What Ethnographers Do: Field Notes

Field notes, for the ethnographically uninitiated, are detailed descriptions of field work (interviews with and observations of people).

Contrary to what the name suggests, field notes are not actually taken in the field while we are interviewing and observing people. Rather, while we are in the field we take scratch notes. In our scratch notes we write down things people say, make sketches of rooms, write down the time things happen, keep tallies of how many times that person has kissed a hamster in the last fifteen minutes, and a host of other information that might be useful to our research.

Once we leave the field and are back in our own homes or offices, we turn the scratch notes into field notes. This is a somewhat arduous and painful process, that involves turning things like, “6pham danc NW g 4:16-5 O” into a meaningful description like “At 4:16 pm, there were six people dancing in the northwest corner of the green, each holding small golden hamsters. They danced in a circle for five minutes.” It is best done as soon after the field work as possible, so that the appropriate details can be added in. Memory decays quickly, and then gets fuzzy and confused.

In traditional anthropology, field notes are not things that you share. They are, in many ways, considered the secret formulaes of anthropological research. Most researchers are loathe to share them. As a result, there is little standardization in how field notes are taken or taught, which creates a great deal of difficulty for an up-and-coming ethnographer. This year we have largely been left to our own devices in developing a methodology for writing field notes.

Old Methods

I have experimented with several different methods.

The first, and most natural to me as a writer, is to turn my scratch notes into a narrative. I write it up much like an interview in a women’s magazine. Things like,

“Eleanor plays with her coffee as I sit down across from her at the diner. She appears to be stirring it with a French fry, and looks up, eyes wide, when I accidentally drop my cell phone with a clatter on the table. She hates cell phones, she says.

‘Every time my cell phone rings, it’s someone wanting something. I would love to smash it with a hammer.’ She laughs, throwing back her head, her blonde ringlets shaking.”

I like this method for a couple of reasons. It provides a more cohesive, enjoyable narrative, for one. They’re much nicer to read this way. Further, they’re much nicer to write.

However, they are also somewhat less accurate depictions of an interaction, and I have to be very careful not to put in judgements. For instance, I wanted to write “and looks up, eyes wide startled, when I accidentally drop my cell phone” and “blonde ringlets shaking with mirth“. However, those additional words, which I would use so casually in my own writing, are not descriptions. They are things I have intuitively decided about the situation, based on my long period of living in the world. When someone’s eyes are very wide and their face is pale, that generally denotes some kind of shock or surprise. She is laughing, therefore she is mirthful. However, these are not necessarily accurate interpretations of a situation, and are certainly not uniform across cultures. Maybe her eyes are wide because she replaced her mascara with superglue this morning. Maybe she is laughing not with mirth but with barely surpressed rage which will shortly cause her to Hulk out and kill everyone in the cafe. Ethnographers are supposed to record what is, not record interpretations of what is. (This goal is clearly not entirely possible, as everything we observe is interpreted by our brains to some degree; however, it’s good to try.)

The second method that I have tried for field notes of interviews is direct transcription, in a question and answer format. I would record everything that I asked, and everything the participant said, word for word.

The primary disadvantage here is that is is excruciatingly painful and time consuming to transcribe these things. (I was once so desperate not to do it by hand that I recorded an interview, then called my number on Google Voice, which transcribes voicemails, and played the interview for the Google Voice voicemail in hopes that it would transcribe it for me. This was a tremendous failure.) I would often get so bogged down in getting every word recorded that I would miss paying attention to the ideas behind what the participant was saying.

The New Method

Alicia and I are trying a new and different method, utilizing the power of spreadsheets.

This is our new template:

And here is our template in use (participant’s name has been changed to protect his privacy)

We fill these in using a three-pass process.

• Pass 1: Scratch notes taken in the field
• Pass 2: Fill in the first three columns on the spreadsheet
• Pass 3: Tag each row with primary, secondary, and tertiary tags.

If both Alicia and I were present for an interaction, we would have two versions of these notes–one that she had done, and one that I had done. They will both be in the same file, on different tabs (or sheets, as they’re called in MS Excel) labeled by our initials.

There are several advantages to this approach.

Uniformity:

Writing field notes the same way every time makes it easier to write, read, and process them.

Sorting:

The tag columns allow us to read the notes chronologically (as they are written) or sort by tag and read them by subject area.

Merging:

Because they are uniform and in column format, it’s much easier to merge the notes of two researchers (or maybe even more!

Quantitative/Qualitative Fusion

Spreadsheets can easily be exported to CSV files for text mining.

Occurences of words or tags can be computed easily, so that each field note file can be tagged for relevancy. For instance, let’s say a file had 20 tags, 10 of which were about hamsters and three were about llamas. That could be tagged with a 50% hamster relevancy and a 15% llama relevancy. Then, if you were later doing a project in which you needed information about hamsters, you could search through all of the field notes of all of the projects you’ve ever done, and say, “Return all field notes with greater than a 30% hamster relevancy.”

Over time, a machine learning algorithm can learn what lines of field notes you are tagging what way, and you can develop an auto-tagger.

Easy Analysis

Most importantly, it allows for easier analysis. When you are ready to analyze the notes, you simply output everything below row eight to a single file for all of the field notes, then sort by the tags. In short, it removes the affinitization step that takes so long in traditional ethnographic analysis.