Create a Google Form Survey for a Quantitative Thesis

If you need to create a Google Form survey for distributing a questionnaire for your thesis, start with the research model and indicators instead of opening Google Forms and entering questions immediately. A polished form with no screening questions, the wrong scale, or no control over responses will leave you with a difficult data file to clean later.

The process below follows the steps you actually need to complete: determine the sample size, design the questionnaire, create the form, collect responses, and prepare the data before importing it into SPSS. For more guidance on choosing a tool, you can also see the Survey topic, online survey questionnaires, and survey forms.

What a survey means in quantitative research

In quantitative research, a survey collects structured data from a group of people through a standardized questionnaire. Each respondent sees broadly similar questions, selects coded response options, and contributes data that can be organized into columns for analysis in SPSS, SmartPLS, or another statistical program.

For a thesis, Google Forms is only the tool that displays the questionnaire and collects responses. It does not create a scale, prove that the questionnaire is valid, or replace Cronbach's Alpha, EFA, or regression model testing. You still need to identify the independent variables, dependent variables, indicators, and coding scheme before creating the form.

A quantitative questionnaire usually contains an introduction, screening questions, respondent classification information, and groups of indicators. If your study measures attitudes or agreement, a Likert scale is a common choice. The Likert scale was introduced in Likert's work (Likert, 1932), but the number of response levels should match your research design and your supervisor's guidance.

How to determine the survey sample size

Decide on the sample size before distributing the form. You can consider the number of indicators and the planned analysis, then compare them with the research population, access to respondents, and your supervisor's requirements. For factor analysis, a commonly used rule is 5 to 10 observations for each indicator (Hair et al., 2010).

If you run a regression with m independent variables, a reference rule is a sample size of at least 50 + 8m (Tabachnick and Fidell, 2013). For example, a model with 5 independent variables requires at least 90 observations under this formula. This is only a starting point because you also need to allow for invalid questionnaires and the possibility that indicators will be removed during analysis.

When the population is finite and its size N is known, the Yamane formula is often used in Vietnamese theses:

n = N / (1 + N × e²)

Here, n is the required sample size, N is the population size, and e is the allowable error. This formula is attributed to Yamane (1967). Do not use this formula while also taking another number from a sample thesis without explaining how you chose between them.

You should collect more responses than the minimum sample size. If you need 200 valid questionnaires, targeting 220 to 240 responses can help cover incomplete responses, unusually fast answers, or questionnaires that fail the screening criteria. In Chapter 3, state the minimum sample size, the number of questionnaires distributed, the number returned, and the number considered valid.

Design the questionnaire for this research objective

Before creating the form, prepare a variable codebook. For example, the construct “Perceived Usefulness” may contain the indicators PU1, PU2, and PU3, while “Intention to Use” contains IU1, IU2, and IU3. Each indicator needs specific wording, a clear response option, and an identifiable code.

Write a short introduction with participation criteria

The opening section should state the research topic, the time required, the academic purpose of the data, and the participation criteria. Avoid a long introduction that causes respondents to leave before reaching the main items.

If your study concerns digital banking app users, a screening question might ask whether the respondent has used the app within a specified period. Respondents who do not meet the criteria should be sent to the end of the form instead of answering questions that do not apply to them.

Choose the correct question type in Google Forms

Use Multiple choice when respondents may select only one option, Checkbox when they may select several options, and Linear scale or Multiple choice grid when measuring agreement. For groups of indicators, consider how the data will be exported to a spreadsheet so that each indicator remains a separate column.

Do not place too many indicators in one grid when the wording is long or respondents mainly use phones. Check the preview on a small screen. A form that is difficult to read increases the likelihood that respondents will abandon it or select the wrong row.

Keep the Likert scale and coding consistent

A 5-point Likert scale can be coded from 1 to 5, with 1 meaning strongly disagree and 5 meaning strongly agree. If you use 7 levels, keep the same logic across the entire measurement section. Do not make 1 mean “very dissatisfied” in one section and “very satisfied” in another.

If you have reverse-coded items, mark them immediately in the variable codebook. When importing the file into SPSS, reverse-coded items need to be recoded before reliability testing and factor analysis. A note in Google Sheets does not replace the recoding step in the analysis file.

Test the form before sending it widely

Send the form to a few people whose characteristics are close to those of the research sample. Check the completion time, spelling, required questions, section branching, and exported data structure. If you edit a question after real data has already been collected, record when the edit was made so you can determine whether the versions remain comparable.

You can also consult the guides on creating a survey form and Google Form surveys when you need to compare question types and basic settings.

Deploy the data collection

After testing the form, choose channels that reach respondents who match the study criteria. You may share the link through class groups, professional communities, email, or social media, but state clearly which population is being invited. A link shared too widely can produce a convenience sample that does not match the research population.

Google Forms works well when you need to create a form quickly, use it for free, and export the data to Google Sheets. You still need to check access permissions, the response status, and the setting that limits each person to one response. If you collect email addresses, consider whether your study requires anonymity and tell participants clearly how their information will be used.

DoThesis has fillform.info, our own form tool for collecting responses to Vietnamese questionnaires. It is an alternative when you want to manage the survey process in one place. Google Forms remains a suitable option if you already work comfortably within the Google ecosystem.

During data collection, track the number of valid responses rather than looking only at the total number of submissions. A response may be invalid because the respondent is outside the target population, left many items blank, or selected the same option throughout the questionnaire. Record the start date, end date, distribution channels, and number of responses per day for the methods section.

Do not change questions on your own while collecting data without discussing the change with your supervisor. If a change is necessary, save a copy of the data and record how many responses belong to the old version and how many belong to the new version.

Clean the data before importing it into SPSS

When you close the form, download the data as Excel or CSV and preserve an untouched original copy. Use a duplicate for cleaning. Do not delete data directly from the original file because you may later need to check a case that was removed.

Remove responses with too much missing information

Check the indicator columns for rows with many blank cells. Even when Google Forms marks questions as required, data may still be missing in optional questions or classification fields. Set the exclusion rule before applying it, such as removing responses that do not answer the main section of the questionnaire.

Check for straight-lining

Straight-lining occurs when a respondent selects the same level for nearly all indicators. This is a sign that requires review, not automatic proof that the questionnaire is invalid. Compare completion time, response content, and attention-check items if your questionnaire includes them.

Check the target population and unusual responses

Filter cases that fail the screening question or whose age or classification information falls outside the research scope. Check for extremely short completion times, repeated response patterns, and duplicate records. Each exclusion decision needs a reason, a count, and a date.

Standardize column names and codes

Rename columns with short codes such as PU1, PU2, IU1, and IU2 instead of keeping the full question wording. Create a codebook containing the variable name, variable label, measurement scale, and coding values. When importing the data into SPSS, indicators should use the Numeric type if you want to run Cronbach's Alpha, EFA, or regression.

Do not include columns such as timestamp, email address, or open-ended responses in the indicator group unless they belong to the model. You can retain them in the original file for checking, but consider privacy requirements before sharing the dataset.

How to present the procedure in Chapter 3

In Chapter 3, describe the questionnaire in a way that another reader can verify. State the source of the scale, how the wording was adapted, the number of indicators for each construct, the scale used, the target respondents, and how the questionnaire was distributed.

You also need to present the sample size and sampling method. If you use the Yamane formula, report the values of N and e used. If you base the sample on the number of indicators or independent variables, state the relevant rule and source. This section should match the data file because the committee may ask why Chapter 3 reports 250 questionnaires while the file contains only 218 rows.

How to write this in your thesis: “The questionnaire consisted of [number of sections] sections. Section [section name] measured [construct name] using [number of indicators] indicators, coded from [first code] to [last code]. The indicators were measured using a [5/7]-point Likert scale, ranging from [lowest level] to [highest level]. Data were collected online from [start date] to [end date]. A total of [number] responses were received, of which [number] valid responses were used for analysis.”

After data collection, Chapter 4 presents the cleaning results, Cronbach's Alpha, EFA, regression, or SmartPLS results. Do not place results that have not yet been run in Chapter 3, and do not write the planned sample size as though it were the actual number of valid responses.

Common errors

Creating the form before finalizing the model. You will have to rename variables, add or remove items, and explain why the questionnaire does not match the hypotheses. Finalize the variable diagram and codebook first.

Using self-written items while calling them a validated scale. Reference items need to be adapted to the context, translated appropriately, and tested using your own data. No form automatically guarantees reliability.

Allowing multiple answers for a variable that requires one answer. This usually happens when Checkbox is used instead of Multiple choice. When you find the error, determine whether the data can be recoded or whether the question must be excluded from the analysis.

Having no screening criteria. You may receive many responses without knowing how many respondents actually belong to the research population. Screening questions should appear before the indicator groups.

Deleting data without keeping a log. When your supervisor asks why 300 responses became 247 valid questionnaires, you need a record showing how many were removed for each reason. Do not send only a deleted file without the original copy.

Importing full question wording into SPSS. Long variable names make the output difficult to read and easy to misinterpret. Use short variable codes, and place the full wording in Variable Label or the codebook.

Frequently asked questions

Do I need a Google account to create a Google Form survey?

Google Forms usually requires you to sign in to a Google account to create and manage a form. Respondents may not need an account if you set access permissions to allow anyone with the link to submit a response. Test the actual settings in another browser before distributing the form.

Is creating a Google Form survey free?

You can create a basic form and export data to Google Sheets without paying a separate fee for each form. Some limits depend on the account and how you manage the data. For a thesis, the more important issue is whether the form reaches the right population and exports a clean file.

Should I use Google Forms or another survey tool?

Google Forms is suitable when you need to create a form quickly, already use Google Sheets, and want to share a simple link. DoThesis's fillform.info is our own tool for Vietnamese survey forms, and it is suitable when you want to manage response collection through a more centralized process.

Choose a tool that lets you control access, save data copies, and export a clear codebook. Changing tools halfway through the collection period usually creates additional risks involving questionnaire versions.

What survey sample size is enough?

There is no single correct number for every study. Consider the number of indicators, the number of independent variables, the population size, and the analysis method. You can refer to the rule of 5 to 10 observations per indicator (Hair et al., 2010), the 50 + 8m regression rule (Tabachnick and Fidell, 2013), or Yamane's finite-population formula (Yamane, 1967).

State the calculation clearly in Chapter 3 and collect additional responses to cover invalid questionnaires. Do not simply copy a number from another thesis when its topic and model differ from yours.

What should I do if Google Forms stops accepting responses or limits responses?

Check the response switch, access permissions, the one-response limit, and the account's storage capacity. Open the link in an incognito window to simulate the participant experience. If the form has stopped accepting responses, turn responses back on before sending a new link.

If people responded during the technical problem, save the existing data and record when the incident occurred. Do not create additional data to make up the difference because survey data must come from real responses.

Should every question in the form be required?

The main indicators should usually be required to reduce missing data, but consider sensitive questions and classification questions carefully. If every question is required, respondents may leave the form when they do not want to provide a particular piece of information.

Test the form on a phone, review the section branching, and check the exported data before official distribution. Then open the data file, prepare the codebook, and save the original immediately after collection ends. The next step, running the analysis on the real .sav or .csv file, can be completed in DoThesis M4 analysis.