Structured Listening

Today’s Agenda

• What a number cannot tell you
• Immersion, and structured listening
• Three modes of watching
• Field notes, and the rhythm that keeps them

Three Numbers, One Climb

Sodapoppin, November 2018

• 27,934, then 28,076, then 28,203 concurrent viewers
• Two hundred sixty-nine viewers in 122 seconds
• The dataset records that climb to the second
  • What it cannot tell you is what the climb was

Same Rows, Four Events

• A raid: another streamer sending their audience over
• A viral clip on Reddit, pulling in strangers
• The streamer returning from a break
• A game update, or a good night accumulating

The row looks identical in all four cases. Today is about how you come to know which one it was.

Signatures You Learn by Watching

• A raid has a signature
  • A near-vertical jump in the count
  • A wave of near-identical greetings in chat
• A viral clip has a different one
  • A slower swell of arriving viewers
  • They do not know the room’s conventions
• You bring the labels, and only if you looked

Immersion and Structured Listening

• Immersion: sustained attention before analysis begins
  • Here it takes the form of structured listening
  • Watch, read chat, take notes, with an ethnographer’s discipline
• The data is historical, the platform is live
  • 5.5 days in 2018, and the site is open right now
  • The behaviors it recorded are still there to watch

Skipping It Costs You

A coder who has never watched Twitch reads the log as plain text and:

• Sees LULW as a typo or a username
• Sees thirty identical posts, not copypasta
• Sees “@sodapoppin same” and cannot tell agreement from reflex
• Sorts everything confidently into buckets that mislead

Thick Description

• Thick description captures meaning, not only events (Geertz, 1973)
  • Rich enough to say what it meant to those present
• Thin counts messages, thick knows the count
  • Raids, inside jokes, and reflexive replies
• Structured listening is how thin becomes thick

The medium changes and the discipline does not.

Mode One, Casual Watching

• Initial exposure and nothing more
  • Open several streams across the range
  • A large gaming channel, a small one, Just Chatting, art
• Do not take notes, and do not code
  • Let the medium be unfamiliar
  • What surprises you, and what do you not understand?
• The goal is exposure replacing assumption

Mode Two, Focused Watching

Choose a few streams and watch with one dimension in mind per pass:

• Chat behavior: pace, who it addresses, when it changes
• Viewer trajectories: climb, hold, or fall, and why
• Host moves: read aloud, react, ignore, or steer

Brief notes after each pass. You are still describing, not coding.

Mode Three, Analytical Watching

• Watch more streams, asking what recurs
  • Do chat behaviors cluster with kinds of stream?
  • Does competitive-game chat differ from art-stream chat?
• Which cases resist easy description?
• The contours of your coding scheme show up here

Field Notes Exist or Not

• Observation without a record is watching television
• Field notes are written, dated, and specific
  • Not polished prose, but thinking on paper
  • They catch an observation while it is still fresh
• The notes do not need to be good, they need to exist

Observational and Methodological Notes

• Observational: what you saw
  • “Chat moved in bursts, near silence while they talked”
  • “Then a flood every time they paused”
• Methodological: a measurement problem
  • “The bursts are answers, and performance for the room”
  • “A coder reading text cannot see the question”

Theoretical and Comparative Notes

• Theoretical: links an observation to a framework
  • “The burst pattern fits uses and gratifications (Katz et al., 1973)”
  • “Chat volume responds to host moves, it is not a trait”
• Comparative: how cases differ
  • “Gaming chat ran fast, art chat ran slow”
  • “They are not the same object”

The Rhythm

• One plain-text document, dated, one entry per session
  • The Hub template carries the four note types
• A note every fifteen or twenty minutes
  • The record has to keep pace, because memory will not
• Your codebook gets assembled out of these notes

Observation Tests Your Question

• Your prospectus commits you to some distinction
  • Observation will confirm it, or complicate it
• The classic case is directed against broadcast
  • Watch a while and you find messages aimed at neither
  • Replies, arguments, and inside jokes between regulars
• A two-way split gains a third term

That is not a failure of your prospectus. It is the cheapest possible moment to find out.

Before Next Time

• Due this week: Topic Selection and Research Questions
• Next week: the Project Prospectus
  • Submit the version you can defend, not the one you love
• Start your field-notes document tonight
  • Template on the Hub, dated entries, four types
  • Log two observation sessions, one gaming and one not
• Read the rest of Chapter 7

Questions?