With deductive reasoning, one uses known information, assumptions, or generally accepted rules to reach a conclusion. In statistics, a data scientist uses this concept (in an attempt) to repair inconsistencies and/or missing values within a data population.
To the data developer, examples of deductive correction include the idea of converting a string or text value to a numeric data type or flipping a sign from negative to positive (or vice versa). Practical examples of these instances are overcoming storage limitations such as when survey information is always captured and stored as text or when accounting needs to represent a numeric dollar value as an expense. In these cases, a review of the data may take place (in order to deduce what corrections—also known as statistical dataediting—need to be performed), or the process may be automated to affect all incoming data from a particular data source.
Other deductive corrections routinely performed by the data scientists...