Memes of Resistance, Election Reflections, and Voices from Drug Court: Social Justice, Ethical Cataloging, and Digital Humanities at Utah State University

Editor's Notes

1. As a metadata cataloger, we are charged with creating descriptive data to objects that usually fall well  out of our expertise.  My own personal lack of information include  logging, irrigation, Utah history-- but, in most cases, the biggest challenge is to try and find names or places to provide context to a photo rather than just “logger in the forest."  I also have the luxury of turning to our University Archivist to help fill in some of those information gaps.  Up until now, I have rarely been presented with the additional challenge of providing descriptive metadata to collections that highlight the underrepresented groups. 
2. My goal was to create descriptive met creating descriptive metadata for drug court was to apply sensitive terminlogytoward this group people.
3. Again, not being an expert on the subject matter, I could, however, rely on the Library of Congress Subject Headings but the average searcher does not always use these standardized terms. I wanted to avoid inserting my own jargon or “slang” so I relied on the curator, Randy Williams, who became immersed in the drug court culture to help fill in my information gaps.  Randy also gave quick terms which I would have had to listen to the audio or review the transcripts to find.  Randy even took it another step further in consulting with a former participant of drug court to refine some of the search terms.
4. The Election Reflections collection that Randy spoke about provided a different set of issues that we needed to consider when we addressed representation in cataloging and metadata.  Unlike collections that are gathered in person, the process for gathering this collection was one step removed and mediated by an electronic survey rather than a human.  I am breaking down our considerations into three quick concepts Consent, context, consistency, and concern.
5. Participants were asked to fill out the survey questions with the understanding that they would be used preserve the sentiments and reactions of citizens to the recent election. This was followed by a series of questions about the person, such as what their political affiliation was, their demographics, and how they would describe themselves. Information from each respondent was then transformed into a PDF and loaded into a digital collection.
6. The entirely digital method of this process (from data collection to the eventual digital collection) allowed us to act in real time and take an intriguing snapshot of northern Utah sentiment during a interesting political time. But it also provided some challenges, since we needed to think through the ramifications of what we were asking and what we intended, particularly as there was no person to explain to participants what would be done with their information. We (Randy) thought through IRB implications and we all combed over the questions we would like, using pre-existing standardized options such as the census records for demographic data. However, we went through the responses, a couple of things occurred to us that we needed to carefully consider as we put the collection online. The first and foremost was consent. There were multlple places within the survey that notified participants about what would be done with the information they supplied. The survey began with a n explanation of the project and ended with a consent form. But even with this formality in place, we tried to be explicit with the information that would be included and excluded. For instance, the survey asked respondents for information such as their name.  Knowing that respondents might not understand the ultimate goal for their collections, the survey asked respondents to both provide their real name and also the name they wish to be associated with their reflection.  It was carefully noted that the first name they entered would not be used in the collection. Some respondents put the same content in both fields, others put in funny or exasperated names.  Neither field was required, so some respondents had nothing at all.    Now for the Context. Seeing how this information is used in the final product might have benefit participants, so they could ultimately decide how their reflection would be portrayed to the public. But as this was our first try at this type of interaction, we had nothing to show them what their reflection would look like. Take for instance, the difference between these two reflections from Dustin and Gobsmacked yet hopeful. And also consider their self-descriptions they supplied. Would they have supplied this if they were not entering characters into a random text box? Critical cataloging asks us to examine our practices and empower those who are described to have a say in how they are depicted. During our next iteration, we canl have samples to show the final look of the PDFs, which may empower participants to control how they want their reflections to be perceived
7.   Another interesting issue that came up was how to preserve the information as closely as possible to the way it was supplied – or Consistency. We reviewed the entries to make sure that they were all coming across as they were entered in Qualtrics (the original survey software).  We noticed a few abnormalities such as larges string or paragraphs of strange characters.  We believe this was caused when respondents copy and pasted text from other places such as Facebook or Twitter.  The CSV export also did not preserve paragraph breaks in the text and after reviewing the differences, we felt it was important to preserve those, as it affected the meaning of the entries. While most of the process was built on automation and batch processing, this parttook a bit of time because we had to go back into Qualtrics (which retained the formatting) and verify that entries came across as intended. For some free text fields, such as location and place names, we did standardize entries to help facilitate searching later on. This included capitalizing place names, correcting the spelling of obviously mis-spelled words or moving them from one field to another.  (For instance, some respondents put "Logan" in the State field instead of the City field.)  We only did this for locations we were absolutely certain of.   One intriguing issue came up with Consistency was the idea of embracing In-Consistency. When we sent out the survey, we neglected to standardize the Political Affiliation with a controlled vocabulary where participants could choose from a set list of responses - and instead, allowed anyone to enter poltical affliations in any way they saw fit. This lead to a nightmare for faceting (as seen in the example in the long column on this slide) but also became an interesting case study in the complexity of human perception of their own political motivations. We decided to not include a controlled vocabulary in the future. Lastly, I’ll just speak to our final Catloging ”C” - Concern. Since our survey was a mediated interaction with participants, we wanted a set of human eyes to review all of the entries for red flags such as threats of harm or self-harm. We made a strict point of not censoring any responses for content or language. But we wanted to ensure that we didn’t overlook areas of potential concern, particularly for such a emotional election. In the end we only restricted one entry, simply because it was a URL that pointed to a personal Facebook page post.  It was felt that there was potential for exposing the respondent in ways the respondent did not anticipate. There you have it – our four “C” of ethical cataloging with mediated digital collection. Ensure consent, Provide Context, Be Consistent (and sometiem In-Consistent) and be Concerned.