In our modern world, data and data analysis are increasingly important. Data is often used either to get a message across or to support and verify the message or their instruction. We have all become very familiar with data sets as the COVID-19 pandemic is described and tracked.
Science Data. Image Credit: Kite_rin/Shutterstock.com
This article will look at how should we analyze data. Data is a vital component of scientific study and the advances in computers have made the acquisition, storage, and use of data more accessible.
We need to start by recognizing that data can be split into qualitative and quantitative data. These types of data are captured and interpreted in different ways. Although qualitative and quantitative data and data analysis are different they can be interrelated. For example, you can determine what quantitative data to measure by analyzing your qualitative data.
Quantitative data provides evidence and predictions whereas qualitative data provides explanations and context.
Qualitative data usually measures characteristics, for example, hair color. It is non-statistical and often unstructured and is used to answer why something behaves in a certain way. Qualitative data is stored in text and documents, audio and visual recordings, observation notes, and interview transcripts rather than in graphs and charts. Qualitative data is usually gathered using open questions such as "Why?" "How?" and "In what way?"
Qualitative data is used to develop theories and hypotheses and to gain an initial understanding of a phenomenon. It can also be used to test these theories and to determine what quantitative data to gather in further analysis.
Qualitative data is largely about language and measures, how things feel or act, and people’s motivations and opinions.
Quantitative data is rigid and measured in numbers, which means that it can be analyzed using databases, spreadsheets, and graphs. It is usually generated by tests, experiments, surveys, and market reports and is usually gathered by measuring specific parameters or by using closed questions.
Quantitative data can be discrete or continuous. Discrete data is data that cannot be broken down for example, the number of students in a class whereas continuous data is data that is continuously changing, an example being the speed of a train on a journey that may be constantly changing.
Since qualitative data is not measured by numbers, it normally requires human interpretation. It is less easy to store qualitative data in a database. It can be done in a NoSQL database where documents can be stored and retrieved for analysis. Qualitative analysis usually requires human interpretation of data.
Qualitative data analysis can be divided into different parts as outlined below:-
- Content analysis where the data is analyzed for specific content and terms.
- Narrative analysis where data in the form of surveys, interviews, or notes are analyzed to see if there is a common thread in the responses.
- Framework analysis where data is assigned a hierarchy so that it can be coded according to predetermined priorities.
- Grounded theory analysis where data from a single case are analyzed and additional data sets are examined to see if they add to the original theory.
The data that is observed may then be examined to determine if there are relationships between different observations. This would most likely need to be done by human analysis. Some software may be able to do some of the analysis by identifying specified terms in the data collected.
The advances in Artificial Intelligence (AI) will create opportunities to improve qualitative data analysis. Computers and algorithms are much quicker than humans at identifying and sorting key phrases but they are not good at interpreting sentiment. It remains true that a machine will do exactly what it is told to do so the human factor in setting up algorithms and AI programs is still crucial. The use of artificial intelligence and computers in qualitative data analysis will continue to improve with time. The use of AI is likely to increase but human oversight and intuition will continue to be vital in many situations.
In quantitative analysis, the numbers associated with the data are entered into a database or spreadsheet, and the different attributes can be analyzed and quantified. For example, if we look at deaths by Covid we can look at daily mortality, average daily mortality, cumulative mortality, and others. This data can be presented as numbers, graphs, or charts depending on what is required.
In contrast to qualitative analysis, quantitative analysis is much easier to automate. Databases and algorithms can gather and sort numbers fairly easily and very quickly. This facilitates the extraction of data sets and can be automated allowing the computer to search, extract and present the data quickly. The human factor is only required to tell the computer what to search for and how to present it.
Quantitative vs Qualitative Data Analysis
The decision to use qualitative or quantitative analysis is not a simple either-or question. Many studies, particularly in social sciences, will need to employ both types of analysis, even though the two types of data are often used for different types of study. They can however be used in tandem.
For example, quantitative data analysis might indicate how prevalent a phenomenon identified in a qualitative study is. Qualitative data could show why the phenomenon identified in a quantitative study is prevalent.
As mentioned above qualitative data can be used to determine what data should be measured in a quantitative study.
One key factor remains whatever type of analysis is being undertaken. The interpretation of the data by experts will always be required and communication skills are important when explaining data to the general public. These skills are unlikely to be automated by machines.