Wikipedia talk:Numbers need citations

MfD Result Notice

This page was the subject of an MfD discussion closed on 16 October 2006. The result was keep, mark as rejected. Xoloz 15:18, 16 October 2006 (UTC)[reply]

Rounding of statistics and figures

Sometimes rounded statistics and figures are used in articles. For example, the round figure of 150 million kilometres for the Earth-Sun distance is often used, or the phrasing "over 2 metres high" is used for an animal's size. Also, figures give to n decimal places are often rounded to give an easily understandable figure. These are obvious over-simplifications, but how should these sort of "figures" be referenced? Carcharoth 09:49, 2 September 2006 (UTC)[reply]

I would argue to use your best judgment when reviewing articles, but that perhaps the examples you've cited fall under the "self-evident" or "common knowledge" category. While a figure such as 150 million kilometres is fine for a ballpark Earth-Sun distance, and in and of itself, might not require a reference, it would still be preferable to write the statement as something like "...per modern estimatations, the distance is believed to be approximately 150 million kilometres (source)." Or, better yet, a more precise estimate with a source. Dr Chatterjee 12:15, 2 September 2006 (UTC)[reply]

Actually, I'm talking about cases where you just want to mention the figure in passing. In the Earth article, you will find an accurate figure for the size, plus discussion of the orbital characteristics. In the global warming article you may find a passing comment about the Earth-Sun distance where the exact figure is not needed. Different articles, different presentation of the same fact. Carcharoth 01:55, 3 September 2006 (UTC)[reply]

A question

Does Britannica have each numerical figure sourced? If not, why should we try? Grue 10:00, 2 September 2006 (UTC)[reply]

Britannica's numbers aren't writable by the interweb populace. Fredrik Johansson 10:16, 2 September 2006 (UTC)[reply]

Along the lines of what Fredrik is saying, Britannica is written by paid or freelance experts in the fields of each article compiled. They are permitted to use original research where applicable, because the level of trust in their credibility is there. Furthermore, these articles are subjected to extensive peer review among other encyclopedists and experts in each field. They are, basically, water-tight before publication. And, as Fredrik pointed out, they are not editable by anyone off the street. That last aspect is precisely why Wikipedia sometimes gets a bad rap in academic circles: there's no proof of expertise, or expert authorship, in any of the articles written. The closest thing we can come to expert proof is to cite our claims extensively. And numbers/figures/etc would seem to require such citations more than most facts. Dr Chatterjee 12:15, 2 September 2006 (UTC)[reply]

Regarding your last assertion, why so? Christopher Parham (talk) 19:55, 2 September 2006 (UTC)[reply]

Well, in my opinion, numerical figures are the most easily prone to exaggeration, subtle alteration, inaccuracy, etc. Thus, numbers should always have a credible-source citation to back them up. They are among the most highly-"fudgeable" of data. Dr Chatterjee 20:56, 2 September 2006 (UTC)[reply]

In my experience this doesn't seem true at all. This proposal places a strange emphasis on figures over other information that is neither necessary or useful. All information should have a credible source backing it up; the most important items to specifcally cite are those which are controversial or contested. Specific figures range considerably from contested to uncontested and range similarly in the urgency with which they require specific citation. IMO this proposal as a separate guideline would simply confuse our efforts. Christopher Parham (talk) 03:54, 3 September 2006 (UTC)[reply]

m:Instruction_creep Geni 16:35, 2 September 2006 (UTC)[reply]

Chinese whispers with numbers

Another reason for making it a stronger guideline to cite source for numbers is that a game of Chinese whispers with numbers can result in strange results. A number can be quoted accurately from a source, but it can then degenerate as it is rounded up or down, converted to different units, multiplied with other numbers (a process approaching original synthesis), or even just simple calculation errors can be made. This can sometimes lead to stunningly inaccurate figures that don't match the original source. Quoting a source helps avoid this, and, crucially, allows verifiability. Carcharoth 01:52, 3 September 2006 (UTC)[reply]

Not another policy please

This is yet more redundant instruction creep. There is no need whatsoever for yet another policy which is nothing more than an overdetailed regurgitation of other policies. If we have ever more policies the result won't be that the principles will be applied better, but simply that there are more pages that hardly anyone reads because it is all just too much. Keeping the number of policies down makes the key points easier to identify and apply. Sumahoy 00:02, 4 September 2006 (UTC)[reply]