A representative sample is a group of participants in a study who are, overall, similar to the broader population under study. Getting a good representative sample is important for ensuring study validity.
The characteristics of the sample should match the characteristics of the larger population that the researchers want to investigate.
For example, researchers who want to understand how poverty effects the social and emotional development of children will select families in lower income brackets because they are representative of families in poverty.
Types of Representative Samples
There are two types of samples: homogeneous and heterogeneous.
1. Homogenous Samples
A homogeneous sample means that the people in the sample are all the same on one or more characteristics.
For example, they may all fall within a narrow age-range or SES class.
2. Heterogeneous Samples
A heterogeneous sample is diverse. The people in the sample vary on a wide range of demographic variables such as gender, age, race, and SES class. Therefore, a heterogeneous sample is more representative of the population of a society than a homogeneous one.
However, sometimes the population is not society, but a subgroup of that society. For example, if studying the job satisfaction of IT workers, then that is the population under study. Therefore, the sample used in the study should match the characteristics of that specific population (a homogenous sample), not of society in general.
Representative Sample Examples
- Sampling Families in Poverty: Researchers want to understand how poverty effects the social and emotional development of children, so they select families in lower income brackets to study.
- Sampling People in a Region: A political campaign wants to know what message is most effective with voters in a specific region of the country, so they field-test campaign slogans only in those areas of the country.
- Sampling Age Groups: A cereal brand is developing a new breakfast cereal for kids, so they conduct taste tests using children ages 5-9 years old
- Sampling Video Gamers: A video game company has designed a new game console in varying styles, so they let experienced gamers test the equipment in their lab and then have them rate each version’s level of appeal
- Sampling Baby Boomers: Coca-Cola has a new beverage product that targets baby boomers, so they send product samples to households that fit that demographic and then conduct phone surveys later
- Sampling Genders: Educators want to encourage females to become math and science majors, so they develop lesson plans and activities specifically for teenage girls
- Sampling Consumers: Psychologists want to understand how people process information presented in online commercials, so they randomly recruit people in supermarkets across all areas of a large city
- Sampling Patients: A pharmaceutical company has developed a treatment for arthritis, so they contact doctors’ offices to identify study participants
- Sampling New Parents: A car company has restyled its minivans, so they invite new mothers and fathers to take a test-drive for a weekend
- Sampling a Diverse Population: The government wants to encourage all drivers to wear their seatbelts, so they field-test various PSAs at high schools, manufacturing plants, and corporate offices.
1. Fundamental Consumer Demographics
The basis of effective advertising and marketing is understanding consumer demographics, so getting a representative sample of consumers is extremely useful.
There are two sides to this coin. First, this means a company must have a detailed picture of their customers’ profile. On the other side of the coin, it means the company must be able to target its advertising and marketing efforts where it will be seen by those customers.
The fundamental characteristics of a population consist of many segments, including: age, race, ethnicity, gender, marital status, income, location, education, and employment.
Segmenting customers based on these variables is useful because different segments of the population often react to ads differently. They also have vastly different needs and expectations regarding products and desired price points.
Therefore, when designing a new ad campaign and marketing strategy, a company relies on many sources of data to create a very detailed customer profile. When it comes time to implementing the campaign, the company then directs their ads to a sample that is as representative of that profile as possible.
2. The WHOQOL
There is no greater challenge in the world of developing representative samples than when it serves a global purpose. This is the case with the development of the WHOQOL, which stands for the World Health Organization General Quality of Life Assessment.
This is a scale that is designed to “…assess respondents’ perception and subjective evaluation of various aspects of the quality of life” (Saxena & van Ommeren, 2005, p. 975).
Unfortunately, developing a single scale regarding such an abstract concept as “quality of life” is quite difficult. This is a concept that is very much a function of culture. Not only are there differences between Western and Eastern cultures, but within those broad categories there can also exist vastly different conceptions of what makes for a quality life.
Since it is impossible to create a representative sample of every culture on the planet, researchers had to work within the constraints of reality. Therefore, samples were selected from 15 centers around the world, including: Bangkok, Bath, Madras, Melbourne, Panama, St. Petersburg, Seattle, Tilburg, and Zagreb.
Sample questions were presented to focus groups of citizens and healthcare professionals in each of those cities. Eventually, 100 questions were selected in the final version based on the opinions of the various samples.
3. Representative Sample in the Era of Social Media
In addition to the incredible level of detailed information a company can acquire regarding basic demographics, social media offers another layer of depth that is truly astounding, maybe even a little shocking.
Companies can collect data from a seemingly endless string of sources: Twitter comments, facebook posts, and news articles clicked on, just to name a few social media examples. Companies can also identify the time of day when certain consumer demographics are online, the keywords they use when searching, and what kind of videos they like.
So, if a bank wants to target consumers who are interested in receiving a loan for a car or new home, are in a certain SES class, and live in the bank’s service area, there are many programs available that will compile a list of potential customers that match that profile.
The bank can then direct their ads towards a sample of potential customers that is highly representative of the population the company wants to target.
4. Using Random Selection to Ensure a Heterogeneous Sample
The principle of random selection means that each person in a given population has an equal chance of being selected. Having a randomly selected sample means that the participants in a study are not all similar on any one characteristic.
That is, if a certain percentage of the population is female, aged 24-36 years old, educated, and consist of three particular ethnic groups, then the people in the sample will be representative of that population.
This is extremely important when conducting research if the goal is to be able to generalize the study’s findings to the population. However, if for some reason a sample is not selected randomly, then the sample may be skewed on at least one characteristic. Maybe all the participants are in a very narrow age range or only represent one ethnic group.
For a practical example, lets consider a study on the memory capacity of children in the third grade. Because of practical limitations, the researchers go to a nearby primary school and select 50 names at random from the school’s register. There is software available to do this efficiently.
However, because it is only one school, in one district, in one city, in one region of the country, the sample is not truly representative of the population of all third graders.
The results of the study may not be applicable to students in other regions of the country, or in other SES classes. If the study was conducted in the heart of a large metropolis, the results may not apply to students in rural areas.
Random selection helps ensure that the sample is heterogeneous, but to do so, the procedure needs to be applied wide enough.
Another way to address this problem is to use stratified sampling, which ensures important sub-groups are proportionately represented in a study.
5. Psychological Research in “WEIRD” Nations
According to a study by Henrich et al. (2010), between 2003 and 2007, a vast majority (96%) of samples used in psychological research came from countries with only 12 percent of the world’s population.
Since psychologists often want to understand the mental processes of “human beings,” this statistic means that the samples being used are a far cry from being representative of the population (i.e., Homo sapiens).
This is a huge blow to the generalizability of research. Moreover, the implications are much more significant given the fact that the problem has existed for far longer a period of time than just from 2003 to 2007.
As quoted in Science Daily, University of British Colombia Psychology and Economics Prof. Joe Henrich states: “The foundations of human psychology and behavior have been built almost exclusively on research conducted on subjects from WEIRD societies.”
To further make the point, the study involved comparing the results of a range of studies in a comparative database and found substantial differences in the results between studies conducted in WEIRD countries and those that involved the rest of the species. The studies included research on memory, conformity, spatial reasoning, and visual perception, just to name a few.
The implication is that results from Western, Educated, Industrialized, Rich and Democratic (WEIRD) cultures may not be representative of the human population at all.
Psychological research depends heavily on studying human behavior via a sample of research participants. In some cases, that sample should be large and varied, so that it is representative of the broader society.
In other cases, however, it is important that the sample is representative of the specific population being studied. We see this in situations in which a company wants to test a new product or marketing campaign on its typical customer profile. Therefore, the company would prefer a sample that is representative of their existing customers instead of one that is representative of the general population.
Henrich, J., Heine, S., & Norenzayan, A. (2010). The weirdest people in the world? The Behavioral and Brain Sciences, 33(2-3), 61-83. https://doi.org/10.1017/S0140525X0999152X
Machery, E. (2010). Explaining why experimental behavior varies across cultures: A missing step in “The weirdest people in the world?”. Behavioral and Brain Sciences, 33(2-3), 101-102. https://doi.org/0.1017/S0140525X10000178
Saxena, S., & van Ommeren, M. (2005). World Health Organization Instruments for Quality of Life Measurement in Health Settings. Editor(s): Kimberly Kempf-Leonard, Encyclopedia of Social Measurement, 975-980. https://doi.org/10.1016/B0-12-369398-5/00508-9
University of British Columbia. (2010, June 30). Psychological research conducted in ‘WEIRD’ nations may not apply to global populations. ScienceDaily. Retrieved September 27, 2022 from www.sciencedaily.com/releases/2010/06/100630132850.htm