A description of a research instrument should let the reader establish how data were collected, what was measured and how responses became results. Saying “an author-developed questionnaire was used” is not enough. Explain the instrument’s origin and version, structure, response options, use and limitations. This guide shows how to describe an original questionnaire and an existing scale without claiming measurement properties that have not been investigated.
What should a research instrument description include?
A useful description connects the instrument’s design with the particular study. Readers should understand why its questions address the research question, whom they were intended for and how the information can be interpreted. This section does not need to reproduce the whole methodology chapter. Explain aims, sampling and fieldwork more fully in their own subsections, while making the necessary connections here.
- Name and origin: developed for your study, an existing instrument, or a modified version.
- Purpose and coverage: the characteristics, opinions, experiences or behaviours to be described.
- Version: language, edition, instruction source and identification of the file actually used.
- Structure: sections, numbers of questions or items, routing, response options and completion time, if established.
- Scoring and analysis: response codes, total or subscale scores, missing data and limits of interpretation.
- Preparation: checks actually completed, revisions, conditions of use and the location of supporting documentation.
Start by collecting the documents rather than polishing a paragraph. Open the form actually distributed, the response export and the scoring instructions. Check whether the question count matches the instrument. If routing meant different participants saw different questions, explain this rather than presenting one apparently precise count as applicable to everyone.
If you are still defining aims, research questions and variables, first read the guide to planning your research methodology. This article addresses the next task: documenting the specific instrument you plan to use or actually used. It does not replace a scale author’s instructions or your seminar’s requirements.
An original questionnaire or an existing scale: state its actual origin
An original survey questionnaire is developed for your particular project. Literature can inform the areas covered, but distinguish inspiration from a concept from copying specific wording. “Developed by the author based on the literature” should refer to identified publications and clarify what was independently created. Entering someone else’s questions into a form does not make those questions your own.
An existing scale is an identified instrument with its own name, intended use and documentation. Not every questionnaire found online is a standardised scale. Previous use in another thesis does not establish its quality. Consult the original publication or manual instead of relying entirely on another student’s description.
If you combine a published scale with your own background questions, describe the two components separately. Readers need to know which items belong to the original instrument and contribute to its score, and which only describe participants. Do not add age, degree subject or an original question to a scale score without a justified procedure.
How to describe the structure of your own questionnaire
First identify the subject being investigated and show the questionnaire’s logical sections. One section might cover library use, another the perceived availability of materials, and a third participants’ background characteristics. Include item numbers if they help readers compare the description with an appendix. Section names should not imply measurement of a broader concept than the questions actually address.
Distinguish closed single-choice questions, multiple-response questions and open questions. For multiple responses, state whether participants could select any number of options or faced a limit. Explain who saw conditional questions. State whether answers were mandatory, because preventing participants from skipping questions affects how data were collected.
For rating options, report the meanings of the points, their direction and the reference period. “A scale from 1 to 5” does not explain whether 1 means “never”, “strongly disagree” or “very poor”. A single frequency question does not automatically become a multi-item psychometric scale. Using five options also does not justify adding all questionnaire responses together.
Illustrate the connection between items and research questions with one concrete example. When investigating access to literature, frequency of library visits provides context, whereas perceived ability to obtain a needed book addresses availability itself. These are different kinds of information. Do not label the whole section “information literacy level” if the instrument only collects self-reported opinions.
What to check in an existing instrument’s documentation
Identify the full name, original authors, authors of the adaptation used and exact language version. Check the intended population and documented applications. An instrument used with adult employees may not have the same justification for studying adolescents. A similar name and identical item count do not establish that two files contain the same version.
- Keep the source of the questionnaire and instructions, including an access date for online materials.
- Check the licence, permission or other conditions for use, reproduction and inclusion in an appendix.
- Establish any user qualification requirements or requirements to obtain a manual.
- Compare the instructions, items, response options and reference period with the version actually used.
- Separate the original authors’ findings and norms from calculations performed on your own sample.
If documentation is missing, do not replace it with guesswork. Identify what is unavailable and discuss the choice before collecting data. Without scoring instructions, you may be unable to calculate the score intended by the authors. Unclear conditions of use need clarification from the appropriate rights holder. A publicly accessible file is not, by itself, evidence of an unrestricted licence.
Translation, shortening and changing items need separate documentation
The International Test Commission guidelines for adapting tests address appropriate permissions, documentation of changes and evidence of comparability. Translation alone does not establish equivalence. Name the modification and justify its scope; do not automatically transfer the original instrument’s validation findings to an altered version.
This also applies to seemingly small edits: removing an awkward item, changing five response options to four, replacing “the past two weeks” with “usually”, or converting interviewer administration into a self-completed form. Keep a version record and discuss the implications with your supervisor. Do not call a shortened questionnaire an “official version” if it was created only for your project.
Scoring, coding and missing responses
Explain what becomes the analysed variable. In an original questionnaire, individual questions may be analysed separately without a total score. An existing scale may produce a sum, mean or separate subscale scores according to its documentation. State the rules clearly enough to trace the path from the response file to the variable used in analysis.
A technical example unrelated to any particular test: for responses coded 1, 2, 3, 4, 5, reversing direction can be expressed as a new value equal to 6 minus the original value. Thus 1 becomes 5 and 2 becomes 4. This applies only when reversal is required by the scoring key. Do not reverse an item merely because its wording includes a negative.
A missing response is not zero. Distinguish missing data, “not applicable”, “do not know” and the genuine lowest value. If the instructions permit scoring with a specified amount of missing data, state that condition. Otherwise, do not invent a rule after inspecting the results. For routed questions, explain that a designed skip is different from refusing to answer.
A higher score may indicate more of a characteristic, but it does not automatically justify “low”, “medium” and “high” labels. Such boundaries need a basis. Do not divide the possible range into thirds and present this as the authors’ norms. If you create categories for descriptive purposes, identify and justify them while preserving the original scoring information.
Piloting and checking comprehension do not replace validation
Before the main study, check whether participants understand the questions and can select an answer. CDC describes cognitive interviewing as a way to examine the response process. Ask someone testing the form how they understand a term and why they selected an option. This can reveal a mismatch between the author’s intention and the respondent’s interpretation.
A trial run also checks form operation: routing, mobile display, response saving and export. The U.S. Census Bureau presents different questionnaire pretesting methods serving different purposes. There is no single mandatory pilot sample size given here for every study. Match the scope to your project and describe it explicitly.
Record who checked the instrument, what was examined, what feedback was collected and what changed before the final version. Do not claim a pilot that never took place. If only your supervisor reviewed the form, describe a consultation rather than a participant pilot. If trial responses entered the main dataset, explain that decision and the compatibility of the versions used.
Reliability and validity: what can your own sample support?
Internal consistency is one aspect of assessing a multi-item scale. Cronbach’s alpha alone does not establish validity, unidimensionality or suitability for every use. The COSMIN 2.0 manual distinguishes these properties. It primarily concerns reviews of patient-reported outcome measures; it is not a universal checklist of compulsory analyses for every student questionnaire.
In practice, do not calculate one alpha for unrelated questions about age, visit frequency, device type and opinion. First establish whether the items are intended to form one scale and whether the measurement approach justifies that assessment. Stating that questions were analysed separately may be more defensible than adding an arbitrary coefficient “for the whole questionnaire”.
Separate two levels of information: published properties of the particular instrument version and results obtained from your data. For published findings, cite the source and study context. For your own findings, describe your sample and procedure. Do not present a manual’s coefficient as if you calculated it in your thesis, or invent a missing value to complete a template.
Example 1: describing an original survey questionnaire
The following is a fictional writing example, not a report of a real study or a questionnaire ready to copy. Replace the numbers and scope with information from your own project. Notice that it does not claim norms, diagnostic use or established validity.
Students’ opinions on the availability of library materials were collected using an original survey questionnaire, version 1.2. It contained 18 questions: four about library use, ten about perceived availability of materials and four describing participants. It included single-choice questions, two multiple-response questions and one open question. Availability ratings used the options very poor, poor, neither good nor poor, good and very good. The question about ordering books from outside the library was shown only to participants reporting use of that option. Questions were analysed separately; no overall availability score was calculated. Before the study, five members of the intended population checked question comprehension and form operation. The term electronically available material was clarified and one routing rule was corrected. These checks did not constitute full psychometric validation. The final questionnaire and coding rules are provided in Appendices A and B.
This description makes the design traceable, but a real thesis also needs sources supporting the areas covered. If five people did not test the questionnaire, remove that claim and report the checks actually performed. Attach version 1.2 itself, not a newer questionnaire revised after data collection. A version number is useful only when it identifies an actual file.
Example 2: describing the use of an existing scale
The second example is a template with fields to complete. It does not invent a test name, authors or coefficients. Enter only information confirmed in the documentation for the version used. If there are no subscales or norms, do not add them merely because another thesis template includes them.
[Specific construct] was measured using [full name and abbreviation], developed by [authors, year], in [language, version, adaptation authors and source]. The choice was justified by [fit with the study aim and population]. The instrument contains [number] items forming [total score or specified subscales]. Respondents answer using [full response labels and reference period]. Scores were calculated according to [instruction source], including [applicable reverse-coded items and missing-data rule]. Higher scores indicate [meaning stated in the documentation]. The study used [the unchanged version or an exact description of modifications]. Conditions of use were checked against [licence or confirmation]. Evidence on measurement properties comes from [identified publication and population]. In the study sample, [analyses actually performed, or a statement that they were not conducted]. The instrument was used to support [the scope of inference in this thesis].
In the final text, replace the fields with a continuous description. Remove inapplicable alternatives, supply real references and check each statement against the instructions. This template organises documentation; it does not establish that a scale is appropriate for your study. Justify that decision before using the instrument, not only when writing the chapter.
Common mistakes: check the description before submitting your chapter
- Too little detail: “the research used a survey”. Improve it by identifying the questionnaire, its origin, structure and use.
- An unsupported label: “an original, standardised questionnaire” without a standardisation process. Describe the actual preparation and limitations instead.
- Misattribution: “alpha was 0.85 in this study” when the value comes from a publication. Attribute it to the source and the relevant population.
- Undisclosed changes: “the original was used” although items were removed. Identify the changes and their implications.
- Apparent validation: “the instrument is valid because several people checked it”. Report comprehension checks without overstating their meaning.
- Inconsistent files: 20 questions in the description, 22 in the form and 24 export columns. Distinguish questions, sub-items and technical columns.
- An unclear score: “more points means better”. Explain what the direction means and the basis for interpreting it.
Before submission, compare the description with the instrument line by line. Verify the title, date and version, authors and sources, item count, instructions, responses, routing, coding and appendices. Finally ask whether a reader can understand how each result in the findings chapter was produced. If not, supply the missing step rather than expanding a general definition of a survey.
The appendix, form and dataset must refer to the same instrument
The appendix can show the instrument, while the chapter explains its use. Do not replace the description with “the questionnaire is attached at the end”. Equally, do not copy every question into the main text. Refer to the appendix when discussing structure and retain the information needed to understand measurement in the chapter. Include the full text of an existing instrument only when its conditions permit this.
For an online survey, retain a copy of what participants saw: instructions, question layout, options and routing. A screenshot of the first page does not document the whole instrument. A response export is not a substitute either, because it can omit instructions and full labels. Identify the platform as the delivery method rather than the questionnaire’s author. Google Forms is not a methodology or evidence of measurement quality.
Prepare a data dictionary for the analysis. For each column, record the question number, code meanings and missing-value labels. Several columns may represent one multiple-response question, while several items may produce one subscale variable. This explains why column counts need not equal question counts. A dictionary makes it easier to check scoring and audit the calculations described later.
Separate public instrument documentation from participants’ responses. Email addresses, names, identifiable open responses and technical identifiers should not automatically enter an appendix. Use appropriately anonymised materials for consultation. However, do not silently remove methodological information such as recruitment, routing or reasons for excluding responses. Protecting privacy and documenting the study transparently require separate decisions.
Finally, compare three files: the final questionnaire, data dictionary and instrument description in the chapter. All should refer to the same version. If you discover a discrepancy after fieldwork, record it and assess its implications; do not edit the archived form to make it appear consistent from the outset.
Frequently asked questions
Can I use a ready-made survey?
Yes, if it fits the objective and the group being studied, and its terms of use permit this. Check the source, conditions of use, and method of interpretation. A questionnaire found online does not become a validated instrument just because another student has used it.
When is scale reliability analysed and regression applied?
For multi-item questionnaires, scale consistency can be assessed, for example using Cronbach's alpha coefficient. An alpha result is not standalone evidence of the tool's validity. If the question involves several predictors, I will check the possibility of applying regression. I will define the scope after evaluating the sample size, variables, and model assumptions.
Can you check an already written methodological chapter?
Yes. I can analyse its structure, the link between the objectives and research questions, and the consistency of the descriptions of the methods and research instrument. Include your supervisor's comments and the questionnaire or interview guide itself so that I can assess the whole, rather than the description alone.
What should a research instrument description in a thesis include?
State the instrument’s name and origin, purpose, language version, structure, response options and scoring rules. Explain its conditions of use, checks performed and limitations. Describe the questionnaire actually used rather than a revised version created afterwards.
Does piloting a questionnaire mean the instrument has been validated?
No. Trial completion can help check question comprehension, response options and form operation. It does not automatically establish all measurement properties. Report what the pilot actually covered and the resulting changes; do not call it full validation without appropriate studies.
Can my own translation of an existing scale be described as its Polish adaptation?
Do not equate translation alone with a documented adaptation and established equivalence. First check for an existing language version and the conditions for use and modification. If your project uses your own translation, describe how it was actually prepared and its limitations. Do not automatically attribute the original instrument’s properties to it.
Need help checking your instrument and its description?
Send the questionnaire in the version actually used or planned, your thesis topic, aim, research questions, any existing scale documentation and your supervisor’s comments. State whether data collection has already started. This makes it possible to distinguish changes still feasible before fieldwork from limitations that need an honest account afterwards.
For connecting the description with the overall research design, see Help with your thesis methodology chapter. To discuss developing or checking the questions themselves, see Questionnaire design for your thesis. If the issue concerns scoring, dataset structure or calculations, the appropriate scope can be agreed after reviewing your materials: Statistical analysis for bachelor’s and master’s theses.
Send your instrument and explain what you need
Sources and how they are used
The following materials support the distinctions discussed in this article. They do not replace a particular scale’s manual or university requirements. The writing examples and checklist are StudentSupport’s editorial work. Sources were checked on 22 October 2026.
- CDC, CCQDER. Cognitive Interviewing.
- U.S. Census Bureau. Appendix A2: Questionnaire Testing and Evaluation Methods for Censuses and Surveys.
- International Test Commission (2017). The ITC Guidelines for Translating and Adapting Tests, second edition, version 2.4.
- Mokkink, L. B., Elsman, E., Terwee, C. B. (2024). COSMIN Manual, version 2.0.
