The Prospectus and Pre-Registration

Today’s Agenda

• The six components of a prospectus
• Two worked prospectuses off one corpus
• What the draft does not need yet
• A failed draft, and the feasibility test
• Pre-registration, and what it binds

One Page, Six Decisions

The prospectus

• What you hand in: one page, next week
  • The topic, the reason, the method
  • Chapter 6’s model runs 250 words
• A diagnostic, not busywork
  • Cannot state it briefly, cannot start it
  • Better found now than in November
• You leave knowing which piece you cannot write

The Six Components

• Title and questions: one line, then one to three
  • Not ten questions
  • The title hints at the key variables
• Framework and gap: two to three sentences each
• Method and contribution: three to four, then one to two

Title and Question

Model prospectus

• Title: directed and broadcast chat, gaming versus not
• RQ: does the directed proportion differ between them?
• What the pair buys you
  • The title names the variables
  • The question names the comparison

Between them they have already ruled out most of what you could have done, which is the point.

Framework and Gap

• Framework: uses and gratifications (Katz et al., 1973)
  • Audiences pick media to satisfy needs, social ones included
  • How viewers participate is one observable trace
• Gap: self-report surveys dominate this literature
  • (Sjöblom & Hamari, 2017; Hilvert-Bruce et al., 2018)
  • Whether chat traces back them up is unexamined

Method in Four Sentences

• Sample and unit: 1,500 messages, stratified
  • Balanced across gaming and non-gaming
  • The unit of analysis is each message
• Coding and reliability: three categories, one check
  • Directed, broadcast, or unclassifiable
  • Second coder on 10 percent, Krippendorff’s alpha
• Then the directed proportion gets compared

Same Data, Different Theory

• Title: parasocial markers, large audiences versus small
• RQ: do smaller audiences show more parasocial language?
• Framework: parasocial interaction (Horton & Wohl, 1956)
  • Smaller-channel viewers are the more socially motivated ones
  • (Hilvert-Bruce et al., 2018)

Same corpus. Different study entirely.

The Second Method Paragraph

• Sample: 1,000 messages, top and bottom quartile
  • Quartiles by average viewer count
• Three markers, coded present or absent
  • Second-person address, familiarity, emotional disclosure
• Reliability and comparison
  • Second coder on 10 percent, Krippendorff’s alpha
  • Each marker’s rate compared between groups

Pick the Workload

• One variable against three
  • The first prospectus codes one, the second three
  • Three means three codebook decisions and three checks
• The better study is the heavier term
  • Roughly three times the coding hours
  • Neither is more correct, and both are defensible
• Make the trade visible now, not in November

What You Do Not Need

• The prospectus is pre-statistics work
  • No test, no p-value, no R code yet
  • “Will be compared” is all the method needs
• What you commit to is structure
  • Question, theory, unit, coding scheme, comparison
• Waiting for the statistics is not a reason

A Prospectus That Fails

Title. Twitch and Society.

RQ. How does Twitch chat affect viewers’ sense of community, and what is the role of parasocial relationships and social identity across different streams and genres? Also, are streamers becoming more popular over time?

Framework. Parasocial interaction, social identity, and political economy.

Method. I will look at some chat logs and see what patterns emerge.

Why It Fails

• The title and question fail first
  • No variables are named
  • Three or four questions stacked together
• The theory and method cannot be judged
  • Three theories pull in different directions
  • Nothing in the method is specific enough
• The contribution could be pasted into any project

That draft was not lazy. It was ambition before discipline.

The Feasibility Test

• Can you get the data, and analyze it?
  • The working corpus ships inside the v2v package
  • Fifty channels is manageable, 1,690 is not
• Does the method match the question?
  • Content analysis reaches content, never why someone felt
• Have you controlled scope creep?
  • Narrow until it hurts slightly, then narrow again

Prospectus as Pre-Registration

• Pre-registration commits the plan before the data
  • Hypotheses, methods, and analysis plan, publicly
  • The Open Science Framework timestamps it
• Without a fixed plan, the plan drifts
  • Researchers adjust until something crosses significance (Simmons et al., 2011)
  • The source of researcher degrees of freedom
• It closes them by making the plan inspectable

Your Prospectus Is Your Record

• It records what you planned
  • Written before any coding or analysis
• It is what you check yourself against
  • When you are tempted to chase a surprise
  • A gap is visible, not automatically wrong
• Chapter 2’s transparency argument, aimed at your project

Before Next Time

• Due: Topic Selection and Research Questions, this week
• Next week: the Project Prospectus
  • Draft all six components badly, then cut
• Read Chapter 7 on structured listening
• Watch two streams, one gaming and one not
  • No notes yet, just watch

Questions?