Written and maintained by CASRAI Editorial Board
Last updated
A quantitative survey question is a closed-ended item whose answer is a number or a choice that is coded as one, so responses can be counted, averaged and compared. This guide gives quantitative survey question examples taken word for word from a fielded federal instrument, the CDC’s 2024 Behavioral Risk Factor Surveillance System (BRFSS) questionnaire, then shows how the response codes must be handled before any statistic is computed, with a worked calculation.
What is a quantitative survey question?
A quantitative survey question asks for a count, an amount, a frequency or a position on a fixed scale, and the answer is stored as a number. That is what separates quantitative survey questions from open-ended ones, whose free-text answers need coding before they can be counted (see open-ended vs closed-ended questions). Quantitative research survey questions usually sit inside a research questionnaire and feed the kinds of variables described in quantitative data examples.
Not every closed question gives the same kind of number. A count of days is a numeric measurement; a “poor to excellent” rating is an ordered category. The levels of measurement decide which statistics are valid, and the ordered-category items have their own page: examples of ordinal survey questions. This page concentrates on numeric-entry and fixed-format items.
Quantitative survey question examples from the CDC BRFSS
The items below are quoted from the 2024 BRFSS Questionnaire (CDC), with wording and codes last checked on 2026-10-29. They are fielded items, not teaching inventions.
Numeric entry within a bounded range (Healthy Days)
Item CHD.01 asks: “Now thinking about your physical health, which includes physical illness and injury, for how many days during the past 30 days was your physical health not good?” The response is a two-digit number of days from 01 to 30, plus special codes: 88 None, 77 Don’t know/not sure, 99 Refused. The interviewer is told that 88 may be coded if the respondent says “never” or “none”. Item CHD.02 repeats the stem for mental health, “which includes stress, depression, and problems with emotions”, with the same response format.
Frequency with a choice of unit (alcohol)
The questionnaire first defines the unit inside the stem: “One drink of alcohol is equivalent to a 12-ounce beer, a 5-ounce glass of wine, or a drink with one shot of liquor.” It then asks: “During the past 30 days, how many days per week or per month did you have at least one drink of any alcoholic beverage?” The answer is entered as days per week (code prefix 1) or days in the past 30 days (code prefix 2), with 888 for No drinks in past 30 days, 777 Don’t know and 999 Refused. A follow-up asks, “on the days when you drank, about how many drinks did you drink on the average?”, with a numeric count and the codes 88 None, 77 and 99.
What these real items have in common
| Design feature | Where it appears in the BRFSS items |
|---|---|
| Fixed reference period | “during the past 30 days” in CHD.01, CHD.02 and the alcohol item |
| Bounded numeric range | Number of days entered as 01 to 30 |
| Definition inside the question | “Physical health” and “mental health” defined in the stem; one drink defined as a beer, glass of wine or shot |
| Separate non-substantive codes | 88 None, 77 Don’t know/not sure, 99 Refused (888, 777, 999 on the three-digit item) |
Examples of quantitative research survey questions by response format
The following four patterns cover most quantitative survey questions. The wording below is illustrative and written for this page; it is not quoted from any instrument.
- Numeric entry: “In the past 7 days, on how many days did you exercise for at least 30 minutes?” (answer 0 to 7).
- Frequency categories: “How often did you attend a lab meeting in the past month? Never / 1-2 times / 3-4 times / 5 or more times.” Categories must not overlap and must cover every possible answer.
- Yes/no: “Did you submit a grant application in the past 12 months?” Coded 1/0, a yes/no item gives a proportion when averaged.
- Rating scale: “How satisfied are you with your institution’s research office? 1 (very dissatisfied) to 5 (very satisfied).” Rating items are ordinal in strict terms; see the Likert scale guide for how they are treated in analysis.
Design rules for closed-ended survey questions
Pew Research Center’s “Writing Survey Questions” sets out rules that apply to every item above. Closed-ended questions “should include all reasonable responses”, and the response categories “should not overlap”. Pew recommends keeping the number of answer choices small, “just four or perhaps five at most”, especially in telephone surveys. Order matters too: respondents more often choose items heard later in a list on the phone, and tend to choose items at the top of a list when self-administering. Ordinal scales are generally not randomized because “the order of the categories conveys important information.”
A double-barreled question asks respondents to evaluate more than one concept; the fix is to split it into two items. Pew also reports a 2008 test in which 58% chose the economy when it was offered as a closed option, compared with 35% when the question was open-ended, a reminder that the format changes the answer. The trade-offs are covered in advantages and disadvantages of surveys.
Worked example: coding special values before computing a mean
Suppose ten respondents answer CHD.01. This data set is made up to show the arithmetic. Raw codes, in order: 5, 88, 12, 30, 77, 3, 88, 99, 10, 2.
Step 1, recode. 88 (None) means zero days, so it becomes 0. 77 (Don’t know) and 99 (Refused) are not day counts, so they become missing. The valid values are 5, 0, 12, 30, 3, 0, 10, 2 (n = 8).
Step 2, mean. Sum = 5 + 0 + 12 + 30 + 3 + 0 + 10 + 2 = 62. Mean = 62 / 8 = 7.75 days.
Step 3, standard deviation. Squared deviations from 7.75 are 7.5625, 60.0625, 18.0625, 495.0625, 22.5625, 60.0625, 5.0625 and 33.0625, which total 701.5. Sample variance = 701.5 / 7 = 100.21; SD = 10.01.
Step 4, median. Sorted: 0, 0, 2, 3, 5, 10, 12, 30. The middle two values are 3 and 5, so the median is 4 days.
If the codes had been averaged as typed, the mean would be 414 / 10 = 41.4 “days”, impossible for a 30-day window. Dropping the 88s as missing instead of recoding them to 0 gives 62 / 6 = 10.33, which overstates the average because the people with no unhealthy days were removed.
Interpretation
| Result in this example | Reading |
|---|---|
| Mean 7.75 days, median 4 days | The mean is higher than the median because one respondent reported 30 days; the distribution is right-skewed, so report both. |
| SD 10.01 days | Larger than the mean, which signals wide spread rather than a typical value near 7.75. |
| Two of ten answers missing | Report n = 8 of 10 valid responses, so readers can judge how much the mean depends on the recode. |
| Two respondents at 0 days | Zero is a real value here, not missing; a ratio-level count where 0 means none. |
Assumptions and common mistakes
- Treating special codes as data. 77, 88 and 99 look like numbers; left in, they inflate every statistic.
- Treating “None” as missing. It is a valid zero in a count item.
- Overlapping or incomplete categories. “1-3” and “3-5” overlap; Pew’s rule is that categories should not overlap and should include all reasonable responses.
- Double-barreled wording. “How satisfied are you with the pay and the working conditions?” cannot be answered with one number; split it.
- Averaging ordinal ratings without saying so. A mean of a 1-5 rating treats the steps as equal, which the scale does not guarantee.
How to report it (APA example)
Report the valid n, the mean and the standard deviation, and say how special codes were handled. For the worked example: “Respondents reported a mean of 7.75 days of poor physical health in the past 30 days (SD = 10.01, Mdn = 4; n = 8 valid responses; two ‘Don’t know’ or ‘Refused’ responses were treated as missing).” For how to cite the survey itself, see how to cite a survey in APA.
How to calculate it in Excel, R or Python
Excel
Put the raw codes in A2:A11. In B2 enter the formula below and fill down, then use the two summary formulas.
B2: =IF(OR(A2=77,A2=99),"",IF(A2=88,0,A2)) Mean: =AVERAGE(B2:B11) SD: =STDEV.S(B2:B11)
R
x <- c(5, 88, 12, 30, 77, 3, 88, 99, 10, 2) x[x %in% c(77, 99)] <- NA x[x == 88] <- 0 mean(x, na.rm = TRUE); sd(x, na.rm = TRUE); median(x, na.rm = TRUE)
Python (pandas)
import pandas as pd
s = pd.Series([5, 88, 12, 30, 77, 3, 88, 99, 10, 2])
s = s.replace({88: 0, 77: float("nan"), 99: float("nan")})
print(s.mean(), s.std(), s.median())
All three return a mean of 7.75, a standard deviation of 10.01 and a median of 4.
Frequently asked questions
What are quantitative survey questions?
They are closed-ended questions whose answers are numbers or are coded as numbers, such as a count of days, a frequency category, a yes/no code or a rating, so responses can be summarized statistically.
Are yes/no questions quantitative?
Yes, when coded 1 and 0. The mean of the codes is the proportion answering yes. The variable is binary, a categorical type, so it supports counts and proportions rather than a meaningful mean of amounts.
Is a Likert item quantitative?
It produces numeric codes, but strictly the levels are ordered categories; whether to analyze them as numbers is an analysis decision that should be stated. See the Likert scale guide.
How many answer options should a closed question have?
Pew Research Center advises keeping answer choices to a small number, just four or perhaps five at most (especially in telephone surveys), with categories that do not overlap and cover all reasonable responses.
Sources
- Centers for Disease Control and Prevention. 2024 BRFSS Questionnaire. cdc.gov/brfss/questionnaires/pdf-ques/2024-BRFSS-Questionnaire-508.pdf (items CHD.01, CHD.02 and the alcohol items; checked 2026-10-29).
- Pew Research Center. Writing Survey Questions. pewresearch.org/writing-survey-questions (checked 2026-10-29).
