Convenience vs voluntary response sample

By Jude Wallis · Published August 11, 2026

The difference is who does the choosing. In a convenience sample the researcher takes whoever is easiest to reach. In a voluntary response sample the subjects put themselves in by answering an open invitation. Both are biased, and each one leans for a different reason.

AP Statistics: Unit 1 (topics 1.12 Potential Problems with Sampling). Neither method uses a chance mechanism, so both are diagnosed under topic 1.12 (Potential Problems with Sampling), which follows topic 1.11 (Random Sampling), where the Fall 2026 course puts the methods that do. Topic 1.12 sits in Unit 1, which is 20% to 30% of the multiple-choice section.

Convenience vs voluntary response: the short answer

Both are nonrandom ways to end up with a sample, and one question tells them apart: who assembled it?

In a convenience sample the researcher assembles it, and the rule is ease of access. The first 40 students in the cafeteria, shoppers at one mall entrance, the teacher's own class. The subjects did nothing; they were simply within reach.

In a voluntary response sample the subjects assemble it. The researcher opens a channel, a link, a call-in line, a mail-back form, a comment card, and takes whoever chooses to use it. Nobody was selected.

Both are forms of selection bias, the family where the lean enters while the sample is being put together rather than when people answer. So the label alone is not what makes the pair worth separating. They are worth separating because they lean for different reasons, and that changes how you work out the direction of the error, which is the part free-response questions actually score.

What a convenience sample is

A convenience sample includes whichever individuals are easiest to reach, with no chance mechanism deciding who is in. The researcher stands somewhere, or asks a group already in front of them, and stops when the sample is big enough.

That shortcut always encodes something. People at a gym at 6 a.m., people at a farmers market on Saturday, people in one AP Statistics class: each group shares whatever put them there, and everyone outside that place and time had no chance at all. Nothing in the procedure was chosen with an agenda, which is why it feels harmless, and not choosing on purpose is not the same as choosing by chance. The convenience sample entry works through why that intuition fails and what it does to a confidence interval.

Because everyone outside the researcher's reach had zero chance of selection, the damage behaves like undercoverage: a whole slice of the population is missing from the start. On the exam, name convenience sampling as the method, then describe who was shut out.

What a voluntary response sample is

A voluntary response sample is made up of people who chose to respond to an open invitation. Nobody was drawn from a list, so nobody was selected and then lost.

That matters for one specific reason. Self-selection is not neutral: taking the trouble to answer an unsolicited invitation is itself evidence of feeling something about the topic. The people who feel nothing in particular do not bother. So the sample fills up at the ends of the opinion scale and thins out in the middle whenever the only thing on offer is the chance to be heard.

The silent majority in an open poll is not nonresponse. A response rate is answers divided by people selected, and here nobody was selected, so there is no rate to quote and no list to chase; the voluntary response sample entry has the full argument and why it decides the repair. Compare nonresponse bias, where a real list exists and follow-up can still pull the estimate back toward the truth.

The differences side by side

Feature	Convenience sample	Voluntary response sample
Who assembles the sample	The researcher	The subjects
The selection rule	Ease of access	Willingness to answer an invitation
Who is shut out	Anyone not at that place and time	Nobody formally, but the indifferent stay away
What predicts the lean	Whatever trait made those people reachable	Strength of feeling about the topic
Direction of the lean	No fixed sign; argue it from the shortcut	Toward strong views, often the unhappy side
Is there a response rate	No, there was no frame to draw from	No, nobody was selected
Cured by a bigger sample	No	No

The first row drives every other one. When the researcher picks, the sample inherits the researcher's shortcut. When the subjects pick, the sample inherits their motivation.

Which way each one pushes the estimate

This is the useful half of the distinction, and it is where the two stop behaving alike.

Convenience bias has no sign of its own. It follows whatever made those people easy to reach, so you have to argue the direction from the specific shortcut and cannot read it off the label. A teacher estimating mean nightly sleep for all 1,200 students by surveying her own two AP Statistics classes will come out too low, because those students carry heavier course loads than the school as a whole. Hand the identical survey to a first-period study hall instead and the estimate can move the other way. Same method, opposite error. The worked example below moves one interviewer across town and flips the sign by construction.

Voluntary response bias does have a predictable lean, because its selection rule never changes: it selects on motivation. A district posts an open link asking parents to rate a new bus schedule. Parents whose child now waits forty minutes in the dark write in. Parents whose morning got easier mostly do not, so reported satisfaction comes out too low.

Be careful how far you push that. When the only reward for answering is being heard, you can assert that strong feeling is over-represented and the indifferent middle is missing. Which end wins depends on which side is motivated, and for a change, a cost, or a complaint channel that is usually the aggrieved side. When a proposal has enthusiastic backers and quiet opponents, an open poll overstates support instead. So say who is motivated here, then give the direction; what is selection bias sets out the three-part sentence graders look for.

The classic mix-up and how to avoid it

Read the scenario for one thing: which side moved first. If the researcher went to the subjects, it is convenience. If the researcher put something out and waited, it is voluntary response.

Wording in the problem	Which one
"interviewed the first 25 people leaving the gym"	Convenience
"handed the survey to her own class"	Convenience
"listeners were invited to call in"	Voluntary response
"a link was posted on the school website"	Voluntary response
"comment cards were left by the exit"	Voluntary response

Three errors cost the most points. First, calling the people who ignored an open link nonresponse; they were never selected. Second, treating a large count as reassurance: an open poll with 100,000 answers is still self-assembled, and a bigger convenience sample is a more precise version of the same wrong number, worked through in does a bigger sample fix bias. Third, stopping at "the sample is biased" with no direction attached, which is the habit how to identify the type of bias is built to break.

One thing neither flaw wrecks is an experiment. Randomly assigning treatments to a group of volunteers, or to whoever was convenient, still supports a cause-and-effect conclusion about those subjects. What it cannot do is carry the result to a wider population, since that right comes from random selection; see can you generalize results. And if the plan is still being written rather than diagnosed, how to choose a sampling method is where the four random methods get compared.

One town, two flawed methods, two different errors

A town has 4,000 registered voters and is asked to approve a bond to rebuild the public pool. Of the 1,200 voters who live in the neighborhood around the pool, 70% support the bond; of the other 2,800 voters, 35% support it. (a) Find the true proportion of supporters. (b) An intern interviews the first 100 people who walk past the pool on Saturday morning, where 90% of passers-by are pool-neighborhood voters. What proportion of support is this method aimed at? (c) Suppose the intern had instead stood outside a supermarket across town, where only 5% of passers-by are pool-neighborhood voters. Now what? (d) The town also posts an open online form. Supporters answer it at a rate of 5% and non-supporters at 20%. What does the form report?

(a) Count supporters in each group. Pool neighborhood: $1200 \times 0.70 = 840$ . Everyone else: $2800 \times 0.35 = 980$ .
(a) Add them and divide by the population: $840 + 980 = 1820$ , so $p = \frac{1820}{4000} = 0.455$ . Just under half the town supports the bond.
(b) This is a convenience sample: the intern took whoever was within reach at one place and one time. The pool-side crowd is 90% pool-neighborhood voters against 30% in the town, since $\frac{1200}{4000} = 0.30$ .
(b) The method is aimed at $0.90 \times 0.70 + 0.10 \times 0.35 = 0.63 + 0.035 = 0.665$ , which overstates support by $0.665 - 0.455 = 0.21$ , or 21 percentage points.
(c) Same method, different shortcut: $0.05 \times 0.70 + 0.95 \times 0.35 = 0.035 + 0.3325 = 0.3675$ . Now it understates support by $0.455 - 0.3675 = 0.0875$ , about 8.75 percentage points.
(c) Note what changed. Nothing about the procedure, only where the intern stood. Convenience bias takes its sign from the shortcut, so the direction has to be argued from the place and time.
(d) The form is a voluntary response sample. Non-supporters number $4000 - 1820 = 2180$ , so $2180 \times 0.20 = 436$ of them answer, against $1820 \times 0.05 = 91$ supporters.
(d) Total answers: $436 + 91 = 527$ . The reported statistic is $\hat{p} = \frac{91}{527} \approx 0.1727$ (p-hat), which understates support by $0.455 - 0.1727 = 0.2823$ , about 28 percentage points.
(d) The 527 answers are $\frac{527}{4000} \approx 0.132$ of the town, and that 13.2% is not a response rate. Nobody was selected, so the 3,473 silent voters are not nonresponse and there is no list to follow up.
Compare the two failures. The convenience estimate moved 21 points up or 8.75 points down depending only on the location. The online form leaned one way for a reason built into it: opposing a bond is more motivating than supporting one, so opponents answered at four times the rate.

(a) $p = 0.455$ . (b) The pool-side convenience sample is aimed at 0.665, overstating support by 21 percentage points. (c) The same method run across town is aimed at 0.3675, understating by 8.75 points, so convenience bias has no fixed direction. (d) The open form reports $\hat{p} \approx 0.173$ from 527 answers, understating support by about 28 points, because non-supporters answered at 20% against 5%.

Name the method and the direction in four scenarios

For each study, say whether the sample is a convenience sample or a voluntary response sample, and say which way the estimate is pushed, or why the direction cannot be pinned down. (a) A researcher estimating the mean daily step count of adults in a city hands step trackers to the 28 members of her running club. (b) A restaurant estimating what share of its diners found the service quick prints a survey link on every receipt; 210 diners scan it. (c) A newspaper estimating what proportion of the city's adults still read a print paper interviews people in the public library on a Wednesday afternoon. (d) A museum estimating how its 9,000 monthly visitors rate a new exhibit leaves comment cards by the exit and gets 300 back.

(a) The researcher chose a group she could reach, so this is a convenience sample. Running club members walk and run far more than city adults do, so the estimate of mean daily steps comes out too high.
(b) Diners decided for themselves whether to scan, so this is a voluntary response sample. Name who is motivated: a slow meal leaves a diner with a complaint to file and a quick one usually does not, so the share reporting quick service comes out too low.
(c) The reporter went to a place and took whoever was there, so this is a convenience sample. People spending a weekday afternoon in a library read print far more than the city as a whole, so the estimate of print readership comes out too high.
(d) Visitors chose to fill in a card, so this is a voluntary response sample. The 300 cards are $\frac{300}{9000} \approx 0.033$ , about 3.3% of visitors, and again that is not a response rate because nobody was selected.
(d) Say only what the mechanism supports. Delighted and annoyed visitors both write; indifferent ones walk out. So the cards reliably overstate how strongly the 9,000 visitors feel, and the exhibit's average score is pushed toward whichever of those two groups is larger, which the museum cannot tell from the cards alone.
Check the pattern across all four. In (a) and (c) the direction came from the trait that made the group reachable, and in (b) and (d) it came from what motivates a person to answer.

(a) Convenience sample, overstating mean daily steps. (b) Voluntary response sample, understating the share who found service quick. (c) Convenience sample, overstating print readership. (d) Voluntary response sample: it overstates how strongly visitors feel, and the score leans toward whichever strong group is bigger, which 300 self-selected cards cannot reveal.

Frequently asked questions

Which is worse, a convenience sample or a voluntary response sample?

Neither is safe, and the honest answer is that they fail differently. A convenience sample can be badly wrong in either direction depending on where the researcher stood, so its error is unpredictable. A voluntary response sample leans toward whoever feels strongly, so its error is more predictable but often larger, because motivation is tied directly to the opinion being measured.

Does voluntary response bias always overstate the negative side?

No. What it does instead is over-represent strong feeling and thin out the indifferent middle. The negative side usually wins when the topic is a change, a cost, or a complaint channel, since annoyance is what drives people to answer. A poll of an enthusiastic fan base tilts the other way. Name who is motivated in this scenario, then give the direction.

Are the people who ignored an online poll a case of nonresponse bias?

No. Nonresponse means a person was selected for the sample and then failed to answer, so a response rate and a follow-up list both exist. In an open online poll nobody was selected, so there is no rate to compute and nobody to chase. That is why voluntary response bias cannot be repaired after the fact, while nonresponse sometimes can.

Can a convenience sample still be used for an experiment?

Yes, with one limit. Randomly assigning treatments to a conveniently recruited group still supports a cause-and-effect conclusion about those subjects, because the random assignment happens inside the study. What it cannot support is generalizing to a wider population, since that comes from random selection, which never happened.

How do I state the direction of convenience bias for full credit?

Say what made that group easy to reach, say how the group differs from the population on the variable being measured, then say too high or too low in context. For example: people leaving a gym exercise far more than the town does, so the estimated mean weekly exercise is too high. The label on its own earns nothing.