On what basis were the answers classified as original research?

Question

Two answers (one of which I wrote) to this question on the commonest boys names in England were recently deleted on the basis they constituted "original research".

I would like to understand why this judgement was reached.

The general rule is:

It is up to the answerer to provide valid, verifiable and potentially replicable evidence, and to show convincingly that it is relevant to the question. Every answer should be suitable for review and voting by our community of experts in evaluating evidence. Answers which require non-trivial specialist expertise or are anecdotal in nature are not acceptable due to the nature of this community.

My answer consists of nothing more than drawing some charts and pictures based on the single definitive data source that actually answers the question. Every fact is directly verifiable by any user as the raw data is linked and is a simple list of names and usage counts. The pictures/charts make the results easier to see but involve no analysis that cannot be replicated or verified by examining the tables in the source.

The drawing of figures from data is explicitly allowed as the answer to this meta question states:

...we allow that as long as it's trivial. E.g. a histogram, scatter plot etc are all fine.

The key to not falling into "original research" is that the answer should stand without the plot and it should only serve a further explanation.

For example this answer of mine is as close as it can get to original research without being so: the plot is a summary and there's independent evidence being presented.

My answer is entirely compatible with this judgement as it involves no extra analysis at all never mind trivial analysis. I have simply visualised the data in the source table (and the visualisation could be reproduced by anyone who knows how to use a spreadsheet and does not require any "community of experts"). Every part of the answer can be independently verified by any site user by examining directly the content of the definitive source data.

So why has the answer been deleted as "original research?

Answer 1

This comment seems to highlight the fundamental disagreement here.

So what you are saying is that I need a peer reviewed reference to point out that "Mohammad" is a variant spelling of "Muhammed"?

No, what you need is a peer reviewed reference (or references) saying

1. "Mohammad" is a variant spelling of "Muhammed".
2. That it is the only variant spelling and names like "Amid" do not count.

3. That the competing European names do not have variant spellings that would increase their counts.

   For example, Mat, Matt, Mateo, Matthew, Mathew, Mathieu, Matthias, and Mathijs are all variants of the same name. And John, Jon, Juan, Jan, Yanni, Shawn, Sean, and Ian are all variants of the same name. These examples should not be viewed as exhaustive. And of course they are arguable. That's the point. Reasonable people may argue that Amid is not a variant of Mohammad but that Ian is a variant of Jon. Or vice versa.

4. Should feminine versions of the same names count? Matilda is a variant of Matthew. Juanita is a version of John (Juan). Christina comes from the same base as Christopher. The claim excludes them—but should it? That seems to favor the claim, as Mohammad has no female variant while several of the more popular Christian names do. Or what about Mary? A very common biblical name given primarily to girls (Mario is a male version).
5. Then to really answer the question, you need peer review of using those assumptions in this context. Although we generally accept the answer as helpful without that.

In the end, you are likely to have a response that says something like, "Under this set of a assumptions, the answer is yes, but under a different set of assumptions the answer would [or might] be no." Because claims like this tend to have certain underlying assumptions. And in some cases, the claimant deliberately picked the assumptions that bolstered the claim.

And that's why we want expert peer review. Because it flushes out what claims are justified and which are not. And it can discover underlying assumptions that are not being measured. For example, that Ian does not count as a form of Jon, even though the Latin spelling of a name pronounced like Jan would have been Ian. Of course, they would have pronounced that like yon rather than ee-un.

I tend to agree that combining Mohammad and Muhammed improves that count. But the same can be said of the various versions of Mat and Jon. There's even an argument that Matthew, Mark, Luke, and John should count together, as those are all references to gospel writers. And that's the flaw with original research. It's hard to tell when to stop. Those are all fine debate points. But we aren't a debating format. Ideally we should be able to just refer to and explain the underlying study.

This is especially true with questions like this, where the base claim is often made out of xenophobia. The implicit claim here is that there are more Muslim children being born in Germany/Berlin than Christian children. That claim is almost certainly untrue, as a higher percentage of Muslim children have some version of the name Mohammad than Christian children have any individual biblical name (Jesus, Christopher, Matthew, Mark, Luke, John, etc.).

TL;DR: Specifically, determining how to group the names (and which names to group) is non-trivial, requiring expert analysis. As explained in this comment:

You derive/choose/specify/find the variants, group by them and count and this is key to the answer but it's obviously your own research. The decision on whether grouping the name or not is a valid research step is again, yours. The decision on whether NOT grouping other names or doing so is still yours. There's obviously tons of stuff in there which is out of your own imagination/rationality and not merely a graph on top of data.