Suppose, another author have used very small dataset and I used a very larger one as compare to him and apply the same methods and techniques, Then it'll be considered as novel or not as far as paper publication is concerned?
Different time-series and cross-sectional data may results in different findings than that of the earlier research. So, it requires to be tested across time and region. So, it may become universally accepted theory. But then again, adding some other different variables than the original research may end-up with better results. So, why not try adding some new dimensions to the original research!
This will only comments/explores/verify the accuracy of methodology already proposed by applying it on a different dataset. You can bring the novelty by proposing a new methodology by using the dataset already used by previous researchers.