Quantitative research is the investigation of an issue using numbers to establish meaning for behavior. In the market research world, numbers give insight into consumer behavior, but quantitative research can and is used to explain other kinds of performance, as well. Scientists and analysts in both the social sciences and the natural ones use numerical data to apply significance to any number of different kinds of phenomena. Here’s a quick overview of quantitative research methods:
Descriptive Research Methods
Descriptive research methods include the basic surveys, questionnaires and polls most of us associate with market research projects. They are used to describe a population, with questions asked of a random sample of people and the answers projected to a larger group in order to more easily identify meaningful patterns and trends. Case studies, ethnographies and observations are also descriptive research techniques. Rather than rely on respondents self-reporting information, researchers use these research instruments to examine existing literature and/or observe certain behaviors in real time. Regardless of the exact tool used, descriptive research methods allow researchers to report on a single population or compare two or more different ones. Questions are objective in nature and can be presented on paper, via the Internet, over the phone or face-to-face.
Correlational Research Methods
Correlational research methods are used when it would be helpful to explore whether a relationship exists between two (and sometimes more) variables (such as price point and sales numbers or advertising exposure and brand awareness). Researchers typically don’t control one variable over another; they simply observe how each exists in relation to the other variable(s). This type of quantitative research method can highlight interesting theories to explore further, but it can’t conclusively prove causation. Thus, a lot of correlational research is used as a foundation for experimental research.
Causal-Comparative Research Methods
When researchers want to assign cause and effect to a situation, they use causal-comparative research methods. With this type of investigation, researchers maintain one variable and manipulate another one. For instance, a company might wish to evaluate whether or not new packaging will affect sales. Instead of asking people if they would likely buy the product in the new packaging (descriptive method) or changing the packaging and watching the sales reports (correlational method), the company might set up a quasi-experiment, presenting the product with the new packaging in one store and keeping the original product in another store to see if the sales between the two stores is any different. The process is not random, however, often making it hard for conclusive validity to be established.
Experimental Research Methods
Unlike causal-comparative research methods, true experiments are both controlled and random. Using our packaging example, researchers might choose to conduct a “real” experiment by randomly choosing samples of people to poll about packaging options, presenting the packaging choices within a controlled environment that is identical for every respondent and each group. In a casual-comparative scenario, sales might be affected by each store’s temperature, the weather, the location of each store and so on. A real experiment controls all possible variables so that true causation can be corroborated more definitively.
Want to Learn More?
If you would like to learn more about quantitative research methods and how to use them to gain actionable insights that strengthen your business products, practices and performance, contact our team at Communications for Research (CFR). We have over two decades of experience crafting quantitative and qualitative studies that get results!
You might also like to download our free survey screening template, “Demographic and Screening Questions: Survey Design Best Practices,” to learn how you can maximize your data collection by properly screening the right sample of people.