The Publisher: White Paper and Poster Studio

Today’s Agenda

• A report that rebuilds itself
• IMRaD, and the three transparency markers
• Two studio blocks: the paper, then the poster
• Publishing, and the checks that run first

The Last Meeting

Studio session

• Two working blocks: the White Paper, then the poster
  • Short input from me between them, the rest is yours
  • I am circulating all evening
• What we build today is what you submit
• Leave with a rendered site, a layout, and a reflection

Done, and Not Finished

• The study has every hard part behind it
  • A question, a theory, a reliability-tested codebook
  • A defensible sample, a clean dataset, three figures, a finding
• And it lives in pieces
  • A codebook here, a script there, output that scrolled past

A study scattered across a dozen files is raw material. Today gives it a form.

A Report That Rebuilds Itself

• The ordinary way is quietly fragile
  • Run, copy numbers, screenshot figures, write around them
  • Change one rule and every pasted number is stale
  • Nothing in the document says so
• The alternative: put the analysis inside the document
  • A Quarto document interleaves prose with live code chunks

Numbers You Never Type

The analysis dataset held `r nrow(analysis)` coded messages,
of which `r sum(!is.na(analysis$is_gaming))` carried a
gaming-status label.

In the rendered report those backtick expressions become the actual counts, so nothing in the sentence can fall out of date.

Figures You Never Paste

#| label: fig-msglen
#| echo: false
#| fig-cap: "Message length by channel type."

ggplot(msglen, aes(message_length, fill = is_gaming)) +
  geom_histogram(binwidth = 5) +
  v2v::theme_v2v()

Those lines sit inside an {r} chunk in the report, so the figure is produced at render time rather than pasted in.

Two Payoffs That Matter

• The report cannot drift, because it is the analysis
• The report can be checked
  • Anyone with the source and data renders the same results (Peng, 2011)
• The idea predates Quarto and R
  • Runnable code and its prose in one document
  • Knuth named it literate programming in 1984

The “one click” is the render button.

IMRaD, and What Each Section Owes

• Introduction: the question and why it matters, your prospectus
• Methods: what was done, in enough detail to repeat
  • Codebook, sampling, reliability, wrangling
• Results: what was found, without interpretation
  • Figures, means, the statistic with df, p, and effect size
• Discussion: what it means, plus the honest limits

An abstract in front, references at the back.

The White Paper, Specifically

• IMRaD structure, APA 7th edition throughout
• Effect sizes and power statistics, not only p-values
• A pre-registration disclosure in the Methods
• A Discussion that situates the finding in your literature

Converting results into IMRaD costs something. Ask what is lost, and what must be added.

Methods, Three Transparency Markers

1. A data provenance statement: where the data came from, and how
2. An intercoder reliability report: coefficient, protocol, threshold
3. The pre-registration disclosure, with its OSF URL

Each signals that the study meets the standard it claims. Anything beyond the plan is labeled exploratory.

What Preregistration Is For

“Preregistration distinguishes analyses and outcomes that result from predictions from those that result from postdictions.”

Nosek et al. (2018, p. 2600)

• Not a promise of virtue, but a way of making a distinction legible
• Draft your disclosure sentence today
  • List your deviations, and say how you report each

Where the 200 Points Sit

• Introduction and framing: 25
• Methods: 55
• Results: 40
• Discussion: 40
• Abstract and references: 15
• Reproducibility and publication: 15
• Reflection: 10

Methods carries the most weight because it carries the transparency markers.

One Site, Many Pages

project:
  type: website

website:
  title: "Twitch Chat Study"
  navbar:
    left:
      - text: "Report"
        href: index.qmd
      - text: "Codebook"
        href: codebook.qmd

_quarto.yml names the site and lists its pages, turning a folder of .qmd files into one navigable site.

Studio Block One

White Paper

• Get index.qmd rendering, even if it is mostly headings
• Move your Results in first, since you already have it
• Replace every pasted number with an inline r expression
• Replace every pasted figure with a chunk that produces it
• Then write the Discussion paragraph reporting size and significance

Going Live

v2v::deploy_portfolio()

Pre-flight checks
  renv library in sync ......... ok
  no uncommitted changes ....... ok
  _quarto.yml valid ............ ok

Rendering 4 pages ... done
Live at: https://username.github.io/twitch-chat-study/

The Three Pre-Flight Checks

• renv in sync: built with the versions the source declares
• No uncommitted changes: live matches what is saved
• _quarto.yml valid: the render will not fail partway
• Each heads off a demoralizing failure
  • The last being a collapse the night a portfolio is due

The checks exist so the last step of a long study is the calm one.

The Poster

• 36 by 24 inches, landscape orientation
• Title, authors, institution, and a 100-word abstract
• Background and hypotheses, then condensed methods
• Key results with one or two theme_v2v() figures
• Discussion and implications, plus a QR code

What Earns Space on a Poster

• A ninety-second pitch, not a paper at a larger font
• Pick the one figure that lands the finding in thirty seconds
  • Give it the most space on the board
• Cut every sentence that shows effort rather than result
• Numbers survive, hedges do not

Your reader is standing up, holding a drink, four feet away.

Studio Block Two

Poster

• Sketch the layout on paper first, in blocks, before software
• Place your chosen figure, then build the text around it
• Write the 100-word abstract last, once the blocks exist
• Generate the QR code from the URL the deploy gave you
• Read your title aloud, and cut it if it takes two breaths

Common Failures in the Last Week

• Rendering on your machine and nowhere else, because renv drifted
• A figure that vanishes in the site
  • Its chunk depends on an object made in your console
• A poster exported at screen resolution, printed at 36 inches
• A QR code pointing at a private repository

Test the whole chain today, on a machine you did not build it on.

The Reflection

• One paragraph, and the portfolio is incomplete without it
  • What turned out harder than expected
  • Where a coding rule proved ambiguous in practice
  • What the dataset could not answer, and why
• Resist the bland version and the anxious one

A reflection is evidence of judgment, not an apology.

Checkpoint Before You Leave

You should have:

• A portfolio site that renders locally with no errors
• A Results section with zero hand-typed numbers
• A published URL, and a QR code pointing at it
• A poster layout with its main figure chosen and placed
• The reflection paragraph drafted, even roughly

From Vibes to Variables

• A question became a design, and a design became data
• Data became figures and a test
• All of it became a site another person can check and extend
• The dataset happened to be Twitch chat, and nothing depended on that

The distance from vibes to variables is the whole of research methods, and you have now walked it once.

Due in Finals Week

Final deadline

• The White Paper and its poster, submitted together, 200 points
• The published portfolio URL, live and public
• Your reflection paragraph, included in the portfolio
• Your final journal entry
• My office hours run through finals week, so bring a draft early

Questions?