Hacker News Reader

Comment by josu | Hacker News Reader

I can't agree with you more. This would have been a nice study if it didn't reach any conclusions; just laid the data, pointed out all the possible reasons why their analysis could be wrong and paved the way for a better study.

One example, their data only covers from 1994 to 2011.

PS: Direct link to the study http://papers.ssrn.com/sol3/papers.cfm?abstract_id=1572085

View on HN

The standard format for a scientific paper includes a discussion section, where you're supposed to draw such conclusions and interpret the results and their significance.

Then this document should not be called a "scientific paper" or/and should never be published.

Sorry, what? It shouldn't be called a scientific paper because it follows the standard format for one?

Then it wouldn't be a study; it would be data collection.

I'm ok with that. But I guess that "data collection" documents don't count towards academia points.

Sort of like:

Here's the data. Analysis is an exercise left to the reader?

I guess what I'm trying to point out here is that while collecting data does take up most of the bulk of the work, data is meaningless without analysis and direction to inform further gathering. That's why the field of study called statistics exists. It's not just about academia points.

On the same vein, the reader should be critical of the analysis; it should not be taken as accepted truth. And the writer should try to predict and address criticisms.

I would agree with you when the analysis is correct. However, no analysis is better than a wrong analysis, especially in cases where politicians may use the studies to pass laws that may hurt citizens. Remeber the whole Reinhart And Rogoff ordeal.

Reply on news.ycombinator.com