What Does 'Data Collection' Actually Mean & What Methods Are Typically Used For It?

When you hear the term "data collection," your mind may immediately jump to nefarious companies harvesting your sensitive personal data with the intent to store and sell it. And while that can be true, data collection is actually a much broader term that refers to a business or individual systematically gathering information for the purpose of evaluating a hypothesis, researching user demographics, recordkeeping, and better understanding just about anything else you may want to make an informed decision on. Data collection is used in some form by nearly all major sectors of industry, including medical, tech, financial, and more.

Methods of data collection vary depending on what type of information you may want to gather. Five of the most common types of data collection are questionnaires and surveys, interviews, focus groups, observations, and document analysis. You may have seen an ad on a social media platform like YouTube or Instagram asking you to pick which advertisement you'd rather see, or which of the following brands you've heard of. This is an example of data collection via surveys, and it's used to understand both your specific preferences and broader market trends.

Data collection methods can be further broken up into categories such as primary and secondary, manual or automated, and quantitative and qualitative. Primary data is collected directly from a source, whereas secondary data collection analyzes existing documents or records to draw new conclusions. Manual data collection involves a person actively gathering information, such as in an interview for a research study. Automated data collection is completed with little human involvement, like a customer satisfaction survey after you place an online order. Quantitative data refers to numerical information like measurements or 1-10 ratings, and qualitative data documents people's observations, opinions, and experiences with a product or process.

A survey is a broad research method that encompasses designing questions, distributing them, and analyzing the data. Questionnaires can be standalone or part of a survey — they're a set of open- or closed-ended questions that are designed and distributed by the researcher to gather information and feedback about a certain product or topic. Surveys are common across several industries because they allow researchers to collect information directly from the target audience. Open-ended questions give you room to share your thoughts, while closed-ended questions may be multiple-choice, yes/no, or numerical, depending on what the researcher wants to measure.

There's a good chance that you've been asked to participate in surveys several times over the past year. After placing an online retail order at Amazon, for example, you'll see a pop-up or email asking "How was your delivery?" with an option for "It was great" or "Not so great." You may also see a QR code or link on a fast food receipt asking for your feedback on the service you received, often accompanied by the chance to win a prize. These surveys collect data to better understand how customers view different aspects of the researcher's business.

Because of their flexibility, surveys can be either automated or manual, and can measure both quantitative and qualitative data, even in the same survey. A paper questionnaire asking students a few quick questions about their university and an automated questionnaire on a mobile app are both survey formats. Typically, surveys collect primary data, but a researcher may also analyze existing survey results as a secondary source to draw conclusions.

Interviews are similar to surveys but involve direct communication between the researchers and the participants. Instead of simply filling out a form to answer questions, this method of data collection can take place in person, over the phone, or via video call and allows the target audience a chance to openly respond with their thoughts and feelings. Interviews can be structured, semi-structured, or unstructured. Structured interviews adhere to a predetermined set of questions that are asked to everyone the same way. Semi-structured interviews follow a base set of questions but allow the interviewer to ask dynamic follow-up questions. Unstructured interviews aim to gather information or feedback on a topic and are largely conversational, with the most room for exploration.

Each method has its own use case. A structured interview is great for collecting quantitative data, whereas an unstructured interview allows researchers to get participants' off-the-cuff thoughts and reactions. Interviews, like surveys, are also used in several industries, from healthcare and tech to retail and service. Pew Research Center conducted semi-structured interviews with 13 journalists and editors to better understand what 'news' means to Americans today, asking questions such as, "As you look at the technological changes and consumption habits in the information industry, do you believe that the definition of news is changing/evolving?"

Interviews are particularly useful because of the depth of data they can capture, but because they are typically manually conducted, they can be very resource-intensive (especially when working with a large pool of participants). Like surveys, interviews are initially primary data sources, but transcripts can later be interpreted and used as secondary sources. They can also span both quantitative and qualitative questions.

Focus groups are similar to interviews, but involve small groups of around six to 10 participants gathered together with a moderator to discuss a topic and record the responses. The moderator guides the group, either in person or via video call, by asking questions and encouraging conversation. All participants have a chance to share their thoughts and feelings, which can spark new ideas that one-on-one discussions may not uncover. It also provides the moderator with a broad view of how the group reacts as a whole.

Focus groups, like the data collection methods mentioned previously, are applicable to a wide range of industries, including tech, healthcare, and many more. In an American Journal of Lifestyle Medicine study titled "Patient Expectations: A Qualitative Study on What Patients Want From Their Healthcare Providers to Support Them in Healthy Aging," researchers conducted 22 one-hour focus groups among a diverse group of 79 participants. The discussions resulted in three qualitative conclusions: Patients want respect and collaboration, information on alternatives to medication, and information on preventative care.

Focus groups are almost always conducted manually and are usually aimed at collecting qualitative data. Quantitative data may be collected as well, though it is not usually the focal point of the discussion. While the moderator collecting data directly from participants yields primary data, a focus group recording can serve as a source of secondary data in later research.

Rather than relying primarily on asking questions to participants, the observation method of data collection involves watching people's behaviors and interactions to record findings. Observation can take many forms depending on how much the researcher guides the participants and interacts with the environment — we'll discuss three common observation methods here. In naturalistic observation, the researcher does not interfere or provide any guidance to the demographic or area they're researching. Researchers simply watch and record, such as when observing how shoppers at a supermarket navigate aisles and how much time they spend in each. 

Participant observation is similar, but involves the researcher directly participating in the process they're studying and possibly interacting with other subjects. Going back to the supermarket example, a researcher engaging in participant observation may work alongside staff to better understand how employees greet customers, manage inventory, and go about daily responsibilities. Controlled observation is much more structured, putting voluntary participants in a predetermined setting such as an office or lab and asking them to complete a task or react to stimuli. These methods are not necessarily mutually exclusive, either. An observation can be both participant and controlled, for example.

Observations are useful because they allow researchers to understand what people actually do naturally, rather than what they may think of as an answer to an interview question. They're typically conducted manually, though camera systems and sensors can be used for automated observation, and they can produce both qualitative and quantitative data.

In data collection, document analysis is exactly what it sounds like. Researchers use existing records, reports, social media posts, interviews, and other data to identify patterns, make informed decisions, and create resources like infographics. This offers some significant advantages — and poses a few important disadvantages — over data collection methods that require a researcher to ask questions directly to participants. Document analysis lets a research team review large batches of data without recruiting participants or conducting interviews, surveys, or focus groups themselves.

Researchers can easily obtain both quantitative and qualitative data using this method, provided the studies they're looking for already exist. An Amazon seller, for example, may study user reviews on a product to determine qualitative data such as which complaints arise most often and what aspects their customers are most keen on. As for quantitative analysis, a researcher might analyze a company's financial reports from the past several years to glean insights into revenue and expenditure changes. But this can only be done if the documents already exist.

This is one of the major downsides of document analysis: The specific topic a researcher is looking for may not have been previously studied and recorded. The recorded data could also be unreliable, outdated, or inaccurate (like in the case of fake Amazon reviews). This is why it's vital for a researcher to vet and verify the sources of information they use. Document analysis can be primary data, though it often relies on secondary data that was originally studied for another purpose.