Can you assure students that only response themes from the Doctoral Experience Survey will be shared with fields? If not, can we withdraw or edit our survey responses?
Date: April 2021
I am concerned about the recent response to a question about survey confidentiality in Ask a Dean.
While I appreciate the transparency in describing how the results will be collated, I found some details really concerning. I submitted my response to the biannual graduate student survey some time ago and provided written responses assuming full confidentiality, as was promised in the survey emails. However, the recent response to student concerns about anonymity includes the statement “do be mindful of writing anything that may disclose your identity” with regards to written responses. Even though I did not name myself, I am certain that the DGS of my field would know who I was from my written response at the end of the survey. I submitted that response with the understanding that results were fully anonymized and that only the Graduate School would see the ‘raw’ written responses.
Can you assure me and other graduate students that only the ‘themes’ of the written responses will be discussed with faculty in each field? If not, how can we withdraw or edit our survey data to protect ourselves and our professional futures? And when exactly will the results from this year’s survey be shared with fields?
A deeply concerned student
Dear Concerned Student,
First, thank you for responding to the Doctoral Experience Survey! Every reply is important for informing efforts to improve the graduate student experience and I appreciate that you took the time to share your thoughts, including in the space for written comments. I do understand why it is important to know that your reply will be kept confidential and I’m glad you asked about it.
As the previous Ask a Dean column about survey confidentiality indicates, within the Graduate School, confidentiality of the survey results is maintained through a combination of limited access, secure storage, masked identifiers, and reporting in aggregate. Handling of “raw” datasets (i.e., individual-level results) is limited to the staff who are directly responsible for data management, analysis, and reporting.
For written responses on the survey, an initial analysis involves parsing comments for patterns, themes, and suggestions — which are then reported such, without the individual comments (the Graduate School, Office of Diversity and Inclusion, Cornell Health, and Office of Global Learning are examples of units that may use the results of this analysis to inform their work).
A subsequent step involves redacting any text that could potentially identify the respondent. In some cases, the redacted information may be replaced by notes that indicate what has been removed. For example, “my advisor Dr. Wise” may be replaced with “my advisor [redacted name]”, “my experience as a Black and first gen student” replaced with “my experience [redacted personal identifiers]”, and “my research on human resource management for dairy farms” replaced with “my research [redacted specific topic]”. Occasionally, when a comment is interlaced with information that could potentially identify an individual, the entire written response will be redacted or paraphrased to conceal the student’s identity. After redactions are complete, the comments are included with the results to be shared with the Graduate School leadership team, as well as with the director of graduate studies (DGS) and graduate field assistant (GFA) for each graduate field.
For written comments to be helpful for informing or motivating change, they need to be detailed and specific enough to be relevant to the distinct graduate fields while also concealing the identity of individual student respondents. Creating this balance is a challenge for the Graduate School staff who redact the original text. It’s also the reason why I suggest that students be mindful of writing anything that may disclose their identity – because the process of redaction will err on the side of protecting student confidentiality, even when that means obscuring insightful comments. If comments do include potentially identifying information, the full comment will be considered in the thematic analysis, but the redaction process may obscure much or all of its substance before it is shared with anyone outside of the Graduate School.
In some cases, to clearly communicate their meaning, students may choose to write comments that could disclose their identity. For example, some experiences, by their very nature, are entwined with a student’s gender, race, area of research, role as a parent or student leader, or other identifying information. In those situations, the student’s feedback won’t be completely lost; the full comment will still be considered in thematic analyses, but it won’t be reported in full text outside of the Graduate School.
I hope the explanation I’ve provided will alleviate concerns about your DGS being able to identify you from your written comments on the survey. If after considering this additional information, you would like to withdraw or edit your comments, please contact me (email@example.com) directly to make those changes. The survey will close for data collection on May 10 and the Survey Research Institute will transfer the final dataset to me two weeks after that. I expect that the quantitative results from this year’s survey will be available to DGSs and GFAs in late June, with the comments (after redaction) available in August.
Anne M. Laughlin
Graduate School Director of Assessment