Hacker News Reader

Comment by lotharbot | Hacker News Reader

> "the whole premise of 'couldn't find evidence thus fact didn't exist' was fishy"

Not really. The one thing he's correct about is that, if NINA was a real and widespread phenomenon, it would show up in certain pieces of documentation of the era -- newspaper want ads, for example (we actually have fairly large archives from the relevant era; it's not like something from thousands of years ago where it's plausible that no documentation would have survived.) His problem was that he didn't look deeply enough, and once he drew the initial wrong conclusion, he invented a totally wrong theory to explain it.

View on HN

> it would show up in certain pieces of documentation of the era

True. I think what he assumed was that it would show in other types of documents (like discrimination lawsuits, news showing revolt of the Irish against this policy, etc)

But I suspect this was very common and accepted, unfortunately.

Except that, as Fried, showed, it did. Not quite anti-discrimination lawsuits, since those laws didn't exist at the time, but there's an example of some Irish filing a libel lawsuit against a newspaper for printing a NINA advertisement, and of an Irish protest against the posting of a NINA sign at a mill.

Actually the mill was a different situation; the mill owner adverstised the jobs and large numbers came to apply, Irish and non Irish. The non Irish came armed (guns and knives) and forced the mill owner to put up a sign saying "No Irish Need Apply". The Irish being unarmed decided to leave the scene.

> he didn't look deeply enough, and once he drew the initial wrong conclusion, he invented a totally wrong theory to explain it.

My reading is the opposite. He started with a theory and then did a shallow search to bolster his pre-conceived theory, ignoring or discounting any findings that didn't fit.

Following the links from the OP, I see the first researcher states there are fewer than 2 NINA ads per decade in the Brooklyn Eagle, while following the example of the search I counted more than 8 ads in Brooklyn Eagle from the 1870ies before giving up the point as discredited.

I wonder if the big reason for these completely different results in the underlying counting experiment is that the later research was done after ocr had improved so much that the searches had very different samples of the papers.

Reply on news.ycombinator.com