What validates a theory?

In the simplest terms, data validates a given theory. If the dataset supports the theory, we can say that the theory works or that we can validate it. The data can be obtained through experimental studies and/or numerical modeling. If the theory's predictions match the data obtained either experimentally or numerically, it can be said to be valid.
For example, if (in theory) we say that prices of day-to-day use items rise every year, we just need to collect the price data of some such items over a few years and draw comparisons. If we see that prices of the items have indeed increased every year as compared to the last year, it can be used to suggest that our theory is valid.
Similarly, we say that distance traveled is a product of the speed of an object and time. This can be validated by driving a vehicle at a constant speed for a certain time and measuring the distance. The distance would be equal to the product of the speed of the vehicle and the time of travel.
Thus, supporting data validates a theory.