the question, ‘Taken all together, what has the impact of COVID-19 outbreak been on your research?,’ there were 4 (0.7%) respondents reporting ‘very positive’ and 18 (3.3%) reporting ‘somewhat positive.’

The primary independent variable was a dichotomous indicator, reflecting whether researchers engaged in HSR or not. This was determined based on respondents’ self-reported research types that they engaged in. Assessed covariates included principal investigator status (response categories: yes/no), number of years engaging in research, and use of experimentation or interventions. At the end of the survey, respondents reported their primary affiliation school/college, faculty status (response categories: faculty/staff/others) and sociodemographic characteristics (age, sex, race, and ethnicity). Primary affiliation was coded into multiple binary variables representing each school/college.

Analyses

Sociodemographic characteristics of researchers and their research roles and experiences were described using frequency and percentage or mean and standard deviation and were compared based on engagement in HSR (e.g., Chi-square test for categorical variables and independent group t-test for interval variables). Each outcome variable was described using frequency and percentage, and bivariate analyses (i.e., bivariate logistic regression) were performed to examine the association between the outcome variables and engagement in HSR. For each outcome variable, multivariable logistic regression models were performed by adjusting for principal investigator status, number of years engaging in research, and use of experimentation or interventions. All statistical analyses were performed using SAS 9.4 (SAS Institute Inc., Cary, North Carolina).

Study respondents

The majority of respondents were ages 35–64 years old (n = 362, 78.8%), male (n = 269, 56.6%), white (n = 394, 90.8%), not Hispanic (n = 424, 91.0%), faculty (n = 415, 78.3%), and currently serving as a principal investigator (n = 420, 81.6%). The respondents were from various fields of study including agriculture and life sciences (n = 78, 15.2%), health sciences (n = 60, 11.7%), liberal arts (n = 56, 10.9%), engineering (n = 55, 10.7%), sciences (biology, chemistry, mathematics, physics & astronomy, and statistics) (n = 51, 9.9%), education and human development (n = 46, 8.9%), and other fields (n = 168, 32.7%). On average, the respondents have engaged in research for over 19 years.

Compared to the respondents who did not engage in HSR, those who engaged in HSR were significantly more likely to be females (n = 106, 61.3% vs. n = 100, 33.1%, p < .001) and faculty (n = 163, 84.9% vs. n = 252, 74.6%, p = .02). Compared to those who did not engage in HSR, a greater percentage of respondents engaging in HSR were in education and human development (n = 44, 24.0% vs. n = 2, 0.6%, p < .001) and less percentage in agriculture and life sciences (n = 17, 9.3% vs. n = 61, 18.4%, p = .006), engineering (n = 6, 3.3% vs. n = 49, 14.8%, p < .001), or science (n = 4, 2.2% vs. n = 47, 14.2%, p < .001). While a greater percentage of respondents engaging in HSR were in health sciences (n = 28, 15.3% vs. n = 32, 9.7%), the statistical significance was marginal (p = 0.06). The respondents who engaged in HSR were also less likely to report using experimentation or interventions (n = 122, 51.3% vs. n = 252, 63.8%, p = .002) and had fewer years of experience engaged in research (16.0 vs. 20.8 years, p < .001).

Impact on research in academic setting

The majority perceived (n = 504, 92.1%) and expected (n = 494, 96.1%) at least some impact of COVID-19 on research, reported the impact to be negative (n = 430, 78.8%), and made changes in their research focus because of COVID-19 (n = 245, 45.5% somewhat and n = 64, 11.9% a lot) (Table 2). Compared to the respondents not engaging in HSR, the respondents engaging in HSR had significantly higher odds of perceiving (Odds Ratio (OR)  =  2.25, 95% Confidence Interval (CI) = [1.07, 4.74], p = .03) and expecting (OR = 4.86, 95% CI = [1.50, 15.79], p = .009) severe impacts of COVID-19 on research. Compared to the respondents not engaging in HSR, the respondents engaging in HSR were 2.28 times more likely to make some changes to their research focus and 3.73 times more likely to make a lot of changes to their research focus because of COVID-19. Adjusted associations between the perceived and expected impacts and engagement in HSR were comparable to the observed bivariate associations.

Impact on research activities

The majority of the study respondents reported negative impact of COVID-19 on data collection (n = 387, 76.5%), and over 41% reported negative impact on data analyses (n = 209) and dissemination of the study findings (n = 206) (Table 3). Compared to the respondents who did not engage in HSR, those who engaged in HSR were significantly more likely to report negative impact of COVID-19 on data collection (OR = 2.16, 95% CI = [1.36, 3.45], p = .001). There was not any statistically significant association observed between engagement in HSR and impact of COVID-19 on data analyses (OR = 0.94, 95% CI = [0.65, 1.36], p = .74) or dissemination (OR = 1.35, 95% CI = [0.93, 1.95], p = .11). After adjusting for principal investigator status, number of years engaging in research, and use of experimentation or interventions, the odds of the respondents who engaged in HSR to report negative impact on data collection remained higher than the odds of the respondents who did not engage in HSR (Adjusted Odds Ratio (aOR)  =  2.74, 95% CI = [1.63, 4.62], p < .001). The associations between engagement in HSR and impact of COVID-19 on data analyses (aOR = 0.92, 95% CI = [0.61, 1.37], p = .67) and dissemination (aOR = 1.28, 95% CI = [0.86, 1.92], p = .23) remained statistically insignificant after adjusting for principal investigator status, number of years engaging in research, and use of experimentation or interventions.

Impact on student engagement

The majority of the study respondents reported negative impact of COVID-19 on recruiting students for research activities (n = 327, 70.3%) and providing research-related learning opportunities for students (n = 328, 69.2%), and 43.1% (n = 192) reported negative impact on providing financial support for students. There was not any statistically significant difference observed in impact of COVID-19 on recruiting students for research activities (OR = 1.13, 95% CI = [0.75, 1.71], p = .56), providing financial support for students (OR = 0.98, 95% CI = [0.66, 1.45], p = .93), and providing research-related learning opportunities for students (OR = 0.999, 95% CI = [0.67, 1.50], p = .997) based on engagement in HSR. The association between engagement in HSR and impact of COVID-19 on student engagements remained statistically insignificant after adjusting for principal investigator status, number of years engaging in research, and use of experimentation or interventions.

Impact from social distancing and protective equipment

The majority reported at least some impact of the needs for social distancing (n = 465, 87.9%) or the needs for wearing personal protective equipment (n = 368, 70.9%) on conducting research (Table 4). Compared to the respondents who did not engage in HSR, those who engaged in HSR were significantly more likely to experience a lot of impact of the needs for social distancing (OR = 1.83, 95% CI = [1.02, 3.31], p = .045) or personal protective equipment (OR = 2.37, 95% CI = [1.42, 3.94], p < .001) on their research. After adjusting for principal investigator status, number of years engaging in research, and use of experimentation or interventions, the respondents who engaged in HSR remained significantly more likely to experience a lot of impact from the needs for social distancing (aOR = 2.82, 95% CI = [1.46, 5.45], p = .002) or personal protective equipment (aOR = 2.98, 95% CI = [1.69, 5.24], p < .001) than those who did not engage in HSR.

Limitations

The current study findings are based on self-reported perceptions rather than objective measures, such as research productivity. While the questionnaires used in this study was face-validated by researchers from various disciplines, the questionnaires did not go through a vigorous validation and reliability assessments. Further limitations of the current study include the sample representativeness and study generalizability. While the overall response rate could not be estimated, the response rates among the principal investigators at the main campus location was about 11%. In terms of demographics, the faculty respondents consisted of a higher proportion of White (85%) and a smaller proportion of female (43%), as observed in the institutional and national faculty demographics. The latest institutional diversity data available showed 63% non-Hispanic White and 36% female (Office of Diversity, 2021). Nationwide full-time faculty in degree-granting postsecondary institutions data showed less than 70% White and 44% female (NCES, 2018). There were some over-representation of White faculty members in this study, which may be related to the selection bias or related to the differences in the target population (i.e., faculty engaging in research or teaching vs. faculty engaging in research and teaching).

This study utilized the data collected from a large public Research 1 university system at a single point in time, and hence the study finding may not be generalizable to other types of universities or settings. However, given the common themes in COVID-19 associated changes (examples: social distancing, sanitation, and personal protective equipment), it can be assumed that the observed impact of COVID-19 on research is comparable in academic medicine and other settings and will continue to be a challenge as the pandemic timeline is uncertain. More studies from diverse research settings could help illuminate the impact of pandemics and other prolonged global emergencies on the research enterprise. Another limitation is that the sample had incomplete representation of researchers from diverse fields. While the study team attempted to make the survey relevant to all existing research fields by consulting about the survey questions and wording with researchers from various backgrounds, not all fields use common language and research methods. The study team received some questions from the respondents about the survey language (e.g., ‘secondary data’) and post-survey comments regarding relevancy of some questions (e.g., a question about participant interaction) in the survey to their fields. Therefore, this study focused on the general questions that were likely to be relevant in most fields. In addition, the survey questions are shared (Table 1) for the readers’ information.

Several brief reports and perspectives have anecdotally discussed the impact of COVID-19 on research and scientists (Omary et al., 2020; Yanow and Good, 2020; Wigginton et al., 2020). Despite the limitations, this study adds to the growing body of research based on primary data collected from researchers. First, this study provides important insights regarding the potentially lost learning opportunities for students to be addressed in higher education institutions, with specific focus on research training. Bridging from the lessons learned from the current pandemic, higher education institutions should develop the emergency standards of operation for faculty, staff, and students to enable continued learning and training opportunities. Further qualitative or mixed-methods evaluation of phenomenon can benefit the development of such emergency standards of operation. Lessons learned from this process will make us better equipped to deal with future challenges and the potential negative consequences to our ongoing academic research activities.

Second, the empirical findings in this study showed that researchers engaging in HSR have been disproportionately impacted by COVID-19 and that efforts are needed to further understand and overcome their challenges. For example, the needs for social distancing and wearing personal protective equipment generated additional hurdles for researchers engaging in HSR. Potential solutions might be to consider pandemic or emergency preparation budget as a part of the grant budget or higher education institutions develop emergency fund for research in academic settings.

While such efforts may already be underway in response to the swiftly changing circumstances, a natural progression of the current work would be to identify and examine the existing efforts and provide insights into ways that researchers, research institutions, and sponsoring agencies can adapt their research. These findings have strong implications for the academic medicine community, whether primarily lab-based studies or involving clinical or community-based research involving human subjects.

Next steps include continuing to analyze the ongoing impact of COVID-19, as well as its vaccine deployment, on both our HSR and non-HSR throughout our institution and utilizing our findings to try and find creative ways to mitigate the negative effects of the pandemic we are currently still facing or other challenges that we will face in the future. It is critical to move beyond a simplistic view of negative research impacts to a better understanding of how academic research and associated training can be enriched by meeting challenges of the past year.

Underlying data

The data supporting the findings of the study are not made publicly available due to the data use restrictions stated in the informed consent form (approved by TAMU IRB) and considerations that their containing information that could compromise the privacy of research participants. Given that the data were collected from staff and faculty from a university system there are concerns that a combination of several key demographic characteristics, research fields, and research types can potentially identify an individual. For researchers with specific research/study questions relating to the data, the anonymized data can be available on request from the lead author [MO]. Upon reasonable request, with TAMU IRB approval, we can share whether participants engaged in any human subject research (yes/no) and key outcome data (i.e., COVID-19 impact on research) without any demographic characteristics, research fields, and research type data that can potentially identify an individual.

Extended data

Texas Data Repository: Effects of the COVID-19 pandemic on university-based research conduct and training. https://doi.org/10.18738/T8/W6IGCE (Ory, 2022).

This project contains the following extended data:

Data are available under the terms of the Creative Commons Zero “No rights reserved” data waiver (CC0 1.0 Public domain dedication).