In this post, we’ll define quantitative data, share quantitative data examples, and outline the differences between qualitative and quantitative data (and other data types).
But first, let’s take a step back.
From test scores to satisfaction ratings to tweets, 2.5 quintillion bytes of data are generated every day. But not all data is created equal.
An interview describing your experience at a restaurant generates different data than a survey where you’re asked to rate the service, menu, atmosphere, and price on a scale of 1 to 10.
For analysts working with data sets every day, it’s important to differentiate between types of data and understand how each might play into your analysis. Often, diving into data starts with a particular question you’re trying to answer, such as:
- How do demographics influence purchasing behavior?
- Will a product or service change resonate well with a certain audience?
- How can we improve efficiency by removing operational bottlenecks?
Depending on the nature of the question, as well as your budget, time, and available resources, you may need to collect and analyze qualitative or quantitative data—or a mix of both. Here’s an overview of different data types.
(Meanwhile, check out these publicly available data sets.)
What Is Quantitative Data?
As a data analyst, you will primarily work with quantitative data, such as time, height, weight, price, cost, profit, temperature, and distance.
The definition of quantitative data is simply any data that can be counted or expressed numerically. With quantitative data, we are usually trying to answer questions involving quantity, frequency, value, or size.
There are two main kinds of quantitative data: discrete and continuous. Discrete data has a limited number of possible values (e.g., whole numbers from 1-100). Continuous data has unlimited possibilities, including fractions and decimals (e.g., weights or distances like 1/4 kg or 0.5 miles).
Data can also be organized into different levels of measurement, such as:
- Each variable has a different value but there is no order. For example, in a survey where there are values of gender, male and female may come with a numerical value (male = 0, female = 1).
- Data follows a specific progressive order based on values (for example, degree types like bachelor’s, master’s, and doctoral).
- This data is continuous and has an order along a scale (e.g., ratings of 1 to 5). Each value is equally spaced from the value before and after (e.g., distance between 1 and 2 is equal to the distance between 2 and 3).
- Data is continuous and has an absolute zero. Ratio data is very similar in properties to interval data. A good example is temperature, which can go down to zero degrees.
Data can also be classified as univariate (single variable), bivariate (two variables), or multivariate (multiple variables). Single variable data comes as a list, while bivariate or multivariate data can be expressed with rows and columns. In the case of bivariate data, the two variables relate to each other (for example, age and bone density in X-rays).
How Do Qualitative and Quantitative Data Differ?
Qualitative and quantitative data are very different, but can both be useful for capturing the complete picture of an event.
Qualitative data depends on descriptive words, images, and observations. It is defined as “data that approximates or characterizes but does not measure the attributes, characteristics, properties, etc., of a thing or phenomenon.” Qualitative data is subjective, exploratory, and aimed at increasing understanding of a problem or situation.
This analysis may involve finding keywords from interviews or survey responses based on how often certain combinations repeat themselves. In statistics, we call qualitative data “categorical data.” An example of qualitative data is a description such as, “The weather is cold and rainy during the month of October,” or, “The package arrived on time and the customer service was courteous.”
Qualitative data analysis methods include content analysis (drawing conclusions from text, media, visuals, and objects), narrative analysis (highlighting oral or written stories and experiences), discourse analysis (observing interactions and conversations in a specific environment), and grounded theory. Focus groups, in-person observations, interviews, and archives like literature reviews and newspapers are all methods for obtaining data.
Qualitative data is valuable in market research when understanding the opinions of customers and brainstorming key problems or areas for improvement. It allows researchers to ask more open-ended questions and collect a wider range of potential responses. It is useful for open innovation, brainstorming, and idea generation.
Compared to quantitative analysis, qualitative interviews can be more free-flowing conversations instead of following rigid frameworks, which may allow for more creativity. Since the required sample size is smaller, qualitative research can also be less expensive than quantitative studies.
The downside of qualitative data is that it is largely unstructured and it is time-consuming to mine lengthy recordings or text for actionable insights. Additionally, it can be hard to replicate the results of qualitative research in future tests and to make sure that the people interviewed are representative of the larger population. When interpreting qualitative research, you also need to recognize how your own biases may influence your conclusions.
Today, there are a few tools that make qualitative analysis easier. Search engines like Google and Yahoo use alt text to categorize and group images. NoSQL databases are another means of storing qualitative data.
Quantitative data comes into play when you need to test or validate specific hypotheses. It is more objective than quantitative data and can provide a more representative picture of a larger population.
Some of the challenges with quantitative data are data quality, as most research requires a large sample size of subjects to yield statistically significant results. Furthermore, researchers sometimes fall into the confirmation bias trap, verifying hypotheses instead of generating new ones based on evidence.
Mixed-method studies contain a mix of qualitative and quantitative data. An example would be a company that wants to understand the overall demand for a new product offering and so it surveys the population of an entire city. After identifying the demand, the company then selects some representatives for more in-depth phone screenings and trials to refine its prototype.
How to Obtain and Use Quantitative Data?
Quantitative data can come from experiments, research surveys with ratings or closed questions, and controlled observations.
With quantitative data, researchers are mostly concerned with the accuracy and integrity of the data, trying to remove bias by capturing verifiable and testable results.
When analyzing quantitative data, you will normally have to start by validating the data. According to SocialCops, this involves checking for fraud (confirming that respondents were actually interviewed), conducting screening to verify that the respondents were selected by well-defined criteria and are representative of the overall sample pool, making sure that the procedure was accurately followed, and looking for completeness (answers to enough questions to provide a holistic view of the bigger picture).
After the data validation, you will need to edit the data. Sometimes there are missing data entries or data that was entered incorrectly. Through some simple data validations, you can identify and remove outliers that might give misleading results.
Once the data is cleaned, you can group it into categories for ease of analysis (e.g., breaking into age ranges 0-17, 18-25, 26-35, etc.).
With descriptive analysis, researchers often identify numerical patterns in the data, looking for mean, mode, median, frequency, and range, which helps put the overall data set into context.
With inferential analysis, they can search for relationships between different variables and predict future outcomes. Correlation describes how two variables are related, while linear regression estimates an unknown variable based on a related known variable. Analysis of variance statistically tests the degree to which the means of two groups differ.
Some of the tools used in quantitative analysis include the following:
- SPSS (Statistical Package for the Social Sciences) allows for many types of statistical analysis like chi square, T-test, ANOVA, factor analysis, etc. It is appropriate for complex types of data and applicable to fields of medicine or social science.
- R is an open-source language for manipulating and visualizing data.
- STATA works best with simpler datasets in econometrics and has an easy-to-use interface offering “graphs” and “statistics” features. Like SPSS, it also allows for programming.
- SAS is helpful in business contexts because you can use it for forecasting, improving efficiency and quality, and measuring performance.
- Excel has many formulas that can lead to data-driven insights. You can also learn specific data programming languages within Excel like VBA.
Once data needs to be taken to the next level and re-created, data scientists or data engineers use programming tools like Python, R, and Matlab/ID for coding.
Why Quantitative Data Matters?
Relying on assumptions or precedence (e.g., “this is the way things have been done for generations”) can lead to expensive mistakes. Data-driven experimentation and decision-making is increasingly important for organizations that want to cut costs and increase revenue potential by understanding how or why processes work the way they do.
Springboard’s Data Analytics Career Track can help you characterize data and then make inferences and decisions.
The demand for data analysis skills is on the rise in every industry and, no matter which field you choose, understanding quantitative data will put you on the right track for an exciting career.