Data Collection

Since you've been using StudySmarter to revise physics, you've been feeling so relaxed about GCSEs that you're sleeping great. "That was my best night's sleep in years!" you exclaim as you wake up. However, as the excellent physicist you are, you know this statement is merely a hypothesis. To use the scientific method to test your hypothesis that last night's sleep was indeed record-breaking, you'll need to collect some data. This article introduces the basic scientific principles of collecting data and why these are important to ensure reliable results.

Data collection meaning

The scientific method has enabled countless aspects of modern life, being responsible for the development of medicines, aeroplanes and computers. It defines the process of making a hypothesis, experimentally testing it, analysing results, and using this to iteratively refine hypotheses. The key activity needed to test the hypothesis is clearly the testing - however, to effectively analyse the results of any experiment that we run, we need to collect data from it.

The purpose of any data collection is to collect high-quality evidence that can be analysed to come up with convincing and trustworthy responses to the questions or hypotheses proposed.

Differences between qualitative and quantitative data

Imagine you are ordering a custom-made suit or dress. When the tailor asks you how long you would like the legs to be, you have two options: you could say "long" and hope that the tailor is able to get the length about right for you, or give them a measurement, "30-inch inner leg" and know that they now have your exact size. While both these approaches could get you a great outfit, there is a key difference - the first approach uses qualitative data, while the second approach uses quantitative data.

Qualitative data

Returning to our example of a scientific study on the quality of sleep, you might gather some qualitative data about this by conducting interviews with people who have just woken up. You could ask questions like: "did you sleep well last night?" or "how rested do you feel?". These leave the answer open-ended, so the sleepy interviewee is free to answer however they like.

Qualitative data collection methods

• One-on-One Interviews: On a one-to-one basis, the interviewer or researcher obtains data from the interviewee, by asking them pre-prepared questions. These questions are designed to help the interviewer draw out information useful for making relevant conclusions.
• Focus groups: This type of research is done in a conversation format with a group of people. The group is generally restricted to 6-10 persons, and the debate is moderated by a moderator. The members of a group may have something in common depending on how the data is to be going to be sorted. A researcher doing a study on basketball players, for example, may select a group of people who are basketball players and a group of those who are not. They can then compare the data from the two groups.
• Record keeping: This strategy makes use of previously existing data and credible sources of information. This information can be utilized in future studies. It's like going to the library. There, one can go through books and other reference Materials to gather essential facts for the research.
• Case studies: An in-depth investigation of case studies is used to acquire data in this strategy. This method can be used to evaluate both basic and complicated issues. The strength of this strategy is how well it draws conclusions using a combination of one or more qualitative data gathering methods. A case study could yield quantitative as well as qualitative data.

Quantitative data

In addition to collecting qualitative data about how people's self-reported quality of sleep, you may also want to record or measure some quantitative data. While this would lack the depth and detail of subjective information, a more controlled set of data gathered quantitatively allows for more straightforward numerical analysis and comparison of results. Some data you could record to quantitively measure the quality of someone's sleep could be the duration of the night's sleep, or asking participants to rate how rested they feel on a 1-10 scale.

A major benefit of using quantitative data is that it is objective, and not susceptible to external influences such as bias. The objective nature of quantitative data also allows us to directly compare different results. While two people may have different qualitative descriptions of what makes a good night's sleep, we can directly compare the quantifiable durations that were asleep. Finally, if data is collected in quantitative terms it is much more straightforward to statistically analyse.

Quantitative data collection Methods

• Survey/Questionnaire: In order to obtain data from a group or a large number of people, surveys or questionnaires are often used. Both quantitative and qualitative research can be collected from surveys/questionnaires. By controlling the questions people are asked data can be gathered from humans while restricting the range of answers, which makes the subsequent processing and analysing of data simpler.

• Experiments: Many hypotheses are tested through means of an experiment. This would involve carefully designing an investigation so that the variable(s) being measured are isolated from uncontrolled factors that could influence the result, and recording objective data about the experimental results using appropriately calibrated sensors and measuring equipment.

Types of data collection

There are two main categories of data collection - primary data and secondary data collection.

Primary data

This is data directly gathered by the person or organisation conducting the investigation, for the purpose of conducting the investigation. For example, in our sleep quality study, a source of primary data would be you running an experiment which timed the duration of people's sleep. Alternatively, qualitative data collected through one-to-one interviews would also be primary data.

Secondary data

Secondary data is information that was not gathered for the specific investigation it is being used for, even if it is still useful. Datasets created by a different study or group can be used as a source of data for another investigation, and sometimes this is the best source of data for certain information - for example, if you wanted to find out the average height of 15-year-olds in the UK to compare to the participants in your sleep investigation, this data would be better obtained from a secondary source such as national anthropometry statistics. It's unrealistic to conduct a huge experiment for every study that may need the same dataset, so by conducting a single large-scale study other investigations can use the results as a high-quality source of secondary data.

Derived data

Derived data is data that has been created or calculated using other data, rather than directly collected or measured. An example would be a value for average height - this figure will have been calculated using an original dataset of many individual height Measurements.

Examples of data collection

We have already discussed several examples of different types of data collection throughout this article, but here are some more to provide even more context to your understanding of data collection.

• A crisp company conducting taste trial focus groups on a new type of crisp - a focus group like this would gather qualitative data from participants about their experience of eating the crisp, which may be used to select products for further development or improvement. This data would also be primary, as the investigation is being run specifically by and for the group making the new crisp.
• The crisp company calculating an average crisp consumption per person, based on data of individual diets from a study run by the NHS. This data would be secondary, as the crisp company did not run the original study, and it would also be quantitative, as the average crisp consumption is a numerical value. This would also be derived data, as it is an average value calculated from a source dataset.
• The average weight of packets of crisps measured by the factory quality control - This data would be primary and quantitative, as it is gathered by the same group who are using it and it's numerical. The average value would again also be derived data, as the original dataset contained the weights of each individual packet.