Understanding the mental health of students of color in the U.S. during heightened public xenophobia and bigotry

The identifiable institutional data and the survey data will be stored on the M+ Google Drive. As explained in other sections of this application, survey responses and respondents’ identifiable information (name, email address) will be stored in separate files in separate folders. Non-identifying registrar data (date of birth, sex, race/ethnicity, citizenship status, degree level, year in program, grade point average, and field of study) are linked with the survey data using the student’s unique survey link. However identifying registrar data (e.g. name and email address) are never linked with the survey data. The two files (identifiable data and survey data) will be linkable by the student’s unique survey link (added by the research team as a column in the spreadsheet of identifiable data before recruitment and contained as a column in the survey data file downloaded from Qualtrics). Identifiable data will never be merged with survey responses (i.e., this information will always be stored in two separate files, the only linking piece of data being the survey links). While the study is in the field, it is necessary to retain all identifying information originally provided by the institution in order to contact non-responders. Once data collection has ended, we will destroy the contact information of non-responders (name, email address) but we will retain their other information for the purposes of non-response analysis (date of birth, student status, race/ethnicity, sex, GPA, citizenship, field of study, and school year). We will destroy the identifiable information of non-responders within 10 days of completing data collection at an institution. We will retain the full information provided for responders, again with the identifiable data and survey data in separate files in separate folders linkable only by their unique survey link. This protocol is consistent with previous iterations of our survey studies.

Identifiable data will never be shared with other researchers or institutions (including the Refresh researchers). The data which can be shared with researchers will always be de-identified.

We will encrypt all survey data given its sensitive nature.

Each participating school will have two designated folders on the M+ Google Drive: one for institutional data containing individual identifiers and another for de-identified survey data (e.g., (1) University X identifiable institutional data; (2) University X survey data). The institutional contact (commonly someone from the institution’s Registrar’s Office) will upload the student data file with identifiable information into the identifiable institutional data folder on the Google Drive. Ability to do this will be restricted, with this individual’s permissions removed immediately after the file is successfully uploaded. After successfully uploading this file, only the HMS study team will have access to this file. The study team will then add a column to this file for the unique links (which will have been generated through Qualtrics by the study team). That file will then be used in a mail merge to generate individual emails for each randomly sampled student at each participating institution for the initial recruitment. These emails will be saved in the identifiable institutional data folder (in a separate file from the institutional identifiable data file), as these will contain fields with identifiable information—student’s first name and email address. The school’s other folder (the survey data with no identifiable information) will contain no files until the data collection period has begun. On the day of reminder email #1, the study team will download the current survey data from Qualtrics in a .csv file). This data file will contain a column with the unique survey links of those who have participated or indicated that they do not consent). That one column (of survey links) will be extracted from the spreadsheet (with no other information or columns) and saved as a separate file in the survey data folder. That file (with one column) will then be opened and saved again, this time into the identifiable institutional data folder. Once the single column .csv file of unique links has been saved and opened from its location in the identifiable institutional data folder, the identifiable institutional data file will be opened and the links from the single column .csv file will be removed as well as the unique links of any individuals who have emailed the study team to indicate that they do not wish to participate. The identifiable institutional data file will then be saved with a new date. That file will then be used in a mail merge to generate individual emails for the sample of non-responders (who have not otherwise indicated that they do not wish to participate) for reminder email #1. The emails will be saved in the identifiable institutional data folder and sent out using the process described above. This process will be repeated for reminder emails #2 and #3.

The folder where de-identified survey data will be stored will be accessible to the HMS study team.

Maintaining respondent identifiers (i.e., identifiable information for those who completed the survey) after the original round of data collection will be necessary for conducting follow-up analysis (as described in the consent forms). The HMS study team may follow-up with a sub-set of respondents within three years of original data collection in order to understand the changing needs of college and university students related to the survey topics. We can only study these issues satisfactorily through individual-level longitudinal analysis. Individual-level longitudinal analysis will allow us to analyze how changes over time relate to specific individual characteristics at baseline as well as changes in individual characteristics from baseline to follow-up. The HMS study team may also collect additional academic records such as retention and GPA, which would be linked to the institutional data by identifiable information and then the unique link contained in the identifiable institutional data file would be used to link the academic records data to the survey data without any identifiable information making its way into the file/folder of survey data. The HMS study team would not obtain additional consent to obtain these academic records. The following language is included in the consent form in order to inform participants of this possibility:

Some schools may request additional analysis on how measures from the survey correlate with academic outcomes. In this case, we will access your academic records stripped of any identifying information, solely for the purpose of this research analysis. The aim of this analysis would be to gain knowledge about how to promote successful academic outcomes. We will protect the confidentiality of these records using the measures described throughout this consent form.

If follow-up of additional academic and school records is requested from a participating institution, the study team will use the following protocol:

1. The HMS study team will provide to the school's registrar (or academic information office) the list of student identifiers (university emails) for the students who participated in the survey. We will do this through sharing a M+ Google Drive folder (with only that school's list of students). 
2. The registrar then returns to us the academic records (GPA, credits, degrees conferred, enrollment status) for those students with the identifiers still attached. Again, this transfer will be through Google Drive.
3. The HMS study team will use the linking file (with mapping between identifiers and non-identifying study IDs) to link the academic records to the study IDs, and remove the identifiers, leaving a file with just the academic records and study IDs.    
4. The HMS study team will then link the academic records to the survey data using the study IDs.

The additional academic records and follow-up analysis described in this paragraph will only be collected and conducted at institutions that request this additional research. With this protocol, as with previous iterations of the study, we will not request amendments for each follow-up using academic records. We will still request approval from the IRB in the case of follow-up with students within three years.

Data will be gathered through Qualtrics, using Transport Layer Security (TLS), encryption (HTTPS), password protection, and HTTP referrer checking. Qualtrics has SAS 70 Certification and meets the rigorous privacy standards imposed on health care records by the Health Insurance Portability and Accountability Act (HIPAA). All Qualtrics accounts are hidden behind passwords and all data is protected with real-time data replication. Only the HMS study team will have access to the data through Qualtrics, protected by a password.

The files containing student identifiers and the files that contain linking information to the identifiers will be destroyed when no further analysis is required with identifiable information, or after three years, whichever comes first. The de-identified results will be stored indefinitely on the PI’s M+ Google Drive for future aggregate analyses.