When a researcher is collecting information, the data is described in various ways. One of the simplest classifications of data is into primary and secondary. Primary data is information that the researcher collects herself, so it is firsthand data. Secondary data, on the other hand, is information collated by other people, such as studies done by other researchers. Although primary data is ideal for a researcher, as he can home in on the specific information he is looking for, secondary data is also useful, because it can tell the researcher whether his research is new. An important point to note when considering types of primary data is that individual data types can have several descriptors.
Choosing a good sample is essential to data collection, if the researcher wants his data to represent his area of study accurately. As the world is so vast, it is generally impossible to collect all the relevant data from all the relevant subjects or area of interest, so looking only at a sample is necessary. In academia, a respected type of sample determination is randomised control, which uses computers to split a sample population into two groups that are as similar as possible. One group can be used as a control to represent normality for that population, and to compare the other sample group with the difference under study against. Other valid, but not as reliable, types of sample might not have controls, or might be made up of single individuals with a single matched control.
Time of collection
A researcher collecting data from the past, such as how many times the individuals in her sample have ever been to hospital, for example, is collecting "retrospective data." If she recruits people to her study and then follows them over time to see how often the individuals have to go to hospital, then this is called "prospective data."
Observational versus interventional
Some types of primary data are simply collected from a sample without the researcher making any change to the sample. One-off surveys of householders about typical broadband usage is one example of an observational study, but more complex studies can also be observational, such as following a group of people accidentally exposed to radiation. Interventional data, though, requires the researcher to change an aspect of the sample, such as administering a new drug and then assessing the people involved over time to see what the effects of the drug intervention are.
Methods of collection
Asking people directly is one way of collecting primary data. Ways to do this include via questionnaire, over the phone, using household surveys or interviews with individuals. Group observation is another way of collecting data. For researching information about fields of interest that do not involve people, direct observation is useful, as the researcher has to make her own measurements.
Quantity versus quality
Data comes in various forms, and the forms that are used most often depend on the field of interest of the researcher. Sociologists and psychologists, for example, are interested in abstract concepts like feelings, knowledge and behaviour, which are all types of qualitative data. Quantitative data, on the other hand, involve information that is either numerical or can be easily converted into statistics, and is more likely to be collected in fields like finance and medicine.