An Introduction to Oral History Transcripts and Transcription

Special thanks to Krystl Raven, Peter Scales, Tarisa Dawn Little, Jesse Thistle, Osgoode Society Oral History, Alexandra Giancarlo, Peter Anderson, Janis Thiessen, Shirley Tillotson, Alex Green, Andrea Blackman, Sandra (@khassl), and Terry Smyth for their recommendations and advice, and Jessica Knapp for her feedback on a draft of this blog post!

As you may remember from my interview with Katrina Srigley, Stacey Zembrzycki and Franca Iacovetta, I am an oral history devotee. I also used oral history extensively in my dissertation, thirty-five interviews in all. But as an oral historian, I have always had an uncomfortable relationship with transcripts and transcription. So today I thought I would unpack some of the existing discussion around transcription. Let’s get started!

What is Transcription?

First of all, a definition is in order. A transcript is basically the text version of an oral interview, written out word-for-word. Transcription is the process whereby a transcript is created. And within those two definitions, there is a whole world of complication. But before I get into that, it’s important to know that creating a transcript for each oral history interview is standard practice in most cases.

There are several types or levels of transcription, and which one you choose will often depend on what you plan to do with the transcript once you are finished. There are three basic forms. First, there is Intelligent Transcription. In this form, the transcriptionist will often edit out words like “uhh” and “um” as well as false starts, and will do some basic editing to ensure that sentences are grammatically correct. This type of transcription is mostly used within the medical or business fields, often when the material being transcribed is a dictation. Second, is the Edited Transcription. This form is an Intelligent Transcription, but with more editing of the text (correcting sentence fragments, removing contractions). This form is often used when the transcript will be published, like in a newspaper. Finally, there is Verbatim Transcription. In this form of transcription, everything is included, such as filler sounds, (“uhh or “um”), pauses, interruptions, laughter, etc.. This type of transcription does not feature any editing at all. Verbatim transcription is the preferred format for historians, since it captures the most nuance.

Why Transcribe?

As I noted above, transcripts are standard practice in the field of oral history. There are some important reasons why this is the case. Probably the most important reason is that many feel that this is an essential part of in the historical analysis of an interview. As Terry Smyth noted in the Twitter thread I started on this topic, “bear in mind that this is an important part of the research process, not a burdensome admin task. Keep a notebook open by your side so you can make notes immediately … the data just keep coming, so don’t lose any! Good luck!” There is much that can be gained simply from sitting and listening to the interview over and over again, and transcription can be a good way to both reflect on the content of the interview and begin your analysis.

Second, transcripts are generally considered easier to use than the original recordings. Most researchers prefer to use transcripts, because they are a quick and simple way of mining an interview for relevant information. Consider that many interviews can be between one and two hours each, versus the time it takes a researcher to read through a transcript. Not to mention that listening to an interview requires a significantly higher degree of concentration over a long period than reading does (at least for most academics). Transcripts also require considerably less equipment and/or technical skill, especially when considering that audio-visual technology goes out of date quickly (some of my younger readers may not even know what a cassette tape is. ::cries::); as a result, older equipment may not be maintained, available, or easy to use, and the sound quality of the recording may be degraded. On the flip side, not everyone is technologically savvy, and many folks (of all ages!) find newer technology to be alienating and overly complicated. In comparison, paper, or even a Word document, seems much safer. Finally, transcripts are essential for ensuring that audio-visual material can be made accessible to Deaf and hard of hearing folks.

Third, transcripts can be important for preserving the anonymity of your interviewees. In many cases, interviewees will sign release forms before they begin an interview, and many of these release forms specify that the interviewee will not be identified by name or personal information in any resulting publication. If you intend to donate your material to an archive, it’s important to consider whether or not future researchers can or should have access to the original, unedited recordings, which may contain names and personal information. Creating a transcript is a useful way around this problem, making the interviews accessible while still protecting your interviewees.

Finally, transcripts can be essential for doing some types of historical analysis. For instance, in recent years, more and more historians are using software like NVivo that can aid in quantitative analysis. For instance, NVivo can track how many times particular words or phrases are used. This can be helpful not only when looking for major themes, but also when determining how many of your interviewees attended university (to give you just one example). While I’ve never personally tried NVivo before, I know many historians who are big fans of the software!

The Limits of Transcription

While there are several good reasons to use transcripts, there are some important limits to keep in mind. First, transcripts, no matter how accurate, are always partial and incomplete. Speech does not translate exactly or easily into written language. I use the term ‘translate’ here specifically. We often think of English as a single language, regardless of the form that it takes. But written and spoken English are very different, both in terms of grammar and word usage. What’s more, speech isn’t simply about the words that come out of one’s mouth; it’s also about subtle cues like vocal tone, inflections, expression of emotions (laughing, crying), and even body language. And the written language simply cannot capture these nuances. That’s why text messages and emails can be so fraught. And alas, I don’t think anyone uses emojis in transcripts (though maybe they should!) Of course, the reverse is also true. I’m not sure how one would communicate a semi-colon in spoken form, though I would be curious to know if anyone has tried. But my point is that a transcript is still only one translation or interpretation of an interview.

Second, creating transcripts is extremely time consuming and costly. Depending on the type of transcript, you may be requiring anywhere up to eight hours to transcribe a single hour of recording (roughly 10 to 15 minutes of transcription for each minute of audio). If you are doing the transcribing yourself, this is a considerable time investment, and, depending on the number of interviews you are using in your project, can take a year or more. If you are hiring someone to do the transcribing for you, this can be exceptionally expensive. Do not underestimate the cost of transcription and the amount of time it takes. It is always longer than you expect. A Canadian transcription service, Transcript Heroes, puts existing rates at anywhere between one to two dollars per minute. If you consider that many interviews, as I said before, are around an hour long, that can be $60 to $120 per interview. And the price can increase depending on how detailed you want the transcription to be, how many speakers there are, whether the speakers have accents, etc… So please keep this in mind when budgeting!

Third, the field of oral history can be exceptionally inaccessible for folks with disabilities, and requiring that students or researchers create transcripts is a major reason why. Simply put, the lack of accuracy with respect to existing voice-to-text (or speech recognition) software is a major barrier for folks who require access to assistive technology. As I note below, even the best of these services still have major problems with inaccuracy. Physically transcribing is not any easier, since it requires considerable dexterity and/or the ability to use a foot pedal.

Finally, it’s important to consider the impact of the transcript on the orality of the format. In response to my query for best practices on transcription, Katrina Srigley noted in an email, “There are also those who would ask this question – is transcription required?! Sitting and listening deeply to the recording and finding ways for participant review through listening are also powerful approaches to working with recorded material.” Many scholars have argued that in creating transcripts, we are betraying the spirit of oral history. Oral history is different from other types of historical research by virtue of the format. The oral component of the interview is one that results in unique information that cannot be found in textual documents. What’s more, the reality is that, for the most part, once scholars have access to a transcript, they rarely return to the original recording. And as I noted above, a great deal of information can be lost as a result.

A Happy Medium?

It’s important to note that transcribing or not transcribing are not your only two options when it comes to processing oral histories. For instance, Tarisa Dawn Little noted, “My recommended approach is re-listening to recording and jotting down time sections to go back to. I don’t transcribe the entire interview because that is way too tedious. Often, I’ve transcribed the sections that I feel are most relevant to the study’s focus.” Personally speaking, I chose to instead create indexes of my interviews. What this means is that, as I listened to the interview, I noted down which topics were discussed, and included timestamps. Alternatively, the Centre for Oral History and Digital Storytelling also recommends the creation of chronologies, which are basically like indexes, but the questions are transcribed. As Eve-Lyne Cayouette Ashby noted, this can be really helpful if you plan to ingest the interviews into a database, and want to be able to easily search by question. While these methods are considerably faster, they are still time consuming; each hour of recording can take up to three hours to index.

Transcription Best Practices

General

• Make or use an transcription guide or template to ensure that the formatting, spelling, punctuation, and how you will document pauses, interruptions, and emotions are all done uniformly. You can create your own, or use existing guides and templates. Here are some examples:
  • Baylor University Institute for Oral History Style Guide (PDF)
  • Columbia Centre for Oral History Research Oral History Transcription Style Guide
  • Centre for Oral History and Digital Storytelling at Concordia University Transcription Model (PDF)
  • Centre for Oral History and Digital Storytelling at Concordia University Chronology Model (PDF)
• Keep future readers/listeners in mind. Make sure that the transcript includes as much detail about the recording as possible (like the names of the interviewer and interviewee, time, date, and place, etc…) This is really important contextual information and, without it, transcripts can be significantly less useful.
• Make sure you go back and proofread the transcript, regardless of whether you transcribed yourself, or someone else did. This is the only way to ensure accuracy. Even the best transcriptionists will make mistakes, especially in portions of the recording that are very difficult to understand.

Do It Yourself

• Software that allows you to slow down the recording and/or allow for playback (like ExpressScribe) will make transcription easier.
  • If you plan to do a lot of transcription, you may want to consider investing in a foot pedal, if possible.

When Using a Transcription Service

• Consider providing a word list if you are hiring someone to do transcription for you. For example, many of my interviewees would use Yiddish and Hebrew words, but not everyone is familiar with these. A word list can be extremely helpful in these situations.

Important Questions to Consider When Transcribing

General

• Which type of transcript will be most useful for your particular situation?
• How much of the interview would you like to have transcribed? Do you need to transcribe every word, or just the parts that are more relevant for your research?
• Will the transcriptions be made public or will they be for personal use only? If you will be the only person who is using the transcript, then you may not need to spend too much time cleaning them up and making them readable.
• Who will have access to the transcripts once they are finished?
• How will you ensure the privacy and anonymity of your interviewees? If you are using a transcription service, will the transcriptionist sign a release form or a non-disclosure agreement?
• Are you planning on posting audio-visual clips of your interviews online? If so, consider that transcripts are extremely important for ensuring that these clips can be accessed by Deaf and hard of hearing folks.

When Using a Transcription Service

• Should I use a transcription service?  Online or automatic transcription services are notoriously unreliable, as many have noted. They work best when the audio is exceptionally clear, there are no background noises or interruptions, and the speaker does not have any kind of accent. And, in almost every case, they will require extensive editing. So in the end, they may or may not actually save you any time. However, most of the people I spoke with who used them did find that, in general, they were very helpful.
• Where is the labour coming from? Often the reason why online services are so inexpensive is because pay extremely poorly. Some also rely on off-shore transcriptionists who are paid far less than North American workers.
• Where will the documents be stored during the transcription process? Remember, many Canadian universities now require that personal information is stored only on servers located in Canada.
• Which file formats are acceptable for use? Can you decide which file format you would prefer for your transcript?

When Using Transcription Software/Apps

• How long are you planning to use this software or app? As I noted before, technology changes fast. This means that the software or app that you are using may go out of date very quickly, or vanish all together. For instance, a really promising transcription program, Pop Up Archive, shut down in 2017. Not only did this mean that the service was no longer available, but users had to ensure that they downloaded all of their files before they were lost. Remember, software and app developers are not required to maintain their programs for your convenience.

Transcription Options Overview

These are the most popular options that are most recommended by oral history programs/centres across Canada and the United States.

Voice to Text

• Dragon NaturallySpeaking
  • Platforms: Mac or PC
  • Cost: $150 USD
  • How it works: Essentially dictation software – you talk and it translates your speech into text.
  • Pros:
    • Most reliable software currently on the market
    • No privacy concerns since the transcription is done on your computer
    • Format of the audio file doesn’t matter, since it is played on your computer as normal
    • Text can be easily copied into Word
  • Cons:
    • Expensive
    • It will work best when only one voice is used
    • The software requires training (ie, the more you use it, the greater the transcription accuracy)
    • For best results, you will need to repeat the interview word for word in your own voice
      • Can be used with ExpressScribe to make the process easier (slowing down the audio to make it easier to repeat what is said)
• Other recommendations:
  • Otter (a note-taking app for your phone)
  • Evernote (information management software)
  • Youtube (can generate a transcript if you upload the audio)

Transcription Aids

• ExpressScribe
  • Platforms: Mac or PC
  • Cost: Free
  • How it works: This is a program that allows you to slow down the audio recording to make it easier to transcribe.
  • Pros:
    • Inexpensive
    • Can be used with Dragon NaturallySpeaking (to help improve the accuracy of the transcription)
    • Works with all of the major file formats (including MP3, .wav)
    • Works with Word
  • Cons:
    • Bit of a learning curve
    • Designed for professional transcriptionists
    • Does not create a transcript for you
  • Online Reviews/Tutorials
    • http://www.oralhistorycentre.ca/introduction-express-scribe

Online Transcription Services

• Temi.com
  • Cost: Free for the first file, 10 cents per minute afterwards
  • How it works: You select the audio file you want (works with all major file formats), and upload it to the service. Their computers will then produce a transcription within a few hours (depends on the length of the audio file)
  • Pros:
    • Inexpensive
    • Easy to use
    • Transcripts can be downloaded in all major file formats, including Word
    • Free trial
    • Works well with clear audio
  • Cons:
    • The file is stored temporarily on their servers, which might be a privacy concern. However, you can delete the audio file once you are finished with it
    • Does not work well with recordings featuring multiple speakers, strong accents, or background noise

• Rev.com
  • Cost: $1 per minute (extra 25 cents per minute for time stamps)
  • How it works: You upload your audio file (works with all major file formats) just like Temi, except the file will be transcribed by a professional transcriptionist
  • Pros:
    • Available around the clock
    • Quick turn around (12 hours or less for files under 30 minutes, and can pay for rush service)
    • Extremely accurate
    • Transcript can be downloaded in all major file formats, including Word
  • Cons:
    • The file is stored temporarily on their servers, which might be a privacy concern. However, like Temi, you can delete the file when you are done
    • Confidentiality is a concern. However, all transcriptionists sign confidentiality agreements with the company, and can sign an NDA upon request
    • Can add up
• Other recommendations:
  • Trint (automated transcription service)

For additional information, you can check out:

• H-Net OralHist
• Oral History in the Digital Age
• Oral History Association
• Baylor University Institute for Oral History
• Concordia University. Centre for Oral History and Digital Storytelling
• University of Winnipeg Oral History Centre
• Columbia Centre for Oral History Research
• Archives of Lesbian Oral Testimony
  • See this blog post for more details on what ALOT can offer.

Further Reading:

• Willa K. Baum, Transcribing and Editing Oral History (Walnut Creek, CA: AltaMira Press, 1977).
• “Indexing and Transcribing Your Interviews,” Library of Congress Veterans History Project, April 3, 2009. http://www.loc.gov/vets/transcribe.html
• Elinor A. Maze, “The Uneasy Page: Transcribing and Editing Oral History,” in Handbook of Oral History, edited by Thomas Charlten et al (Lantham, Maryland: Rowman Altamira, 2006), 237-269.
• Doug Oard, “Can Automatic Speech Recognition Replace Manual Transcription?” in Oral History in the Digital Age, edited by Doug Boyd, Steve Cohen, Brad Rakerd, and Dean Rehberger. Washington, D.C.: Institute of Museum and Library Services, 2012, http://ohda.matrix.msu.edu/2012/06/automatic-speech-recognition/
• “Principles and Best Practices for Oral History,” Oral History Association, October 2009, https://www.oralhistory.org/about/principles-and-practices-revised-2009/
• Raphael Samuel, “Perils of the Transcript,” in The Oral History Reader, edited by Robert Perks and Alistair Thomson (New York: Routledge Press, 1998), 389-392.
• Linda Shopes, “Transcribing Oral History in the Digital Age,” in Oral History in the Digital Age, edited by Doug Boyd, Steve Cohen, Brad Rakerd, and Dean Rehberger. Washington, D.C.: Institute of Museum and Library Services, 2012, http://ohda.matrix.msu.edu/2012/06/transcribing-oral-history-in-the-digital-age/
• “Webinar 2 – Interview Transcription,” Concordia University Centre for Oral History and Digital Storytelling, http://storytelling.concordia.ca/content/webinar-2-interview-transcription

Oral history can seem very intimidating, but I hope this guide has made it a little bit less so. While this guide was primarily designed for historians, it should also be helpful for anyone interested in oral history, such as genealogists and journalists. I hope you enjoyed this week’s blog post! If you did, please consider sharing it on the social media platform of your choice. If you would like to see more blog posts on oral history, please let me know in the comments below or on social media! And don’t forget to check back on Sunday for a brand new Canadian history roundup. See you then!

Liked this post? Please take a second to support Unwritten Histories on Patreon!