How to get a representative sample for your survey

When you’re conducting a survey, it is often all-but-impossible to capture feedback from every person in the group or audience you're targeting. Even if you manage to get a survey in front of all of them, you'll undoubtedly have some people who do not respond or aren’t interested in participating.

The good news? You don’t have to survey everyone in your target survey group to get extremely useful and insightful data from your survey. In fact, you can needlessly go overboard when you're surveying large groups of people, which can cost you time and money and add complexity to the process.

Through representative sampling, you can then analyze the results to deliver actionable data and useful insights to support your market research and customer experience efforts.

A representative sample is a subset of data, usually from a larger group, that can portray similar characteristics. Representative sampling helps you analyze larger populations because the data generated contains smaller, more manageable versions of the larger group. We see representative sampling in action during elections, when pollsters survey representative samples of voters to learn which candidates are getting support.

Representative sampling can save you considerable time and money. Through statistical analysis and review of the data, you can capture an accurate representation of a broader customer audience, or specific group within that audience without having to get input from everyone–or even the vast majority of those who could be surveyed.  

Yet making sure you get accurate and credible results takes a good working knowledge of the clear benefits and potential pitfalls associated with representative sampling. When armed with this knowledge, you can strategically use this tool to gain better insights into your customers, and use that information to enhance customer experience or inform better decision making regarding your business.

Imagine you had a group of 300 people–150 men and 150 women–who went through a specific training program. You want to get their feedback about the program to identify any issues and find out which elements of the program participants found most valuable. 

By relying on representative sampling as part of your survey process, you don’t need responses from the majority of those who participated in the training. Instead, from the full group of 300 participants, you could generate a credible representative sample that might include 60 people—30 men and 30 women. Those responses would be representative of the larger group.

Once you receive their responses, you can analyze the results to gain insights that represent the opinions of the full group. Because you have a representative split of men and women who completed the program, you can also assess the responses based on gender.

Instead of having to rely on 300 people to respond to your survey, you can get highly accurate results from just 60 total responses, with an even split between men and women. In turn, you save time and money.

SurveyMonkey Audience can help supply your representative sample by connecting you to your ideal respondents—even if they’re all over the world.

There are a number of established methods to get a representative sample that have been tested and verified over time through academic, scientific, and market research. 

The most common methods include: 

By using one of these methods of capturing representative samples, you'll assure credibility and accuracy of your results.

Probability sampling is a sampling technique in which you choose samples from a larger population using a method based on the theory of probability. For participants to be considered for a probability sample, they must be selected using what’s called random selection. Random selection simply means that everyone in the population you're surveying has a known and equal chance of being selected. So, if you had a population of 1,000 people every person would have odds of 1 in 1,000 of being selected.

Because those conducting the survey or research can’t cherry-pick who is selected to be part of the sample, probability sampling gives you the best chance to capture a sample that is truly representative of the population. In short, probability sampling removes the potential for human bias or sampling error and relies on statistical theory to randomly select a small group of people from the larger population and then predict that all of their responses will match that of the larger population.

It’s always great to keep things simple. And simple random sampling does just that by offering a straightforward path to getting a viable sample group.

Simple random sampling is as easy as assigning numbers to each individual in your group or sample and then randomly choosing numbers through an automated process to determine who will be included in the sample. You can choose your sample numbers via a lottery system or by using number generating software that randomly selects them for you.

This method is one of the best ways to avoid sampling bias, which can often happen with conducting concept testing. Sampling bias can sneak its way into your survey when some members of a population are systematically more likely to be selected in a sample than others. Sampling bias may get you results that look favorable or support a particular viewpoint—and could lead to sampling error, false assumptions and poor decisions.

Simple random sampling greatly minimizes the chance for sampling bias by providing equal odds for every member of the population to be chosen as a participant in the study at hand.

Non-probability sampling is not so random. Rather, non-probability sampling is a technique in which those conducting the research select samples based on subjective judgment rather than random selection. Subjective judgment isn’t driven by established formulas or statistical analysis. Instead, it relies on an individual’s expert opinion or experience to identify those respondents who will be included in the sample.

When you conduct non-probability sampling, not all members of the population have an equal chance of participating in the study. This means each member of the population has a known chance of being selected.

As you may have guessed, non-probability sampling is a less stringent method than probability sampling because it introduces human judgment–and potentially human error or sampling bias. On the positive side, the approach can be a more efficient and effective way to assure that the people who will be surveyed or included in a study are likely to provide the most useful information  Non-probability sampling is a frequent key tool for qualitative research that focuses on nonnumerical data to generate insights and conclusions.

So if it is not as rigorous or reliable, does using non-probability sampling make sense? In certain circumstances, the answer to that question is yes. Non-probability sampling can be especially useful for exploratory studies like a pilot survey (deploying a survey to a smaller sample compared to pre-determined sample size). Further, it is often used in situations in which random probability sampling would be difficult or impossible due to time limits, cost constraints, or other challenges.

Quota sampling can deliver representative results that help your organization make strong data-driven decisions when you need to know more about a specific subset of your target population and don’t have a big budget. 

This method is a nonprobability sampling method in which researchers create a sample involving individuals that represent a specific population. With quota sampling, you can make sure your survey results closely resemble your target population. That way, you’ll get results you can really use. Quota sampling carries similar risk-reward as nonprobability sampling, but it is a very effective way to capture actionable data and insights from a specific audience.

If your sample is too big, your survey and the ensuing analysis can become complex, expensive and time-consuming. In a perfect world, properly conducting a larger survey and getting a higher percentage of responses increases the accuracy and certainty of the resulting data. Yet those added benefits typically aren’t worth the risks presented by taking on a costly, time-devouring and unwieldy survey. More often, representative sampling offers a faster, easier, and accessible way to get valid and statistically significant results.

If your sample is too small, you're more likely to get results that are not statistically significant. For instance, you may end up including a disproportionate number of individuals who are outliers, or not representative of the larger group. This can skew your results and lead to flawed or incomplete data, and, ultimately, poor decisions.So, your aim should always be to find the ‘just right’ representative sample size. When your sample is the right size, it allows for time and cost efficiency while also giving you reliable representative sample statistics. SurveyMonkey makes it easy for you to determine the right sample size with its sample size calculator. By using this simple tool, you can quickly determine your representative sample.

Sampling bias can occur when some members of a population are systematically more likely to be selected in a sample than others. When you hear controversy over the validity of certain political polls, it often focuses on potential sampling bias in which the pollster ends up polling too many individuals who have a particular political leaning.

Sampling bias may get you results that look favorable or support a preferred theory, idea or effort. Yet if those results are skewed and don’t reflect reality, you can end up making poor decisions or harming your organization’s reputation—and potentially yours, as well.

The best way to avoid sampling bias is to conduct simple random sampling, in which samples are chosen strictly by chance. This provides equal odds for every member of the population to be chosen as a participant in your study. If you're going to do nonprobability sampling in which participants are selected based on judgment, be sure that those involved have experience, credibility, and awareness of the need to avoid sampling bias that would taint results and conclusions.

It’s all but impossible to avoid some element of sampling bias in surveys that rely on representative sampling. For instance, people who are extremely busy tend to participate less in surveys. As a result, surveys often don’t fully capture the perspectives from really busy people, meaning the super busy are often underrepresented.

Looking for more ways to reduce bias? SurveyMonkey’s market research survey templates include a variety of question types to help your respondents answer more truthfully. 

Representative samples are important because they ensure that as many relevant types of people as possible are included in your sample, and that the right mix of people are interviewed or surveyed. This helps ensure your results aren’t tainted by bias. It also helps guard against overrepresenting certain groups.

Representative samples provide the following benefits:

Representative sampling can be a key element of getting the most out of surveys by generating useful data and insights to support your market research, customer experience, and other business goals.

Check out SurveyMonkey’s Market Research Solutions for help when you conduct surveys using representative samples. Looking to identify a representative sample? SurveyMonkey’s Audience Calculator makes the process quick and easy.