Read a membership survey without losing the denominator

Start with who answered, not a headline percentage

The membership committee—Eric Reiter, Allyson Lunny and Lyndsay Campbell—ran an online survey in winter 2014. Its four-page summary says about 250 people were surveyed and 47 responded. These are observations about that exercise, not the present membership or a representative census of Canadian sociolegal scholars. The summary does not give the full questionnaire, sampling frame or a nonresponse analysis.

Every question has its own base. Page 1 uses 47 for membership class, 46 for duration and number of other memberships, and 40 for the names of other associations. Page 3 uses 46 for methodologies and 38 for publication destinations. Page 4 reports 44 answers about SSHRC grant applications. Copy the base before doing any division.

First page of the 2014 membership survey with 47 responses and question-specific bases — Original Association survey summary, physical page 1. Used with the owner’s permission.

Make three different kinds of count visible

The affiliation question explicitly permits more than one affiliation. A list of organisations or degrees likewise cannot be read as a partition merely because it appears in bullets. Methodology counts sum to 131 selections across 46 respondents; dividing each selection by 46 produces percentages that can legitimately add to more than 100%. This does not establish which combinations any individual selected.

Question bases and response types in physical pages 1–4
People answering	The methodology question has 46 respondents; do not replace this with the 47 total responses.
Selections	Methodologies include theoretical 27, empirical interdisciplinary 24 and historical/archival 19. They are not mutually exclusive categories.
Unknown or unclear	The SSHRC question totals 35 yes + 7 no + 2 unclear = 44. Keep unclear visible instead of silently excluding it.

Worked calculations: keep the sentence with the number

Reproducible arithmetic; the percentages are editorial calculations, not new survey data
Individual memberships	27 ÷ 47 × 100 = 57.4%, rounded to one decimal. Sentence: 27 of the 47 respondents reported individual membership.
Student memberships	17 ÷ 47 × 100 = 36.2%. The 17 are respondents, not the total student membership in 2014 or today.
Historical/archival methodology	19 ÷ 46 × 100 = 41.3% selected this methodology. It is not a 41.3% exclusive category of scholars.
SSHRC answers	35 ÷ 44 × 100 = 79.5% answered yes. The count includes two unclear answers in the base; it concerns applying, not receiving a grant.
Invitations and response	47 ÷ about 250 is approximately 19%. Since the invitation number is approximate, a precise response rate would overstate the source.

Membership class counts 27, 17 and 3 total 47; methodology selections use a different base of 46 — Original diagram from the 2014 survey, physical pages 1 and 3. Counts are preserved; no current membership estimate is implied.

A five-pass procedure for any summary

Record the source, year, authoring committee and physical page. Identify what the summary does not supply.
For each question, write its exact base and whether choices overlap. Leave unknowns explicit.
Check sums against the stated base only where the categories form a partition. Never force a multiple-choice list to sum to the number of people.
Write the calculation and a complete sentence identifying the population of respondents. State rounding.
Separate what the survey describes from an action you propose. A new priority requires current evidence, not an old percentage alone.

Copy a complete reading sheet

Blank template for your own notes; no responses are collected here
Source and timing	Title / source URL / year of collection / physical page / authoring committee.
Question and unit	Question wording as supplied / people or selections / base / multiple answers allowed?
Computation	Numerator / denominator / formula / rounding / unknowns retained.
Claim and limit	What the summary establishes / missing questionnaire or data / what cannot be generalized.
Possible next step	A specific follow-up question / evidence needed / no invented updated result.

A completed decision note

Question: should a methods workshop offer a route for historical and archival research? Source: methodology question, physical page 3, winter 2014, 46 respondents, multiple selections. Observation: 19 selections for historical/archival work, alongside theoretical, comparative and interdisciplinary methods. Decision to investigate: ask current prospective participants which sources and methodological obstacles they face. Do not claim that 41.3% of today’s membership needs this workshop or that the old categories predict demand.