Multiple-Choice Poll Results: Selections Versus People

Use the free poll builder and results visualizer to enter a complete multi-select tally and choose its percentage basis. This guide works through a creative studio’s session choices, including unknown respondents, missing options, zero results, impossible records, rounding, and exact-count ties.

Name the question, the counted event, and the people base

Multiple-choice poll results become confusing when votes is used for everything. Start by naming the event counted in each row. In a multi-select question, the row contains selections of that option. One person can contribute to several rows while still being one respondent. The total of those rows is a selection total, not automatically a people total. A percentage is only interpretable when the reader knows which of those quantities forms its denominator.

Our fictional Cedar Studio asks: Which creative sessions would you use next month? The options are Stencil printing, Paper construction, and Poster lettering. Respondents may select every session that interests them, with each option counted at most once per person. The source review establishes that 100 people answered this particular question. It records 72 selections for stencil printing, 48 for paper construction, and 30 for poster lettering. These are illustrative observations, not predictions of bookings or attendance.

The count of 100 belongs to people who answered this question, not invitations sent, people who opened an announcement, registered studio members, or attendees at a previous event. Someone who saw the question but made no selection is not part of the declared answered-question denominator. If a broader survey includes this question, its overall completion total is not automatically the correct base either. Establish the relevant respondent count in the source before entering it as people who answered.

The tool handles one poll with two through ten choices and manually entered counts. Choosing multiple choice describes the prepared record; it does not enable a native multi-select feature on a destination platform. You remain responsible for the original response method and the tally. Save a short source description explaining that selections were counted once per option per respondent. Without that context, a correct division could still appear to measure something different from what Cedar actually collected.

Calculate each option’s share of all selections

First add the complete option counts: 72 + 48 + 30 = 150 selections. For the selections basis, divide an option’s count by 150 and multiply by 100. Stencil printing has 72 / 150 x 100 = 48.0%. Paper construction has 48 / 150 x 100 = 32.0%. Poster lettering has 30 / 150 x 100 = 20.0%. These percentages describe how all recorded selections are distributed among the three offered sessions.

A clear chart title is Share of 150 recorded selections. Beneath it, retain the exact counts and the question’s observation period. A statement such as Stencil printing accounted for 48.0% of selections is correct. A statement such as 48.0% of people wanted stencil printing is incorrect for this calculation. The numerator happens to count people selecting that option under the source rule, but the denominator here is all selections, which includes contributions across several options.

This basis is useful when the immediate question is how the selection pool was distributed. It can show which options attracted relatively more recorded selections without requiring a respondent total. It does not describe mutually exclusive groups of attendees. The 48.0% stencil segment and the 32.0% paper segment can contain selections from the same people. Do not infer that these represent separate audiences or use their areas to allocate supposedly non-overlapping groups of participants.

Because every selection belongs to one offered option, exact selection shares sum to one hundred when all counts are complete and the total is positive. Displayed shares may differ slightly after independent rounding. If one row is unknown, the sum of known rows is not the complete selection denominator. Do not call a calculation based only on those available rows the final distribution. Repair the missing count or keep the draft unfinished rather than publishing an incomplete fraction with a complete-looking label.

Use respondents to answer how many people chose each option

Now keep the same counts but change the denominator to the known 100 people who answered. Stencil printing has 72 / 100 x 100 = 72.0% of respondents. Paper construction has 48.0%, and poster lettering has 30.0%. These percentages answer a different question from the selection shares: what proportion of the answered-question base selected each session? The count of 72 did not change. Only the quantity used to interpret it as a percentage changed.

The displayed respondent percentages sum to 150.0%, which is valid for this multi-select question. Some people selected more than one option, so the rows overlap. Do not divide those percentages by their sum to force them back to one hundred and still call them percentages of respondents. That normalization would return the selection basis instead. Use a clear label such as Percentage of 100 respondents, with a note that multiple selections were allowed and totals can exceed one hundred.

The two views can be explained together in one sentence: Stencil printing was selected by 72 of 100 respondents and accounted for 72 of 150 recorded selections. That is 72.0% of respondents and 48.0% of selections. Include both denominators when comparing the figures. Without them, a reviewer might think one chart was recalculated incorrectly or that a new batch of responses arrived between exports. Save the approved version with its chosen basis clearly identified in the handoff.

The complete total also yields 150 / 100 = 1.5 selections per respondent on average in this example. That average does not mean every person chose one and a half sessions, nor does it reveal the exact overlap. The option totals alone cannot identify how many selected both stencil printing and paper construction. Do not invent an intersection chart or a list of unique session audiences from these marginal counts. The original response-level record would be needed for that different analysis.

Distinguish unknown respondents, missing options, and entered zeros

If Cedar has all three option counts but cannot establish how many people answered, leave the respondent count unknown. The 150-selection total still supports the 48.0%, 32.0%, and 20.0% selection shares. Count bars of 72, 48, and 30 also remain useful. Respondent percentages do not. Do not use the largest option count, total selections, announcement views, or membership as a convenient substitute for the missing people denominator. Each would create a different and unsupported interpretation.

A missing option count is a different problem. Suppose the paper-construction count was not supplied. The known values of 72 and 30 sum to 102, but that is only a subtotal. It cannot support the complete selection distribution or a complete top-choice statement. Even a known respondent total of 100 does not fill the missing row. The builder requires complete valid option counts for a complete results chart; keep the unavailable value blank and resolve it at the source.

An entered zero is an observation, not a gap. In a separate record of 20 respondents, counts of 12, 8, and 0 mean the third option was explicitly recorded with no selections. The selection total is 20, so shares are 60.0%, 40.0%, and 0.0%. A blank third field would not justify that result. Use 0 only when the source supports it. Cleaning a worksheet by filling every empty cell with zero can silently change both the total and the conclusion.

Unfinished work can still be worth keeping. Copy the question independently of the results, and save a valid incomplete JSON draft when you need to return later. Do not expect the page to autosave. Record what is missing in a brief source note, but do not turn that note into a numerical guess. A transparent unfinished tally is easier to repair than a polished chart whose percentages conceal the fact that one of its inputs was never observed.

Reject impossible records before drawing conclusions

With one selection allowed per option per person, no option count can exceed the number of respondents. Cedar’s count of 72 for stencil printing cannot coexist with a declared total of 60 people who answered under that rule. Check whether 60 belongs to a different question, a partial source window, or an unrelated attendance record. Do not increase the respondent number just enough to make the form pass. Resolve the source disagreement before interpreting the result.

The opposite consistency check concerns total selections. If every declared respondent answered by selecting at least one option, selections cannot total less than respondents. Cedar’s 150 selections cannot describe 160 people who answered this question. Perhaps 160 is the number who opened a survey, including people who skipped the item. That broader count is not the required denominator. Use the answered-question total only when the source establishes it, rather than treating an overall participation figure as interchangeable.

Check the counting unit when apparent duplicates appear. Repeated mentions in comments are not necessarily additional selections of the same option by a respondent. The tool does not deduplicate people or ballots. If the raw tally lets one person contribute several counts to the same option, it does not fit the respondent-per-option interpretation without separate review. Resolve that tally outside the builder, or describe a different count-based record without pretending its bars measure the percentage of unique people who chose each session.

Use whole nonnegative option counts, within the editor’s range of zero through ten million. A negative value is not a meaningful negative selection, and a fractional person is not an acceptable way to reconcile rounded source percentages. These validation rules establish internal consistency only. A record can pass them while still having a wrong source window or mislabeled question. Compare every input with the source, including the denominator, before treating successful entry as evidence that the underlying collection was accurate.

Handle zero totals, rounding differences, and exact-count ties

An all-zero tally has no positive selection denominator. Zero divided by zero is undefined, so do not publish 0.0% for every option or split the chart into equal shares. Report no recorded selections at the checkpoint. If respondents are also zero, there is no respondent percentage either. If a positive answered-question count is declared with all option counts zero, the record contradicts the rule that respondents answered by selecting at least one option and needs correction.

Independent one-decimal rounding can produce a displayed total just below or above one hundred. Counts of 1, 1, and 1 have exact selection shares of one third each, displayed as 33.3%, 33.3%, and 33.3%. The sum is 99.9%. Counts of 1, 1, and 4 out of 6 produce 16.7%, 16.7%, and 66.7%, summing to 100.1%. Keep the exact counts and explain rounding instead of altering an observation or silently assigning a leftover tenth to one row.

Ties are determined by exact counts, not displayed percentages. Consider counts of 1,001, 1,000, and 7,999, totaling 10,000 selections. The first two shares both display as 10.0% to one decimal, but their counts differ by one. They are not tied. Conversely, two options with exactly 48 selections are tied even if color, chart order, or label length makes one more visually prominent. A top-choice statement must follow the counts, not a formatting impression.

The same rounding principle applies to respondent percentages. Do not use rounded labels as inputs to another calculation or subtract them to claim an exact difference. For a precise audit, return to counts and the original denominator. An independently rounded bar label is a readable summary, not the complete data record. In a narrative, distinguish a descriptive lead from any claim of statistical significance; this tool supplies neither significance testing nor evidence that a self-selected poll represents a wider population.

Build a chart whose labels preserve the calculation

Choose the results view only after reviewing the complete tally. For Cedar’s respondent chart, use the three counts with a respondent total of 100 and explicitly select the respondents percentage basis. The title or accompanying explanation should identify the creative-session question and observation window. Label the bars Percentage of respondents rather than simply Results. Retain the count of each option so a reader can reconstruct the division without relying on the visual length of a colored bar.

For the alternate chart, change the basis to selections and state Total selections: 150. The resulting 48.0%, 32.0%, and 20.0% are not a new poll result. They are another view of the same supplied tally. If sharing both charts, explain their different questions beside them. Do not place them in a before-and-after sequence that suggests a change over time. A changed denominator can make a bar look smaller even when not a single count has changed.

Use a count chart when respondents are unknown and people percentages would invite an unsupported assumption. Its axis or labels should say selections, with counts of 72, 48, and 30. Make clear that the respondent total was not supplied. Count bars do not solve a missing option count; complete valid rows are still required for the complete result. Choose a format that accommodates the whole session label rather than abbreviating Paper construction until only the organizer knows what it means.

Write the accompanying result in plain language: Stencil printing received 72 selections, the largest option count. Among the 100 people who answered, 72.0% selected it; the question allowed multiple selections. This wording reports the observed record without promising attendance or calling the result a representative survey. The question card remains a separate image for asking the question. A static PNG of either view does not host voting, collect new selections, or update when the source changes after export.

Export the exact record and use it for a bounded decision

Cedar can use the result to prepare the stencil-printing session first while checking demand for the other formats through its ordinary planning process. A higher option count gives an ordering signal among these respondents and choices. It does not reveal how the entire studio membership would answer, guarantee filled places, or identify the best schedule. Keep the practical decision proportional to the evidence: prepare a session outline, confirm resources, and ask a separate booking question when actual attendance needs to be established.

Review the record in the multiple-choice results visualizer before exporting. Check all option counts, the selected percentage basis, the respondent field, and the source period. Use the clean PNG for the chart and full copy or TXT for a readable explanation beside it. The artwork carries no forced advertising watermark. A chart without its denominator is not ready to publish just because every bar is visible and the design looks finished.

Download CSV for structured data and JSON for the complete editable checkpoint. The PDF provides a raster report with exact project data attached and a footer linking to the tool. Its visible text is not selectable data, so retain the text exports when the recipient needs to reuse the wording or numbers. Save locally before leaving the page. The project contains authored text and counts, not respondent identities, account files, a live survey connection, or an automatically refreshed results feed.

For a wider account review, the social media report builder can hold manually supplied account metrics and your written interpretation. Keep this poll’s selection and respondent figures clearly distinguished from account reach or engagement there. Nothing transfers automatically between the tools or directly into SMTasker. Finish the handoff with the question, source, checkpoint, exact counts, denominator, and bounded next action, so another person can verify the arithmetic without reconstructing the entire collection process.

Questions about Multiple-Choice Poll Results

Continue with another poll-planning guide

Apply the idea in the Poll Builder & Results Visualizer, or continue with another focused poll guide. Browse more utilities in all free tools.