How do you form good analyses from faulty, biased and unreliable data and methods?

To properly evaluate the reliability of your data, you must first identify its type. If the data collected is heavily based on self-reported info, such as in surveys, it could cause issues with response bias and exaggeration because of social pressures. The same goes for methods that are outdated or have flawed assumptions.

If you have identified any possible problems, it’s important to use additional sources of information in order to correct them. You could do this by looking for alternative data or using triangulation methods. When presenting results, it may be helpful to provide context about the sources of data so that future audiences are able to make their own assessment.

If all else fails then transparency about what was found during analysis should always be provided regardless of how good (or bad) the outcome may have been – this will allow others to evaluate whether there were indeed any significant inaccuracies present within that particular dataset which can help inform future decisions related to similar projects moving forward. In the end, these measures will ensure that a reliable output is produced which can be used for short and long term decisions.