Understanding how the data was collected can reveal its strengths and weaknesses.
Sampling Method: Was the sample a random, representative sample, or was it a convenience sample? A random, national sample will have higher generalizability than a non-random sample from a single city or institution.
Data Collection Instrument: Was the data collected via a survey, an experiment, or administrative records? The method can influence the type of bias present. For example, a self-reported survey might be subject to social desirability bias.
Auditing and Quality Control: Were there any processes in place to audit the data and ensure its quality during collection? The documentation should explain the quality control measures taken.