When researchers use web based analytical tools they should be aware of the statistical properties of data. It is easy to loose track of the limits of these data.
The methods used to collect data may limit valid use and make it risky to use particular dataset to answer the question that the researchers bring.
For example, if a researcher is looking for data about cigarette smoking in a town, using national or state data may provide a benchmark for comparisons, but attempting to analyze the survey data and extract information for the town would likely produce unstable or unreliable results due to the size and methods used to construct samples.
Researchers need to assess the impact of the sample size and its selection procedure on the usefulness of the results.
Data might not be accessible at the level of specificity needed to provide a good answer.
Local area data may not be available
The data might be protected and there for unavailable to safeguard the privacy rights of respondents or subjects or
The data might not have been collected for the time period required.