6.4.3. Validation of results

In order to have condence on the data collected from a sample population, three validation tests are adopted usually.

The initial objective of validation is to ensure that the data we collected is reliable. This is important because if the data is flawed, any conclusions we draw from it will also be flawed. To build this confidence, three critical tests are carried out on the data.

Think of validating data like checking ingredients before baking a cake. Just like you would check if your flour is fresh, your eggs are good, and your measurements are correct, in data collection, we ensure that our data is accurate and consistent before using it.

The rst simply considers the consistency of the data by a eld visit normally done after data entry stage.

The first validation test involves conducting field visits after the data has been entered. This means that someone goes out to the real locations where data was collected to verify that the information matches what was recorded. It helps identify any obvious errors or inconsistencies in the data.

This step is like checking a restaurant's reviews by actually visiting the restaurant. If a review says the service was fast, but you find yourself waiting an hour, then the review (or in this case, the data) might not be accurate.

The second validation is done by choosing a computational check of the variables. For example, if age of a person is shown some high unrealistic values like 150 years.

The second validation involves performing calculations or checks on the collected data to look for inaccuracies. For instance, if the age recorded for an individual seems unreasonably high or low, it will be flagged for review. This computational check helps to catch errors that may not be visible during a field visit.

Imagine you're checking a math test, and one answer is 500 when it should be around 5. This doesn't make sense and indicates an error that needs to be addressed, similar to checking for unrealistic ages in the dataset.

The last is a logical check done for the internal consistency of the data. For example, if the age of a person is less than 18 years, then he cannot have a driving license.

The third validation method uses logical reasoning to check if the data makes sense internally. It looks for contradictions within the data. An example of this would be checking if anyone under 18 has a driver's license in the dataset; if they do, it indicates an error. This ensures all rules about the data are followed.

Consider the logic of a school registration; a 3-year-old cannot register for 12th grade. If such a registration appears, there must be a mistake. Similarly, we make sure the data follows appropriate rules and logic.

Once these corrections are done, the data is ready to be used in modeling.

After necessary validations and corrections have been applied, the dataset is considered robust enough to be used for further analysis and modeling. This final step is crucial as it ensures that the models built on this data are trustworthy and will yield meaningful results.

Think about tuning a musical instrument. After you verify that each string is in tune, only then can you play a beautiful song. Similarly, after validating the data, we can move forward with creating models that will provide valuable insights.

--