00 Plan your survey
A survey study asks a defined group a set of structured questions. It's cheap and fast, but the quality of your questions is the quality of your data — so most of the work happens before you send anything.
What it looks like
- Shape: question → sampling plan → instrument (with the full question wording in an appendix) → response rate → results → bias discussion
- Paper structure: abstract → introduction → methods → results → discussion → limitations → conclusion → references → appendix
- Common mistake: leading questions, and a convenience sample presented as representative. Say exactly who answered and who couldn't.
Start with a question a survey can answer
Surveys are good at measuring what people think, feel, report doing, or have experienced — and how those things relate to each other. They can't directly observe behavior or prove cause and effect.
How do phones affect teenagers?
Among students at two local high schools, is self-reported evening screen time associated with sleep quality?
Define who you're surveying
- Population — the whole group you want to learn about (all students at two high schools)
- Sample — the people who actually receive your survey
- Respondents — the people who actually answer it
The closer your respondents are to your population, the more your results can say about it.
Choose a sampling method
| Method | How it works | Trade-off |
|---|---|---|
| Random | Everyone in the population has an equal chance of being picked | Most representative, but you need a full list of the population |
| Stratified | Randomly sample within groups, such as each grade level | Ensures every group is included; takes more planning |
| Convenience | Survey whoever is easy to reach — your classes, your clubs | Easiest by far, but results may not represent anyone else |
| Snowball | Respondents share the survey with others like them | Reaches hard-to-find groups; the sample can become one-sided |
Most student surveys use a convenience sample, and that's fine — as long as you say so honestly and don't claim your results describe everyone.
Decide how many responses you need
Set a target before you start. For comparing groups, a common rule of thumb is at least 30 responses per group; for describing a population, more is better. Track your response rate — the share of people you invited who actually answered.
Get ethics right
- Start the survey with a short consent statement: what it's for, how long it takes, that it's voluntary, and how answers will be used
- Get parental consent for anyone under 18, if your school requires it
- Keep responses anonymous — don't collect names or emails unless you truly need them
- Let people skip questions and stop at any time
- Ask your teacher whether your school requires approval before you begin
- Ask about self-harm, abuse, or illegal activity without adult oversight and support resources
- Collect names alongside sensitive answers
- Share individual responses — report only group results
Pilot it
Give your survey to three to five people like your respondents and watch them take it. Ask which questions were confusing. Every survey improves after a pilot.
Build a timeline
Weeks 1–2 · Read and plan
- Background reading, research question, hypothesis, sampling plan
Week 3 · Build
- Write questions, find validated scales, set up the form, get approval, pilot
Weeks 4–5 · Collect
- Send the survey, send one reminder, and track your response rate
Weeks 6–8 · Analyze and write
- Clean the data, run your analysis, draft every section, then revise
01 Write good questions
The single biggest factor in a survey's quality. Every question should measure exactly one thing, the same way for everyone.
Use an existing scale when you can
Researchers have already built and tested scales for many things — stress, sleep quality, self-esteem, loneliness. A validated scale has been shown to measure what it claims to, so using one makes your results far more trustworthy than questions you write yourself. Search Google Scholar for “[your topic] scale” and cite the scale in your paper.
Pick the right question type
- Multiple choice — for facts and categories (grade level, how often)
- Rating scales — for opinions and attitudes, usually five points from “strongly disagree” to “strongly agree”
- Numeric — for amounts (hours of sleep last night)
- Open-ended — for explanations in people's own words; use sparingly, since they take longest to answer and analyze
Fix the four classic mistakes
Don't you agree that phones keep students up too late?
On a typical school night, how many minutes do you use your phone after getting into bed?
How satisfied are you with your sleep and your energy during the day?
Split into two questions: one about sleep, one about daytime energy.
Do you use your phone a lot at night?
In the last 7 days, on how many nights did you use your phone within 30 minutes of trying to fall asleep?
0–1 hours · 1–2 hours · 2–3 hours
Less than 1 hour · 1 to less than 2 hours · 2 to less than 3 hours · 3 hours or more
Keep it short — aim for under ten minutes. Put easy, non-sensitive questions first, group related questions together, and save demographic questions for the end.
02 Title page
APA 7 puts a running head at the top of every page, in capitals, with the page number.
RUNNING HEAD: SHORTENED TITLE OF STUDY 1
- Title of the study
- Your name
Teens and Sleep
Evening Screen Time and Sleep Quality Among High School Students: A Survey Study
03 Abstract
A one-paragraph summary of the whole study, including your actual results.
The formula
- The topic and why it matters
- Your research question or hypothesis
- Who you surveyed, how many responded, and what you measured
- Your main results, with key numbers
- What they suggest
Many adolescents report poor sleep, and evening phone use is often blamed. This study examined whether self-reported evening screen time is associated with sleep quality among high school students. An anonymous online survey was completed by 120 students at two public high schools (response rate 40%), using an established sleep-quality scale and questions about phone use before bed. Students who reported more screen time after getting into bed reported worse sleep quality. Although the cross-sectional design cannot show cause and effect, the results support encouraging students to limit phone use at bedtime.
Keywords: screen time, sleep quality, adolescents, survey
- Leaving out the sample size or response rate
- Claiming your results apply to all teenagers
- Writing more than one paragraph, or more than 150–250 words
- Citing sources in the abstract
Write your abstract last, once every other section is finished.
04 Introduction
Move from broad to narrow: the topic, what research already shows, the gap, and finally your question and hypothesis.
Funnel structure
1. The big picture
- Introduce the topic and why it matters
- Name the real-world issue your research speaks to
2. What we already know
- Summarize and compare the key studies, organized by theme
3. The gap
- What previous research hasn't measured, or hasn't measured in your population
- How your study builds on what came before
4. This study
- Your research question, your hypothesis, and a sentence on how you surveyed people
- Any limits on scope
5. Roadmap
- One or two sentences on how the rest of the paper is organized
Theoretical framework
Define each concept you measure — researchers call these constructs — and name the theory that connects them.
Sleep quality refers to a person's own assessment of how well they sleep, including how long it takes to fall asleep and how rested they feel. Research suggests that light from screens and mentally stimulating content can delay sleep, which predicts that more screen time in bed will be associated with worse sleep quality.
Hypothesis
Students who report more screen time after getting into bed will report lower sleep quality than students who report less.
Use APA 7 for references and in-text citations, and cite every piece of information you take from a source.
05 Methods
Describe exactly what you did, in the past tense, so a stranger could repeat your survey without asking you anything.
Participants and sampling
Your population, how you chose your sample, how many people you invited, how many responded, and your response rate. Describe who your respondents were (grade, age range) and who couldn't take part.
The survey link was shared with all 300 students enrolled in 10th- and 11th-grade English classes at two public high schools. Of these, 120 completed the survey (response rate 40%). Respondents were 15–17 years old. Students absent during the two-week collection period did not receive in-class reminders.
Instrument
Describe your survey: how many questions, what they measured, and which validated scales you used, with citations. Put the full survey, with exact wording, in an appendix.
Procedure
How and when you distributed the survey, how consent was obtained, how long it took, and how many reminders you sent.
Data analysis
How you cleaned the data (for example, removing incomplete responses), which statistics you calculated, and which software you used. If you asked open-ended questions, explain how you coded them into themes.
Percentages and averages describe your data. To test a relationship, use a correlation for two numeric answers, a t-test to compare two groups, or a chi-square test for two categorical answers. Google Sheets, JASP, and jamovi are all free.
06 Results
Report what you found, clearly and without interpretation. Save the “why” for the discussion.
Include
- Your response rate and a short description of who responded, often in a table
- Descriptive statistics for each key question — percentages, means, standard deviations
- The result of each test you said you'd run
- At least one clear chart or table, with a numbered title
- How many people answered each question, if it varied
Most students use their phones in bed, which is why they sleep badly.
Seventy-two percent of respondents (n = 86) reported using their phone after getting into bed on at least five of the previous seven nights. Minutes of screen time in bed were negatively correlated with sleep-quality scores, r(118) = −.32, p < .001.
- Interpreting or explaining results here
- Reporting percentages without the number of people behind them
- Listing every question's results when only some answer your research question
- Leaving out results that didn't support your hypothesis
07 Discussion
Explain what your results mean, moving from your findings back out to the bigger picture.
Answer these, in order
- Was your hypothesis supported? Restate the main finding in plain words.
- How does it compare to previous research? Where do you agree or disagree, and why might that be?
- What explains it? Connect your results to the theory in your introduction.
- What else could explain it? Consider alternative explanations — especially other variables you didn't measure.
- Why does it matter? Practical implications and recommendations.
A survey shows that two things are related, not that one causes the other. Students who sleep poorly might use their phones more because they can't sleep. Say “is associated with,” not “causes.”
08 Limitations
Every survey has blind spots. Naming yours honestly makes your paper more credible.
Consider
- Sampling bias — who was left out, and whether your respondents represent your population
- Nonresponse bias — whether the people who answered differ from those who didn't
- Self-report bias — people misremember, estimate, or give answers that sound better (social desirability)
- Question wording — any questions respondents may have misunderstood
- Timing — a one-time survey captures a single moment
The sample may not be representative.
Students who chose to respond may have been more interested in sleep than those who didn't, and screen time was self-reported rather than measured, so it may have been underestimated.
09 Conclusion
Bring the paper full circle. Remind the reader what you set out to do, then show that you did it.
Four steps
- Restate your aim and research question.
- Answer it. Summarize your main findings.
- State your contribution. What your study adds, including any recommendations.
- Point forward. How future research could build on it — for example, with a larger random sample or an experiment.
- Introducing new results or sources
- Repeating the introduction or abstract word for word
- Claiming your results apply beyond the people you surveyed
10 References & appendix
References in APA 7, followed by your full survey.
References
- Only include sources you actually cited
- Cite every validated scale you used, not just articles
Keegan, J. (1999). The First World War. Alfred A. Knopf.
Appendix: your survey
Include every question exactly as respondents saw it, in order, with all answer options and your consent statement. Readers can't judge your results without seeing your wording.
11 Checklist
Run through this before you send your survey, and again before you submit. Your ticks are saved in this browser.
Before you send
The paper
The document