Writing Style (poq-writing-style)
A POQ paper must be readable by both substantive opinion scholars and survey methodologists, and
disciplined to the cap for its type. The word count covers text and notes but excludes figures,
tables, references, and appendices — so the AAPOR disclosure appendix and exhibits do not count
against you. This skill is about reaching the POQ audience and respecting format — not generating claims.
When to trigger
- Drafting the introduction, framing the contribution, or final polish
- Over the word cap and needing to cut without losing the argument
- Writing the abstract (confirm the current requirement — 待核实 on exact cap)
- Aligning citations and format before submission
Reach the POQ audience
-
Front-load the contribution. By the end of the introduction the reader knows the question, the
data, the finding, and whether the contribution is to opinion theory, current opinion, or survey
validity. Do not bury the "so what."
-
Name the methodological move plainly. State the survey design and the error source you address;
POQ readers expect to see coverage/nonresponse/wording/mode/weighting handled, not glossed.
-
Define jargon on first use. A communication scholar should follow a methods paper and vice
versa; spell out acronyms (AAPOR, TSE, RR, DEFF, AMCE, MRP) on first appearance.
-
Argument-first prose. Lead with claims; use design-based evidence to support them. Avoid "the
data show…" without saying what they show and why it matters.
Format & front matter
-
Citations: keep one consistent reference style (manage with Zotero/BibTeX); references excluded
from the word count.
-
Disclosure: ensure the "Appendix A: Disclosure Elements" is present and referenced (see
poq-survey-design-and-measurement); it does not count toward the cap.
-
Anonymize: POQ is double-blind — no author names/affiliations/acknowledgments in the
manuscript, no obvious self-references, strip identifying file metadata.
-
Data Availability Statement: plan the DAS for the endmatter (see
poq-transparency-and-data-policy).
Fit the type cap (text + notes; excludes figures, tables, references, appendices)
- Move detailed methods, full specifications, and extended robustness to appendices (excluded).
- Cut throat-clearing and literature dumps; engage the debate, not every paper (see
poq-literature-positioning).
- Tighten footnotes — they count toward the limit (notes are in the cap).
- Prefer one decisive trend figure to three redundant tables.
Micro-rewrites in the POQ register
-
Claim-first, not data-first. Before: "The data show significant differences across
demographic groups (p < .05)." → After: "Partisanship, not education, drives the gap: the
partisan difference is 23 points versus 4 points across education levels." POQ readers want the
substantive magnitude, then the inference.
-
Name the error source, don't gesture. Before: "We account for potential survey biases." →
After: "Because the panel under-covers non-internet households, we benchmark against the CPS and
weight on the coverage-related variables." A survey scientist can now evaluate the fix.
-
Un-bury the contribution. Before: an intro that spends four paragraphs on the history of
polarization research. → After: paragraph two ends with "We show that item wording, not true
attitude change, produced the apparent trend — a measurement contribution." State the lane
(opinion theory / current opinion / survey validity) in the intro, in those terms.
Two referee tics worth pre-empting: POQ methodologists flag any inference language ("significant,"
"effect") that outruns the design (cross-sectional, nonprobability), and substantive referees flag
acronym walls — one undefined DEFF or AMCE in the intro reads as writing for insiders only.
Anti-patterns
- An intro that never states whether the contribution is to opinion or to survey methods
- Burying the methodological move (how error was handled) deep in the paper
- Padding a Research Note (< 3,000) or Polls in Context (≤ 2,500) toward Article length
- Acknowledgments or self-references that break double-blind anonymity
- Putting disclosable methods in long footnotes (which count) instead of Appendix A (which does not)
Output format
【Contribution stated by end of intro?】opinion / current-opinion / survey-validity [Y/N]
【Methodological move named?】error source + design plain? [Y/N]
【Abstract】present + within current requirement (待核实)
【Word count】Article ≤6,500 / Note <3,000 / Polls ≤2,500 (text + notes)?
【Anonymized + DAS planned + Appendix A referenced】[Y/N]
【Next】poq-transparency-and-data-policy
Supplementary resources
