Grade by Grade
E. Culminating Data Management Investigation · Culminating Data Management Investigation

Designing an Ethical Survey

E1.2 E1.3About 70 minutes
Learning goals
  • I can apply ethical principles (consent, voluntary participation, anonymity, minimal risk, permission) to my data collection.
  • I can write clear, neutral survey questions that each measure one thing.
  • I can recognize and fix leading, loaded, double-barrelled, vague and overlapping-interval questions.
  • I can choose question types that give data I can analyse with the tools I planned.
  • I can pilot-test a survey and revise it.

Words to know

informed consent
Agreement to take part given by someone who has been told the purpose of the study, what they will be asked, how answers are used, and that they may refuse or stop.
anonymous
No one, including the researcher, can link an answer to the person who gave it.
confidential
The researcher may know who answered, but promises not to reveal it and reports only combined results.
leading question
A question whose wording pushes the respondent toward a particular answer.
double-barrelled question
A question that asks about two things at once, so a single answer cannot be interpreted.
pilot test
A trial run of a survey with a few people from outside the sample, to find confusing questions before the real collection.

People are not just data points

Every answer on a survey comes from a real person who trusted you with it. Collecting data ethically protects that person and also produces better data: people who feel safe answer more honestly.

The curriculum asks you to develop an ethical survey as part of your plan. In Organization of Data for Analysis you learned how to write survey questions and the principles of ethical research. Now you will apply them to your own project, step by step.

Five ethical principles for a student survey

  1. Permission: a survey at school needs approval from your teacher and usually the principal. Follow your school's and school board's rules; some boards require a formal review, or a parent's or guardian's consent for participants under 18.
  2. Informed consent: tell people the purpose, what you will ask, how answers will be used, and how long it takes. Taking part must be their choice.
  3. Voluntary: no pressure, no rewards that feel like pressure, and the right to skip any question or stop.
  4. Privacy: make the survey anonymous if possible (no names, no student numbers, no handwriting that could be recognized, no tiny groups that identify someone). Keep the forms or files secure, and report only combined results.
  5. Minimal risk: avoid questions that could embarrass, upset or expose people (health conditions, family income, illegal acts). If you need such information, use published secondary data instead.

Data about First Nations, Inuit or Métis communities have additional principles of community ownership and control (such as OCAP®, which you met in Organization of Data for Analysis); a student project should use published data shared by those communities rather than collect its own.

1. Introduction and consentpurpose, anonymous, voluntary, skip any question2. Screening questionis the respondent in the population?3. Main questionseasy and neutral first, one idea per question4. Background / lurking variablesonly what the analysis needs5. Thank you and how to return itwhere, by when, who to contact
The structure of a good survey. Consent comes first, the questions the analysis needs come next, and nothing extra is asked.

Writing questions that measure what you mean

  • One idea per question. "Do you work and play sports?" is double-barrelled: split it.
  • Neutral wording. "Don't you agree teens work too much?" is leading. "Should students be forced to stay up late..." is loaded.
  • Precise words and time frames. "Do you sleep enough?" is vague. "On a usual school night, about how many hours do you sleep?" is measurable.
  • Response options that do not overlap and cover every case. "0-5, 5-10, 10-15" leaves 5 and 10 in two groups and has no option above 15.
  • Prefer exact numbers for numerical variables (you can group them later). Use closed questions (choices) for categories and open questions sparingly.
  • Order matters. Start easy and neutral; put the questions you most need early; keep it short (a few minutes).
Poor: choices 0-5, 5-10, 10-15, 15+ (where do 5, 10 and 15 go?)05101520250-55-1010-1515+Better: ask for the exact number, then group it yourself later0510152025Groups made afterwards: 0 to less than 5, 5 to less than 10, 10 to less than 15, 15 or more.
Top: overlapping choices put the boundary values (circled) in two categories, and the dots below show that exact answers each have one clear place. Group them afterwards with non-overlapping intervals.
Worked example

The model investigation, Stage 6: Amira's survey

Write Amira's survey, following the principles. Explain why each question is there.

  1. Introduction: "This anonymous survey is part of a Grade 12 Data Management project about sleep and part-time work. Please do not write your name. Taking part is voluntary: you may skip any question or stop at any time. Only combined results will be reported. Putting your completed survey in the box in the library means you agree to take part. It takes about 3 minutes. Questions? Ask Ms. Ferreira in the math office." (Approved by the teacher and principal.)
  2. Q1. Are you a Grade 12 student? Yes / No. (Screening: only Grade 12 answers are analysed.)
  3. Q2. During the school year, do you usually do paid work (a part-time job)? Yes / No. (Categorical variable for Hypothesis 1.)
  4. Q3. If yes: in a usual week, how many hours of paid work do you do? ____ hours. (Numerical variable for Hypothesis 2; exact number, not intervals.)
  5. Q4. On a usual school night (Sunday to Thursday), about how long do you sleep, from falling asleep to waking up? ____ hours ____ minutes (to the nearest 15 minutes). (Response variable; uses the operational definition.)
  6. Q5. On a usual school night, which TWO activities take most of your evening? Homework / Sports or practices / Paid work / Family responsibilities / Screen time for fun / Other. (Lurking variables for the discussion.)
  7. Q6. On how many school nights in a usual week do you work after 6 p.m.? 0 / 1 / 2 / 3 / 4 / 5. (Helps explain any link between work and sleep.)
  8. Thank you. "Please put this in the sealed box in the library by Friday."

Six short, neutral questions. Each one has a purpose in the analysis plan, the survey is anonymous and voluntary, and nothing sensitive is asked.

Worked example

Example 2: Fix these questions

Identify the problem and rewrite: (a) "Since sleep is so important, how many hours do you sleep?" (b) "How many hours do you work and study each week?" (c) "How much do you sleep? A lot / Normal / Not much" (d) "Do you sometimes fall asleep in class, which is against the rules?"

  1. (a) Leading: the opening suggests a "good" answer, which may inflate reported sleep. Rewrite: "On a usual school night, about how many hours do you sleep?"
  2. (b) Double-barrelled: one number for two activities cannot be separated. Ask two questions: hours of paid work, hours of homework.
  3. (c) Vague categories: "normal" means different things to different people, and the answer is not numerical. Ask for hours to the nearest 15 minutes.
  4. (d) Loaded and risky: it asks people to admit breaking a rule, which invites dishonest answers and could embarrass them. Remove it or ask neutrally ("In a usual week, how often do you feel very sleepy during morning classes? Never / 1-2 days / 3-4 days / Every day").

Neutral wording, one idea per question, precise numerical answers, and no questions that pressure or expose people.

Worked example

Example 3: The pilot test

Amira gives her draft survey to four Grade 11 students (outside her population) and her parent. What does she learn, and what does she change?

  1. Two pilot testers wrote ranges for Q4 ("7-8 hours"). Change: add "(write one number, for example 7 h 30 min)".
  2. One tester who babysits a sibling for pay was unsure about Q2. Change: "Paid work includes regular babysitting or other jobs you are paid for."
  3. Everyone finished in under 3 minutes, and no one felt uncomfortable.
  4. Pilot answers are not included in the real data: the testers are not in the population, and the questions changed.

Piloting found two confusing questions before they could spoil the real data. Amira records these changes in her log and mentions them in her report's method section.

If your data are secondary or simulated

Ethics still applies. With secondary data, respect the licence and cite the source; never try to identify individuals in a data set. With a game of chance, design it for fun with no real money or prizes of value (gambling for money involves laws and real risks), and get permission before running it with others.

Just for fun

A pilot test is a dress rehearsal for your survey. It is the one time in statistics when you want people to find your mistakes.

Hands-on activity

Your investigation, step 4: draft, pilot and revise

You need: your plan, paper or a free online form tool, and three or four people who are not in your sample (family members are fine).

  1. Write your introduction and consent statement using the five principles.
  2. Draft your questions. Beside each, write which variable it measures and which analysis uses it. Delete any question with no analysis.
  3. Check each question against the list: one idea, neutral, precise time frame, non-overlapping and complete options, exact numbers where possible.
  4. Pilot it: watch your testers answer (time them), then ask "Was anything confusing? Did any question make you uncomfortable?"
  5. Revise, and record every change in your project log.
  6. Show the final version to your teacher for permission before collecting any data.

If you use secondary data: instead of a survey, write the exact table names, variables and filters you will download, and test them by downloading a small piece.

Your turn

1

What is the difference between anonymous and confidential?

2

Rewrite to remove bias: "Most experts say teens should work less. How many hours do you work?"

3

Why is "Do you like your job and your boss?" a poor question?

4

Fix these response options for hours of homework: 0-1, 1-2, 2-3.

5

Give two things an introduction to a survey must tell respondents.

6

Why were Amira's pilot answers left out of the real data?

More practice: warm-up

1

Name the five ethical principles from this lesson.

2

Is "Don't you agree that homework is too long?" leading, double-barrelled or neutral?

3

Should a student survey include names?

4

What is a pilot test?

5

Who should approve a survey at school before it is used?

More practice: on your own

1

A student plans to ask, "What is your family's yearly income?" to study income and screen time. Explain the ethical concern and suggest an alternative.

2

In a class of 25, a survey asks students their age, gender, the bus route they take and their sleep. Even with no names, why might it not be anonymous?

3

Explain how asking for the exact number of hours worked, rather than intervals, helps Amira's analysis.

4

Write a neutral Likert-scale question about how rested students feel.

5

A survey offers a $50 gift card to one person who completes it. Discuss one benefit and one ethical or statistical concern.

Challenge

1

Amira's Q5 asks students to check "TWO activities". Explain how she should organize and analyse these answers, and one limitation of this question type.

2

Explain why the order of questions could bias answers, using an example related to sleep.

3

A student wants to run a dice game at lunch for a probability project, with a small prize for winners. List three ethical or practical safeguards.

Check your understanding

Quiz

1.

Which question is double-barrelled?

2.

A survey is anonymous when:

3.

Which response options for weekly work hours are best?

4.

What is the main purpose of a pilot test?

5.

Which question is the most ethically risky for a student survey?

6.

Informed consent means respondents:

For parents and teachers

Big idea (E1.2, E1.3): an ethical survey protects people and produces honest, usable data. Each question measures one thing in neutral words and connects to a planned analysis.

How to help: Volunteer as a pilot tester. Answer the survey while your teen watches, then say honestly which questions confused you or felt too personal. Ask them, "Which part of your analysis uses this question?" for each item. Make sure they have the school's permission before collecting any data.

Common difficulties:

  • Questions that ask for opinions when numbers are needed, or numbers grouped in overlapping intervals.
  • Asking for names or details that make a survey identifiable.
  • Too many questions ("just in case"), which lowers response rates.
  • Sensitive topics: steer these toward published secondary data.

Safety note: a probability game must not involve real-money gambling; points or small treats (with school permission and allergy awareness) are fine.