When planning a survey, one practical question comes up early: should you collect data online or on paper? Online surveys are fast, cheap and easy to analyse. Paper surveys feel old-fashioned, yet many researchers still use them, and for good reasons. The honest answer is that neither mode is always better. The right choice depends on your population, your topic, your resources and the kind of data quality you need.
This post compares online and paper surveys across the factors that matter most for data quality, explains mixed-mode designs, and gives you a simple way to decide which approach suits your study.
Why survey mode matters
The survey mode is the way questions are delivered and answered. Mode affects:
- Who can be reached (coverage).
- Who chooses to respond (response rate and non-response bias).
- How people answer (measurement effects, such as honesty on sensitive questions).
- How complete and clean the data is (missing values, errors, data entry).
- Cost and time.
Survey methodologists often describe these as sources of total survey error. A good mode choice reduces the most important errors for your particular study.
Online surveys: strengths
Speed and cost
Online surveys can reach many people quickly with low printing and postage costs. Responses arrive in real time, and there is no manual data entry.
Built-in data quality features
Survey platforms allow:
- Skip logic and branching, so respondents see only relevant questions.
- Validation rules, such as numeric ranges or required formats.
- Randomisation of question or response order to reduce order effects.
- Progress bars and save-and-continue options.
- Attention-check items and timing data to identify careless responding.
Reduced data entry errors
Because respondents enter data directly, you avoid transcription mistakes that can occur when paper forms are typed into a spreadsheet.
Privacy for sensitive topics
Self-administered surveys, both online and paper, generally encourage more honest answers to sensitive questions than interviews do, because no interviewer is present. Online surveys completed privately can support this, provided respondents trust the security of the platform.
Multimedia and flexibility
You can include images, videos or interactive elements, and easily reach geographically dispersed populations.
Online surveys: weaknesses
Coverage problems
Not everyone has reliable internet access, digital skills or comfort with online forms. Older adults, people in rural areas, low-income groups and some professional groups may be under-represented. If these groups differ from others on your key variables, your results may be biased.
Lower response rates in some settings
Email invitations can be ignored, filtered as spam or lost in busy inboxes. Response rates for online surveys vary widely and are often lower than for well-conducted paper surveys, particularly when there is no existing relationship with the respondents.
Sampling challenges
Open survey links shared on social media produce convenience samples with unknown selection processes. You cannot calculate a true response rate, and the same person may respond more than once unless you control for this.
Careless and fraudulent responses
Online surveys, especially with incentives, may attract speeders, straight-liners and automated bots. You need screening procedures, such as attention checks, completion-time checks and duplicate detection.
Device differences
Many people complete surveys on mobile phones. Long grids and complex layouts can be difficult on small screens, leading to more dropout or errors. Always test on multiple devices.
Paper surveys: strengths
Better coverage for some populations
Paper surveys can reach people without internet access or with limited digital skills. In community settings, schools, hospitals, factories or rural areas, paper may be the most inclusive option.
Higher response in supervised settings
When you distribute paper questionnaires in person, for example in a classroom or staff meeting, you can often achieve high response rates. Respondents can ask questions, and you can collect forms immediately.
Familiarity and trust
Some respondents trust paper more than online links, especially where phishing and scam messages are common. A printed letter with a university logo can signal legitimacy.
Fewer technical problems
There are no login issues, broken links or device compatibility problems.
Paper surveys: weaknesses
Cost and time
Printing, postage, travel and manual data entry take time and money. Large samples can become expensive.
Data entry errors
Typing responses into software introduces the risk of errors. Double data entry, where two people enter the same forms and discrepancies are checked, reduces this risk but increases workload.
More missing and inconsistent data
Without automatic skip logic and validation, respondents may skip questions accidentally, answer questions that do not apply to them, or tick two options on single-choice items.
Less flexibility
Complex branching and randomisation are difficult on paper, and changes after printing are costly.
Does mode change the answers people give?
Yes, it can. Research on mode effects has shown that the same question can produce different answer distributions depending on how it is delivered. Some well-documented patterns include:
- Social desirability tends to be higher in interviewer-administered modes than in self-administered modes (online or paper).
- Visual layout matters in self-administered modes: the spacing of response options, the use of grids and the position of the midpoint can all influence answers.
- Primacy effects (choosing options near the top of a list) are more common in visual modes, while recency effects are more common when options are read aloud.
Between online and paper, differences are often smaller because both are visual and self-administered. However, layout differences between the two versions can still create measurement differences, so design both versions to look as similar as possible if you use both.
Mixed-mode designs
Many studies combine online and paper modes to get the benefits of both. Common approaches include:
- Sequential mixed mode: invite everyone online first, then send paper questionnaires to non-respondents.
- Concurrent choice: offer both options at the same time and let respondents choose.
- Mode by group: use paper for groups without internet access and online for others.
Mixed-mode designs can improve coverage and response, but they add complexity. You need to:
- Keep question wording, order and layout as similar as possible across modes.
- Record which mode each respondent used.
- Check whether responses differ by mode, and consider mode as a control variable in analysis.
- Describe the design clearly in your methodology chapter.
How to decide: key questions
1. Who is your population?
If almost all members have reliable internet access and are comfortable online, such as university staff or IT professionals, online is often suitable. If access is uneven, consider paper or mixed mode.
2. How will you contact them?
If you have email addresses from a sampling frame, online is easy. If you only have postal addresses or can meet people in person, paper may work better.
3. How sensitive is your topic?
Both online and paper self-administration are generally better than interviews for sensitive topics. Focus on credible confidentiality assurances in either mode.
4. How complex is your questionnaire?
Many skip patterns, randomisation or experimental conditions favour online delivery.
5. What are your resources?
Consider budget, time, printing, postage, travel and data entry capacity.
6. What does your field expect?
Look at recent studies with similar populations and see which modes they used and what response rates they achieved.
A quick comparison
- Cost: online lower; paper higher.
- Speed: online faster; paper slower.
- Coverage of digitally excluded groups: paper better.
- Data entry errors: online fewer; paper more (unless double-entered).
- Skip logic and validation: online much stronger.
- Response rates: depends heavily on population and contact method; supervised paper distribution is often high.
- Risk of bots and fraud: higher online, especially with open links and incentives.
Practical tips to improve data quality in each mode
If you choose an online survey
- Use a personalised invitation sent from your university email address, rather than a generic link posted publicly, whenever you have a sampling frame.
- Test on phones and tablets. Replace large matrix grids with shorter item blocks where possible, because grids are hard to read on small screens.
- Add one or two attention checks, such as “Please select ‘agree’ for this item”, and decide in advance how you will treat respondents who fail them.
- Record completion time and set a rule for identifying unrealistically fast responses, based on your pilot timings.
- Prevent duplicates by using unique links or platform settings that limit one response per person, where this is compatible with your anonymity promises.
- Protect against bots if you use open links, for example by using the platform's built-in bot detection and by checking open-ended answers for nonsense text.
- Send reminders at planned intervals, usually one to two reminders after the first invitation.
If you choose a paper survey
- Design a clean layout with clear numbering, enough white space, and obvious instructions for skipping questions.
- Use ID numbers on each form (not names) so that you can track entry and check errors.
- Plan data entry carefully. Create a codebook before entering data, define codes for missing values, and consider double entry for at least a sample of forms.
- Provide a return method that is easy and confidential, such as sealed envelopes and a locked collection box, or prepaid return envelopes for postal surveys.
- Store forms securely according to your ethics approval and data management plan.
If you use both
- Create the paper version from the final online version (or vice versa) so that wording and order match exactly.
- Add a variable in your dataset for survey mode.
- Compare key variables and missing data rates between modes before combining them.
These small steps take little time compared with the whole project, but they can make a large difference to the quality and credibility of your data.
Reporting your choice in the thesis
In your methodology chapter, explain:
- Which mode or modes you used and why.
- How the mode fits your population and sampling frame.
- Steps taken to protect data quality, such as attention checks, double data entry or pilot testing on mobile devices.
- Response rates by mode, if mixed mode was used.
- Any checks for mode differences.
- Limitations related to coverage or mode effects.
Final thoughts
Online surveys are efficient and powerful, but they are not automatically better. Paper surveys can reach people that online surveys miss and can achieve high response rates in the right settings. The best mode is the one that reaches your target population fairly and collects accurate answers within your resources. Think carefully about coverage, response, measurement and cost, pilot test your chosen approach, and justify your decision clearly. That is what examiners and reviewers want to see.
Which mode are you planning to use, and why? Share your situation in the comments.