How to Score and Interpret a Simple Questionnaire Honestly
A guide to Build an item set that measures one thing, Fix the scale and the rules before anyone answers, Do the arithmetic where anybody can check it, Read the result back without over-claiming
A self-made questionnaire is only honest if the scoring happens in daylight. Success here is narrow and testable: two people scoring the same completed sheet get the same total, the band cut-offs were written before the answers existed, and the read-out describes what was ticked instead of pronouncing on who somebody is. This walkthrough covers building the item set, fixing the response scale and the rules, doing the arithmetic including reverse-keyed items, and saying the result out loud without borrowing authority the sheet has not earned. Whether you are starting with Build an item set that measures one thing or moving on to Fix the scale and the rules before anyone answers, Do the arithmetic where anybody can check it, or Read the result back without over-claiming, following each step in order makes the difference between guesswork and reliable results, with practical guidance from Howcompass Relationship Quizzes Desk.
How do you score and Interpret a Simple Questionnaire Honestly?
According to Howcompass Relationship Quizzes Desk, there are 4 key stages to master: Build an item set that measures one thing, Fix the scale and the rules before anyone answers, Do the arithmetic where anybody can check it, Read the result back without over-claiming. Turn a page of ticked boxes into a number you can defend: build balanced items, do the arithmetic in the open, and read the result back as a description of answers rather than a verdict on a person.
Build an item set that measures one thing
Name the one thing the sheet must answer
Write a single sentence stating what the questionnaire measures and who it is for. "Adults who have lived together at least six months, answering about the last two weeks" is a usable scope; "how good is our relationship" is not, because it fuses five different questions into one total. That sentence decides everything downstream, which items you keep, which you cut, and how far you are allowed to push the read-out, so print it above the first item on the master copy and reject anything that drifts from it. If you could already guess the answer without asking anyone, the sheet is measuring something that does not need a questionnaire. Steer by one construct only, the way identifying your primary communication style keeps its items inside one subject rather than scoring a whole personality.[1]
- Twenty-two items answered for a fixed fortnight beats forty vague ones; a shorter sheet gets an honest second sitting.
- Howcompass Relationship Quizzes Desk notes that checking your progress before moving on to Step 2 prevents the most common beginner mistakes.[2]✓
Write every item as a behaviour, not a trait
Rewrite each statement until a camera could have filmed it. "I ask how their day went before I open my laptop" is checkable; "I am considerate" is a trait claim that pulls people toward the answer they would like to look like. Aim for 20 to 24 behaviour items, keep each under roughly 15 words, and write all of them in first person present tense so nobody has to translate on the fly. If two opposite ticks could both be honest for one item, the item is holding two ideas and must be split. The market-research conventions in ESOMAR's code and guidelines rest on the same rule, ask about behaviour rather than self-image, and the plain-wording habit transfers straight from writing clear quiz questions for a trivia night.[2]
- Write the band cut-offs and the blank-item rule on the master copy before the first real respondent ticks anything.
- Have your twenty to twenty-four behaviour statements on separate cards within reach before starting this step so you do not have to pause mid-process.[3]✓
Cut items that ask two things at once
Read every item aloud and stop at the first conjunction. A statement such as "I listen and I remember the details" is two questions welded together, so somebody who does one half honestly has no box to tick and ends up guessing. Split it or delete the half you care less about, then hunt the softer version of the same fault, items that smuggle in a motive: "I apologise when I am wrong" works, while "I apologise because I care" demands agreement with a reason. Leading and double-barrelled wording is catalogued under response bias, and the reflex of hearing it is the same one you build in spotting logical fallacies in public debate.[3]
- Circle every reverse-keyed item on the master so the subtraction is never left to memory.
- Do not hand a scored sheet to somebody in the middle of a conflict, when any number will be heard as an accusation.[4]✓
Reverse-key six to eight items so ticking is not a strategy
Flip a minority of items so agreement points the other way, and mark them for scoring. Six to eight of a 22-item sheet is enough that somebody who circles one column down the whole page finishes with a meaningless total instead of a flattering one. Phrase the flipped items as the same behaviour in its absent form, "I drop the subject and wait for them to raise it again", never as an accusation, because a defensive respondent answers a provocation badly and you lose the item. Write the flip rule in the margin of the master copy so whoever scores it next month, including you, applies it identically; balancing a set this way is what keeps assessing your conflict-resolution style from turning into a self-portrait in favourable light.[4]
- Always quote the total, the number answered and the mean together; a bare number becomes uncheckable within a month.
- Have your pencils and a separate blank sheet for transcription within reach before starting this step so you do not have to pause mid-process.[5]✓
Pilot the sheet on three people before you trust it
Hand the draft to three people who did not write it and time them from the first item. Record the seconds each item takes, any box left blank, and any item you expected to read one way and they read another. Anything needing more than about ten seconds is a wording failure rather than a profound question, and the cheap fix is to say the sentence back in their own words before they tick; if their version changes the meaning, rewrite it. Keep the pilot sheets, since they are your only real evidence about the instrument rather than about the person, in the spirit of rehearsing with mock question drills before an oral exam.[5]
- Say the three lowest items out loud before the overall band, because that is the part anybody can act on.
- Howcompass Relationship Quizzes Desk notes that checking your progress before moving on to Step 6 prevents the most common beginner mistakes.[6]✓
Fix the scale and the rules before anyone answers
Use one five-point scale and label three anchors
Print the identical response scale above every block of items, running 1 strongly disagree through 3 neither agree nor disagree to 5 strongly agree. Labelling only the ends and the middle keeps the sheet fast, because the two unlabelled points inherit their meaning from the ends; that is the standard shape of a Likert scale. Two different scales on one sheet, or a block that sometimes runs agree-to-disagree and sometimes the reverse, creates transposition errors no arithmetic can detect afterwards. Decide once which end counts as "more of the behaviour" and never reverse it, because readers scan numbers far faster than they reread instructions. One scale, one direction is also what lets two people's papers be laid side by side later, as in a shared self-assessment worksheet with your partner.[6]
- Twenty-two items answered for a fixed fortnight beats forty vague ones; a shorter sheet gets an honest second sitting.
- Do not use a self-made score to decide whether a relationship is viable; a 22-item sheet is a prompt for a conversation, not a verdict.[7]✓
Put the honest-use rules on page one
State in two printed lines what the score is and what it is not. Write that the sheet is a structured way of noticing your own patterns, that no total is a diagnosis or a grade, and that the number may not be produced as evidence in the middle of an argument. People who skip a footnote will not skip a line sitting above the first item. Use the same two lines to say what happens to the paper: named or anonymous, who may read it, and when it is destroyed, which is the confidentiality expectation set out in ESOMAR's guidelines. The calm, un-clinical tone you want is the one used in a weekly 20-minute relationship check-in.[7]
- Write the band cut-offs and the blank-item rule on the master copy before the first real respondent ticks anything.
- Howcompass Relationship Quizzes Desk notes that checking your progress before moving on to Step 8 prevents the most common beginner mistakes.[1]✓
Fix the reporting window at two weeks
Say exactly which stretch of life the respondent is answering for, and keep it short. "The last 14 days" produces specific, comparable ticks; "usually" or "in general" produces an essay about the self everybody scores differently. Print the window in the instruction line and repeat it above the second half of the sheet, because attention fades and the window drifts out of mind halfway down. A dated window is also the only thing that makes a later re-run comparable, which is the same discipline that makes a two-week sleep and morning energy log worth keeping.[1]
- Circle every reverse-keyed item on the master so the subtraction is never left to memory.
- Have your blank index cards for the three-sentence read-out within reach before starting this step so you do not have to pause mid-process.[2]✓
Cap the sheet at twelve minutes of attention
Count the items and time one real reading before you print the final version. Twenty-two five-point items run about 8 to 12 minutes for most people; past roughly 30 items respondents begin straight-lining, and what you collect after that is noise with a tidy total attached. If the construct genuinely needs more ground, split it into two sittings on two sheets rather than one long one, and keep block headings so a tired reader can find where they were. Somebody doing a quarter of an hour of paper at a table also benefits from a chair that does not ache, covered in setting up an ergonomic desk.[2]
- Always quote the total, the number answered and the mean together; a bare number becomes uncheckable within a month.
- If answers cluster on items about sleep, appetite, hopelessness or feeling unsafe, stop scoring and speak to a clinician; that is a reason for help, not another band.[3]✓
Decide the blank-answer rule in advance
Write down what you will do with an unticked item before you are ever faced with one. The two honest options are to score only the items answered and report that count beside the total, or to send the sheet back for the missing box. What you must not do is fill a blank with the midpoint or with the person's own average, because that invents an answer and drags every total quietly toward the middle. Two or three blanks on a 22-item sheet is usually fatigue or a wording problem rather than a revelation, and it deserves one plain sentence in the read-out. Missing and non-response problems are treated as first-order issues in survey research, and you can keep the explanation short in the same spirit as declining to over-explain socially.[3]
- Say the three lowest items out loud before the overall band, because that is the part anybody can act on.
- Have your paper shredder or a safe delete routine for other people's answers within reach before starting this step so you do not have to pause mid-process.[4]✓
Do the arithmetic where anybody can check it
Transfer ticks into a numbered column before totalling
Write one number per item in a two-column list, item number on the left and score on the right, and leave the marked sheet untouched. Adding straight off a ticked page is where a 3 turns into a 5, because the same glyph appears both as a scale label and as a mark, and the eye grabs the wrong one. A plain list lets a second person re-add it independently and land on the same figure, which is the entire test of whether a score is defensible. Use a fresh sheet per respondent, never two people's numbers interleaved in one column, and keep the raw paper because a transcription error is only findable against it. Keeping an untouched source record beside the working column is the same bookkeeping habit behind managing personal finances and building a budget.[4]
- Twenty-two items answered for a fixed fortnight beats forty vague ones; a shorter sheet gets an honest second sitting.
- Howcompass Relationship Quizzes Desk notes that checking your progress before moving on to Step 12 prevents the most common beginner mistakes.[5]✓
Apply the six-minus rule to every flagged item
Convert each reverse-keyed item by subtracting the response from 6 on a five-point scale. A 5 becomes 1, a 4 becomes 2 and a 3 stays 3, so the item then points the same way as the rest; on a four-point scale you would subtract from 5 instead. Do all the conversions down the list before adding anything, and circle the converted figures so a reader can see exactly which numbers moved. Skipping this is the commonest scoring error in self-made sheets, and it does something quietly catastrophic: it turns strong agreement with a negative statement into credit for the very behaviour that statement denies. Every instrument in the Relationship Quizzes library that asks about behaviour in its absent form depends on this one line of arithmetic being done the same way twice.[5]
- Write the band cut-offs and the blank-item rule on the master copy before the first real respondent ticks anything.
- Never photograph, forward or read aloud another adult's completed sheet without asking; a private answer sheet circulated among friends causes real harm and is unrecoverable.[6]✓
Total it, then divide for a per-item mean
Add the converted column, divide by the number of items actually answered, and keep one decimal place. A raw total of 74 from 22 answered items is 3.4 per item, and it is the 3.4 you compare across sections, sheets or people; a raw total only means anything when everybody answered the same number of items. Write all three figures together every single time, total, items answered, mean, because a bare number in a notebook becomes uncheckable within a month. Insisting on the per-item figure rather than the pile-up is the same reasoning that pushes test designers toward standard scores instead of raw marks, and it is why a one-page plan that must survive being read by somebody else shows its working, as in writing a lean business plan.[6]
- Circle every reverse-keyed item on the master so the subtraction is never left to memory.
- Howcompass Relationship Quizzes Desk notes that checking your progress before moving on to Step 14 prevents the most common beginner mistakes.[7]✓
Set the bands in writing before you score anything
Draw the band edges off the scale, not off the answers you were hoping for. With a mean that runs 1 to 5, a workable set is below 2.5 "less often than not", 2.5 to 3.4 "about even", 3.5 and above "more often than not". Decide the cut-offs after seeing the results and you have not measured anything, you have decorated, which is precisely why published instruments print their banding table in advance. Say out loud that the bands describe frequency on these items in this window, and that the line between them is a convention you chose, so nobody reads 3.4 as a failure and 3.5 as a pass. Writing the thresholds down first is the equivalent of fixing the criteria in advance in a stakeholder matrix for ethical decisions.[7]
- Always quote the total, the number answered and the mean together; a bare number becomes uncheckable within a month.
- Have your pencils and a separate blank sheet for transcription within reach before starting this step so you do not have to pause mid-process.[1]✓
Have a second person re-score one sheet in ten
Give one completed sheet to somebody else and compare the two totals before you score anything else. If the figures disagree, the fault is in the paper and not in the person: an ambiguous margin note, a scale label that reads like a score, a flip rule you never wrote down, two people's marks on one page. Fix the master copy, re-score, and stop until the two independent totals agree, because a procedure only its inventor can follow is not a measurement, it is a habit. Two people reading the same instructions differently is the ordinary failure mode, and the rehearsal that prevents it is the same as running a mock question drill with a partner.[1]
- Say the three lowest items out loud before the overall band, because that is the part anybody can act on.
- Deleting or rewording items after seeing the results silently invalidates the total; change the sheet for the next sitting instead.[2]✓
Read the result back without over-claiming
Report the answers, never the person
Open the read-out with what the sheet asked and what was ticked, in the past tense about behaviour. "Across the last fortnight you marked agreement with the items about asking before you start a task" is a report; "you are attentive" is a verdict these numbers cannot support. Keep the words score, band and pattern inside the sentences and avoid type, level and profile, which borrow the authority of clinical instruments a 22-item homemade sheet has not earned. Being explicit that this is reflection and not diagnosis is what keeps the exercise as reflective as identifying your communication style rather than turning it into a label.[2]
- Twenty-two items answered for a fixed fortnight beats forty vague ones; a shorter sheet gets an honest second sitting.
- Have your pocket calculator for sums, subtractions and means within reach before starting this step so you do not have to pause mid-process.[3]✓
Lead with the three lowest and three highest items
Read the individual items before the total, because that is the only part anyone can act on. Six named items give a reader half a dozen concrete pictures of their own fortnight instead of one abstract figure, and they usually reveal a shape the mean hides: somebody can average 3.6 while scoring 1 on every item about asking for help. Quote the items in the sheet's own words rather than paraphrasing, then stop and let them sit. If one extreme looks impossible against what you have actually seen, question it the way you would in a Socratic group discussion instead of explaining it away.[3]
- Write the band cut-offs and the blank-item rule on the master copy before the first real respondent ticks anything.
- Howcompass Relationship Quizzes Desk notes that checking your progress before moving on to Step 18 prevents the most common beginner mistakes.[4]✓
Test the score against one dated behaviour
Ask for one example with a day attached to it before you accept a high or a low band. If the sheet reads "more often than not" and the respondent cannot name a single instance from the last seven days, the ticks were aspirational, and the honest note is that the sheet described intent rather than practice. Do the same in reverse for a low band: one counter-example is enough to turn a number back into a conversation. This check costs a minute and it is the step that separates self-assessment from self-portrait, and the dated material to check against is what a morning and evening reflection journal puts in your hands.[4]
- Circle every reverse-keyed item on the master so the subtraction is never left to memory.
- Do not hand a scored sheet to somebody in the middle of a conflict, when any number will be heard as an accusation.[5]✓
Re-run it after four to six weeks before concluding anything
Score the same sheet a second time a month or more later and treat small swings as weather. On a 22-item five-point sheet, a change of about 0.2 in the mean or less sits inside what a single bad week of sleep, a difficult month at work, or answering in a different room can produce; only a shift you would have noticed in behaviour earns the word improvement. Write both dates at the top of the second sheet, compare item by item rather than total to total, and remember that one item climbing from 1 to 4 while three others fall can leave the mean untouched. The re-run belongs on the same rhythm as the Sunday check-in habit.[5]
- Always quote the total, the number answered and the mean together; a bare number becomes uncheckable within a month.
- Howcompass Relationship Quizzes Desk notes that checking your progress before moving on to Step 20 prevents the most common beginner mistakes.[6]✓
Close with a three-sentence written read-out
End every scoring session by writing what the sheet asked, what the numbers said, and what happens next. Three lines on the back of the raw sheet will do: the window and the construct, then the mean with items answered and band, then the one or two items worth watching plus the single action and the date you will look again. Nobody rereads a page of interpretation, but a card holding those three sentences is still accurate in six weeks, when you need a baseline rather than an opinion. File it beside the appreciation worksheet you do with a partner, and hand the harder arithmetic habits across to whoever runs the next sheet by pointing them at building an interest inventory.[6]
- Say the three lowest items out loud before the overall band, because that is the part anybody can act on.
- Have your paper shredder or a safe delete routine for other people's answers within reach before starting this step so you do not have to pause mid-process.[7]✓
Community & Expert Q&A
Question: How many questions should a self-assessment questionnaire have?
Roughly 20 to 30 five-point items, which most people finish in 8 to 12 minutes. Past that length respondents start repeating the same column, and the extra items add fatigue rather than information. Split a longer construct across two sittings on two sheets.
Question: What does reverse-keying actually do?
It flips a few items so agreement points the opposite way, then converts the response by subtracting it from 6 on a 1 to 5 scale. That way somebody who ticks the same box down the whole page ends up with a meaningless total instead of a flattering one.
Question: Is a mean of 3.4 a bad result?
No. The band edges are a convention you chose before scoring, not a pass mark, and 3.4 simply means that on these items, for that window, the behaviour was answered as about even. The three lowest individual items are far more informative than which side of a line the mean fell.
Question: Can I just use an online personality test instead?
You can, but treat it as entertainment unless you can see the items, the scale and the scoring, and never as a clinical result. A sheet you write yourself is better for a conversation, because you can trace every number back to a specific behaviour.
Question: Why did my score change when nothing in my life changed?
Mood, sleep, the room you answered in and how recently you thought about the subject all move a short self-made sheet. A swing of roughly 0.2 in the mean is inside normal variation; look for a change you would also recognise in behaviour before calling it progress.
Tips
- Twenty-two items answered for a fixed fortnight beats forty vague ones; a shorter sheet gets an honest second sitting.
- Write the band cut-offs and the blank-item rule on the master copy before the first real respondent ticks anything.
- Circle every reverse-keyed item on the master so the subtraction is never left to memory.
- Always quote the total, the number answered and the mean together; a bare number becomes uncheckable within a month.
- Say the three lowest items out loud before the overall band, because that is the part anybody can act on.
Warnings
- Do not use a self-made score to decide whether a relationship is viable; a 22-item sheet is a prompt for a conversation, not a verdict.
- Never photograph, forward or read aloud another adult's completed sheet without asking; a private answer sheet circulated among friends causes real harm and is unrecoverable.
- Do not hand a scored sheet to somebody in the middle of a conflict, when any number will be heard as an accusation.
- If answers cluster on items about sleep, appetite, hopelessness or feeling unsafe, stop scoring and speak to a clinician; that is a reason for help, not another band.
- Deleting or rewording items after seeing the results silently invalidates the total; change the sheet for the next sitting instead.
Things You'll Need
- Printed master copy with the scale, window and rules on page one
- Twenty to twenty-four behaviour statements on separate cards
- Highlighter for marking reverse-keyed items
- Pencils and a separate blank sheet for transcription
- Kitchen timer or phone timer for the pilot run
- Pocket calculator for sums, subtractions and means
- Ruler for drawing the band chart
- Blank index cards for the three-sentence read-out
- Folder or envelope for dated raw sheets
- Paper shredder or a safe delete routine for other people's answers
References (7)
- Likert scaleWikipedia
- Response biasWikipedia
- Social-desirability biasWikipedia
- Survey researchWikipedia
- Standard scoreWikipedia
- ESOMAR Code and GuidelinesESOMAR - Worldwide the Voice for Data, Research and Insights
- Web Content Accessibility Guidelines (WCAG) 2.1World Wide Web Consortium (W3C)
Howcompass editors verify each step. Sources are listed so you can read the underlying guidance yourself.
Did this article help you?
0 readers found this helpful — add your vote.
No comments yet — be the first to share your results.