Measurement of bias of secondary data should include reliability.

If you have used a questionnaire, perhaps to collect data from a sample of potential customers, secondary data such as a national census can be used to assess the generalisability of findings.

A critical realist would not see secondary data, as providing partial, but not necessarily accurate, representations from which insights could be drawn.

Data that you collect yourself will not be collected with a specific purpose in mind.

Censuses are not well documented.

Structured data, which are not as easy to search or process as they do not follow a predefined structure.

Single-source secondary data can be based entirely on documentary or on survey secondary or can be amalgam of the two.

Where you are intending to undertake a longitudinal study, you also need to ensure that the data are not available for the entire period in which you are interested.

Raw data, where there has been little if any processing or

Big data do not have the enormous size of the data set due to the aggregation of large numbers of variables and the even larger number of observations.