4 ms·

Definitly. For me it was enough to see no obvious negative correlation(as was claimed) even though the data is not great and no analysis for other variables was

• kqr 5y ago
  Sorry, I don't follow. How is it enough to see no obvious negative correlation? This diagram says literally nothing since many of the important causal pathways involved are not sufficiently controlled for. Before and after seeing this diagram, you ought to have effectively no change in your belief.
  • ketzu 5y ago
    Original claim was: Internet speeds are low in Germany (and other countries) because of the high age of the people living there. Countries with a younger population place higher importance on the internet and therefore have higher internet speeds. The various data sources on "internet speeds" are not good, age is better docummented and probably less questionable (I never worked with humans to really establish any of that). I was under the impression that good data on internet speeds is hard to get, so I went for worse data to get a first impression. If the skew is not created by bad data (e.g., variable selection bias in different populations) but by other causal relationships, I believe it discredits the factor of population age as the main contributor. I was and am willing to discuss those weaknesses of the data in a discussion. I think I stated with those graphs that I stayed unconvinced of the claim, not that they are conclusive proof. I think of them as a form of exploration to see if a claim is worth further investigation (especially considering the amount of work needed for that).