Examples of Quantitative Questions: How to Write and Use Them in a Questionnaire
What are quantitative question examples?
Quantitative question examples are questions designed so respondents can select a value, an option, or a level that can be coded as a number. After collection, you enter these responses into an .sav or .csv file for descriptive statistics, scale testing, EFA, regression, or a PLS-SEM model.
You can picture a quantitative question like this: “I find this banking application easy to use.” The respondent chooses a value from 1 to 5, where 1 means “Strongly disagree” and 5 means “Strongly agree.” The response is not limited to the words agree or disagree. It becomes a numerical value, such as 4, which you can compare across groups and use in your analysis.
The basic structure of a quantitative question includes the content being measured, the target respondent, the response options, and the coding rule. If you are studying satisfaction, define whether you are measuring satisfaction with the product, service, employees, or the overall experience. A question that is too broad makes it difficult for the item to represent the construct.
If you are defining your research scope, you can also review the Scientific Research topic, Scientific Research, and What is scientific research to connect your research question with the questionnaire and the data you need to collect.
Why quantitative question examples matter in quantitative research
A well-designed quantitative question turns a research concept into a measurable item. For example, “service quality” is a broad construct. You can measure it through several items, such as whether employees respond on time, whether the information provided is clear, and whether the service process is convenient.
The questions also determine whether you can test your hypotheses. For the hypothesis “Service quality has a positive effect on satisfaction,” your questionnaire needs one group of items measuring service quality and another group measuring satisfaction. Each group usually contains several items rather than one broad question.
Common formats include demographic questions, classification questions, single-choice questions, multiple-choice questions, quantity questions, and questions using a Likert scale. Questions about gender or age group are usually used to describe the sample. Likert questions are often used to measure attitudes, perceptions, intentions, or the level of agreement with a statement.
You need to distinguish a research question from a questionnaire item. A research question might be “Which factors affect online purchase intention?” A questionnaire item is a specific statement such as “I intend to continue shopping on this platform.” One research question may require several items to be tested with quantitative data.
Quantitative methods are suitable when you need to measure a sufficiently large number of respondents, compare groups, or test relationships between variables. If you want to understand the context of a concept in greater depth, you can read what qualitative research is to see how the two approaches differ. In a quantitative thesis, however, the questionnaire, coding, and statistical tests still need to remain consistent with the research model.
How many quantitative question examples are enough?
There is no single number that applies to every question. A good question must first measure the intended construct, use clear wording, ask about one idea, and provide suitable response options. After data collection, you can then assess the reliability and validity of the group of items.
The table below lists benchmarks commonly used when processing quantitative data. Each row includes a source, so you do not have to copy an unexplained threshold into your thesis.
| Check | Reference point | Source |
|---|---|---|
| Cronbach's Alpha for the scale | 0.7 or above | (Nunnally, 1978) |
| Cronbach's Alpha in exploratory research | May be 0.6 or above | (Hair et al., 2010) |
| Corrected Item-Total Correlation | 0.3 or above | (Nunnally and Bernstein, 1994) |
| KMO when running EFA | 0.5 or above | (Kaiser, 1974) |
| Bartlett's Test | p-value below 0.05 | (Kaiser, 1974) |
| Factor loading in EFA | 0.5 or above | (Hair et al., 2010) |
| Total Variance Explained | 50% or above | (Hair et al., 2010) |
| Observations per item | Approximately 5 to 10 observations | (Hair et al., 2010) |
These benchmarks apply at different stages. Cronbach's Alpha and Corrected Item-Total Correlation assess the internal consistency of a group of items. KMO, Bartlett's Test, eigenvalue, factor loading, and Total Variance Explained belong to the EFA stage. You should not use an EFA benchmark to conclude that an individual question is acceptable.
Before the official data collection, read through the questionnaire with a few people whose characteristics are close to those of your research sample. The purpose is to identify unclear wording, overlapping questions, missing response options, and an excessive completion time. This is an instrument review step. It does not replace testing with official data.
How to read quantitative question examples in SPSS output
Suppose you are studying “intention to continue using an e-wallet.” The items are coded INT1, INT2, and INT3. Respondents use a 5-point Likert scale. After entering the data, run Analyze > Scale > Reliability Analysis in SPSS, move INT1 through INT3 into the Items box, select Statistics, check Scale if item deleted, and then review the Reliability Statistics and Item-Total Statistics tables.
The following is illustrative output, not the result of an actual study:
| Reliability Statistics | Cronbach's Alpha | N of Items |
|---|---|---|
| Intention to continue using the service scale | 0.842 | 3 |
| Item-Total Statistics | Scale Mean if Item Deleted | Corrected Item-Total Correlation | Cronbach's Alpha if Item Deleted |
|---|---|---|---|
| INT1 | 8.21 | 0.681 | 0.794 |
| INT2 | 8.05 | 0.726 | 0.751 |
| INT3 | 8.14 | 0.694 | 0.781 |
In the Reliability Statistics table, N of Items is the number of items included in that run. Cronbach's Alpha is the Alpha coefficient for the entire group. In this illustrative output, Alpha is 0.842, which is above the 0.7 benchmark commonly cited from (Nunnally, 1978).
In the Item-Total Statistics table, look first at the Corrected Item-Total Correlation column. All three illustrative items are above 0.3, which is consistent with the benchmark from (Nunnally and Bernstein, 1994). The Cronbach's Alpha if Item Deleted column shows how Alpha would change if you removed each item. You should not delete an item simply because the value in this column is slightly higher. Review the item wording, the theoretical basis, and your supervisor's feedback.
After Cronbach's Alpha, if your model includes EFA, go to Analyze > Dimension Reduction > Factor. In the dialog box, move the items to be tested into Variables, select Descriptives and check KMO and Bartlett's test, select Extraction to review eigenvalue and Total Variance Explained, and then select Rotation if your model requires a rotated matrix. Read the KMO and Bartlett's Test, Total Variance Explained, and Rotated Component Matrix tables according to the purpose of each table.
What to do when a quantitative question does not meet the benchmark
If a group of items has a low Alpha, do not immediately delete several items at once. Work through the checks in order so that you can still explain each decision in your thesis.
Check data entry and coding errors
Open Variable View and check Type, Values, Missing, and the coding direction. If a reverse-worded item has not been reverse-coded, its correlation with the other items may be low. For a Likert scale from 1 to 5, the usual reverse-coding is 1 to 5, 2 to 4, 3 remains 3, 4 becomes 2, and 5 becomes 1. Record which item was reverse-coded and why.
Reread the item wording
An item may perform poorly because it asks about two ideas at once, uses terminology respondents do not understand, or does not match the construct being measured. For example, “The application has an attractive interface and processes transactions quickly” contains two separate ideas. A respondent may rate the interface highly but the speed poorly, which makes the response difficult to interpret.
Review Corrected Item-Total Correlation and Alpha if Item Deleted
If Corrected Item-Total Correlation is below 0.3, you may consider removing the item according to (Nunnally and Bernstein, 1994). The decision still needs to be consistent with the content of the scale. After removing an item, run the analysis again and save the results from each round. Repeatedly deleting items only to increase Alpha can cause the scale to lose important content and may create difficult questions at the defence.
Check EFA when items split across factors
If items measuring the same construct fall into several factors, review the Rotated Component Matrix to identify cross-loading items or items with a low factor loading. EFA output is rarely tidy the first time you run it. You can remove an item when you have a clear basis and then run the analysis again, but keep the output and the reason for every adjustment.
If the data cannot be resolved through coding corrections or the removal of an item with a defensible basis, the problem may be in the questionnaire design, the respondent sample, or the suitability of the construct. Software cannot turn a question that measures the wrong construct into a good question.
Distinguishing quantitative question examples from qualitative questions
Quantitative questions usually have response options standardized in advance, and the results can be converted into numbers. Qualitative questions usually allow respondents to express themselves in free-form text to provide opinions or explanations. The two types of questions serve different purposes.
A quantitative example is “How satisfied are you with the service?” with choices from 1 to 5. A qualitative example is “What suggestions do you have for improving the service?” The first question produces data suitable for calculating a mean, standard deviation, or test statistic. The second produces text data that cannot be entered directly into Cronbach's Alpha.
You also need to distinguish quantitative questions from closed-ended questions. A closed-ended question has predefined response options, but not every such option is a quantitative measurement scale. The question “What year of study are you in?” may offer choices 1, 2, 3, and 4, but this is an ordinal categorical variable. The question “How many times have you purchased the product in the past month?” produces a number of purchases and is a quantity variable.
When designing your questionnaire, identify the construct measured by each question, the type of variable, and the analysis method you expect to use. You can also review research fields and research subjects to avoid writing broad questions that are not connected to a specific respondent group.
Common mistakes
The first mistake is writing questions around the research objective without connecting them to the variables in the model. A study on repurchase intention needs a group of items measuring repurchase intention, together with explanatory variables such as perceived quality, perceived value, or satisfaction if those constructs are included in the model.
The second mistake is combining two ideas into one item. Words such as “and,” “at the same time,” and “as well as” are often signs that you should reread the question. If the two ideas could receive different ratings, separate them into two items.
The third mistake is reversing the scale direction without recording it. Respondents may struggle when a series of statements treats agreement as positive but one item treats agreement as negative. If you must use a reverse-worded item, check the coding before running the analysis.
The fourth mistake is using too many demographic questions while leaving too few items for the main constructs. Information about gender, age, and occupation helps describe the sample or support group analysis. It does not replace the scale for an independent variable, dependent variable, mediator, or moderator.
The final mistake is revising the questionnaire after collecting data without updating the variable codes and the methods chapter. Record every change in a tracking table with the variable code, old wording, new wording, reason, and date of the change.
Frequently asked questions
Do quantitative question examples always use a Likert scale?
No. A Likert scale is suitable for attitudes, perceptions, levels of agreement, or intentions. Questions about age, number of purchases, income group, area, and usage status are also quantitative questions when their options are coded and used for statistical analysis.
How many quantitative questions are enough for one variable?
The number of items depends on the construct, the original scale, and the model design. You need enough items to represent the content of the construct rather than adding questions simply to reach a target number. When determining sample size, the benchmark of 5 to 10 observations per item can be consulted from (Hair et al., 2010), but the final sample size also depends on your analysis method and model.
Does a low Alpha mean that I should delete the quantitative question immediately?
Not yet. Check the coding, reverse-worded items, missing data, and question wording first. Then review Corrected Item-Total Correlation, Alpha if Item Deleted, and the theoretical basis. An item should be removed for a methodological and content-based reason, not simply because you want a higher Alpha.
Can one question measure one construct?
It can work for some simple or categorical variables, but latent constructs such as satisfaction, perceived quality, and intention usually require several items. If you want to run Cronbach's Alpha or EFA, a group containing only one item does not provide enough information for those tests.
What is the difference between a quantitative question and a research question?
A research question guides the entire study, such as “Does convenience affect usage intention?” A questionnaire item is a specific statement used to measure “convenience” and “usage intention.” One research question is usually measured through several items in the questionnaire.
How to write this in your thesis: [Insert the research question], [insert the independent variable], [insert the dependent variable], and [insert any mediator or moderator]. Open your research model again, make a list of the independent variables, dependent variables, and any mediating or moderating variables, then match each item in the questionnaire file to one variable. If you need to build the model and complete the questionnaire and scales before running the data, you can use M3 model, hypotheses, scales, and questionnaire for this step on your own topic.