Why Wikipedia should not be used as a single reference source

[ejstubbs]

12 hours ago, The Johnster said:

Wikipedia is what it is, and no decent research should ever be taken from a single source if possible.

 

This.  It's fundamental, and I don't quite understand why the OP decided to throw the spotlight on Wikipedia in particular.  Yes, it's popular and often cited but, as Johnster and others have said, there are a lot of other much less reliable and far more popular online sources than Wikipedia.  In particular, anything that it is posted on FB/Twitter-style social media should be treated with extreme caution until verified from a number of reliable sources of a less transient nature.

 

It should be possible to find out when and by whom the questionable passage was added to that article.  At the top of the article page there is a "history" link, which takes you to a list of all the changes to the page back to when it was first created.  For the article in question there are a lot.  Given the size of the 'travels of D2860' section it should be possible to spot when it was added by looking for a sudden large increase in the article's byte count.

 

However, there are quite a lot of other useful tools available in the article history part of Wikipedia which can help to make it easier to pin down things like this.  In the case of the Class 02 article I looked at the page statistics (linked off the history page) and found that a user called "N1TH Music" had contributed by far the most content to the article.  Clicking on that user's top edits for the article showed a significant dump of information into the article by that user in the autumn of 2021.  By checking the 'diff' links for each of their edits of a reasonable size it becomes clear that the questionable information about D2860's travels was added on 5th November that year: https://en.wikipedia.org/w/index.php?diff=1053680398  Since that user has contributed the vast majority of the information in that article, sadly the questionable accuracy of that edit rather undermines one's confidence in a lot of the other information in that article provided by that user*.

 

I note that the suspect information relating to D2860 is shown as having been removed at 22:12 on 7th September i.e. yesterday, by a user identified only by the IP address "80.1.163.170" - only for it to be reinstated two minutes later by a user called "Saintstephen000".  It was removed again at 01:15 on the 8th by a user named "LRV1007", and reinstated at 02:37, again by Saintstephen000.  Somewhat bizarrely IMO Saintstephen0000 then sent an incredibly patronising message to LRV1007 at 02:39 asking them, when making edits, to "make dam'd sure it's sourced" - which strikes me as being more than a little ironic in the context.  It wouldn't surprise me if LRV1007 had deleted their profile after receiving such an ill-judged put-down.

 

What the above shows is that, for all that may not be perfect about Wikipedia, one thing it does offer is a degree of traceability for the information presented in articles, together with the ability for people with better knowledge to correct inaccuracies in place (something sadly lacking in the FB/Twitter etc world), with changes being recorded and with oversight from both other informed contributors and the editorial staff**.  Dismissing it out of hand as being inherently untrustworthy is, to my mind, a classic case of the perfect being allowed to become the enemy of the good.

 

* I note that N1TH Music has contributed to a number of other locomotive and MU articles on Wikipedia.  One would hope that their 'research' behind those contributions has been rather more thoroughgoing then seems to have been the case for D2860's supposed UK-wide stravaigings.

 

** Not dissimilar in that respect to discussion forums like this one, one might argue.

Edited September 8, 2022 by ejstubbs

[Reorte]

Wikipedia is a good starting point. Yes, it needs to be taken with a pinch of salt but there's a great deal of very good information on there. Of course it's always sensible to check on any further details but there's sometimes an attitude that you shouldn't believe a word on it. No source will ever be completely reliable and Wikipedia is better than many. It's got lots of people checking it over, and although that can cut both ways a glance at the talk page is usually worth a look (at least for a reasonably popular article) to see if there's anything being argued over. Keep an eye on the language used, it's generally worth being suspicious of anything not well written or not written in the overall style.

[Steamport Southport]

14 hours ago, Northmoor said:

In an idle moment at work (my search history at work doesn't contain anything dodgy because it's rightly monitored, but some might find it weird), I was looking at this page today:

https://en.wikipedia.org/wiki/British_Rail_Class_02

 

I did enjoy the bit about the "travels" of D2860, preserved at the NRM, whoever added that entry was clearly on mushrooms at the time.  I think I'd have read more elsewhere about a 4-wheeled shunter travelling several thousand miles around the UK rail network. 

 

Universities teach students that when referencing research sources, you NEVER use Wikipedia as a single source.  This is why, and remember that some previously well-respected standard railway publications have been debunked in recent years.

 

There was one four wheeled shunter that travelled quite exensively on the BR network.

 

51218 was a bit of a nomad.

 

https://preservedbritishsteamlocomotives.com/51218-lyr-68-lms-11218-br-51218/

 

And yes, photographic evidence does exist of it at all those locations including Bristol, on the S&DJR and in South Wales.

 

 

Jason

[Steamport Southport]

Wikipedia is about as useful as you want it to be. I just use it as a basic guide, but some pages and subjects are far better than others.

 

For example I found this page pretty interesting and gives a good overview of Tyrannosaurs. But look how many sources and links there are.

 

https://en.wikipedia.org/wiki/Tyrannosaurus

 

 

 

Jason

[Pete the Elaner]

Wikipedia is the largest & therefore most comprehensive wiki available.

A wiki is: "a website or database developed collaboratively by a community of users, allowing any user to add and edit content."

Wikis are very useful resources for storing information which can be updated at any time. Anyone with an account can correct them (& in wikipedia's case, anybody can create their own account).

This does mean that anyone can add information which is deliberately misleading or, more often, information which they believe to be true, but may not be. This means that a wiki's biggest strength is also its biggest weakness.

The biggest problem with wikis is that people do not update them to correct any errors. I have been in several environments where some have asked others to update them. This misses the point: the whole idea is that if somebody sees something wrong or out of date, they should be able to correct it themselves.

[D9020 Nimbus]

I've encountered significant errors in the past in an article that appeared in a refereed academic journal (in computing) so nothing can be guaranteed accurate if there's only one source. 

[The Johnster]

4 hours ago, Steamport Southport said:

 

There was one four wheeled shunter that travelled quite exensively on the BR network.

 

51218 was a bit of a nomad.

 

https://preservedbritishsteamlocomotives.com/51218-lyr-68-lms-11218-br-51218/

 

And yes, photographic evidence does exist of it at all those locations including Bristol, on the S&DJR and in South Wales.

 

 

Jason

When it was transferred to Danygraig (from Manchester ISTR), it came down under it’s own steam via the Central Wales line, and took several weeks, delayed by hot axleboxes.   It would have been much easier to load it onto a crocodile.  It never went north again in BR service. 

Edited September 8, 2022 by The Johnster

[GDR]

If I want to know anything about railways I dredge through here first. 

 

As for Wrongepedia I once, ok after a bit of a session, looked up the village I came from and added to the "famous residents" bit. So I made myself "a successful Formula 3 driver who never quite made it to F1" while several of my then friends became triallist for Liverpool/respected brain surgeon/convicted member of IRA. All that crap is still on there. 

 

So yes - always recheck anything on Wiki!!

[GDR]

Go on .... someone create a Wiki entry about the famous LNWR Webb 9-wheel shunter...

[njee20]

Wikipedia is an amazing resource, whether wholly reliable in isolation or not. Intentionally adding nonsense to it seems pretty pointless to me.

 

It's just like Abraham Lincoln said; you can't believe everything you read on the internet.

[The Johnster]

No source is reliable unless it can be cross-referenced, especially your memory and other eye-witness accounts, and even then it is important to satisfy yourself that the cross-referred sources do not have a common origin.  Taken on those terms, Wikipedia is not a bad general source of infomation.  

[melmerby]

Information of any sort cannot be 100% guaranteed accurate, no matter how many sources are referenced unless one has the original documentation/photographic evidence etc. pertaining to it.

Even then some original information might have a human error that's crept in it that doesn't match the real information. e.g. un spotted mistaken key press on typewriter or someone writes down a wrong number.

 

Someoneone saying "I saw X at Y location on such and such a date" cannot be relied on because no human brain is 100% perfect at memory recall.

 

A good example of inadvertent errors creeping in, was in Fake or Fortune last week where the commitee that decides on what goes in the definitive catalogue resume on an artist rejected a painting as non authentic and in their description call it an "Oil on Canvas" when everyone could see it was actually on a wooden panel!

In years to come it might be assumed that there are two paintings of the same subject, one on wood & one on canvas

 

I would suggest, start with Wikipedia, then expand out to as many sources as possible to corroborate what's there and then the information gathered is correct as far as can be determined.

 

[Northmoor]

9 hours ago, ejstubbs said:

 

This.  It's fundamental, and I don't quite understand why the OP decided to throw the spotlight on Wikipedia in particular.  Yes, it's popular and often cited but, as Johnster and others have said, there are a lot of other much less reliable and far more popular online sources than Wikipedia.  In particular, anything that it is posted on FB/Twitter-style social media should be treated with extreme caution until verified from a number of reliable sources of a less transient nature.

 

It should be possible to find out when and by whom the questionable passage was added to that article.  At the top of the article page there is a "history" link, which takes you to a list of all the changes to the page back to when it was first created.  For the article in question there are a lot.  Given the size of the 'travels of D2860' section it should be possible to spot when it was added by looking for a sudden large increase in the article's byte count.

 

However, there are quite a lot of other useful tools available in the article history part of Wikipedia which can help to make it easier to pin down things like this.  In the case of the Class 02 article I looked at the page statistics (linked off the history page) and found that a user called "N1TH Music" had contributed by far the most content to the article.  Clicking on that user's top edits for the article showed a significant dump of information into the article by that user in the autumn of 2021.  By checking the 'diff' links for each of their edits of a reasonable size it becomes clear that the questionable information about D2860's travels was added on 5th November that year: https://en.wikipedia.org/w/index.php?diff=1053680398  Since that user has contributed the vast majority of the information in that article, sadly the questionable accuracy of that edit rather undermines one's confidence in a lot of the other information in that article provided by that user*.

 

I note that the suspect information relating to D2860 is shown as having been removed at 22:12 on 7th September i.e. yesterday, by a user identified only by the IP address "80.1.163.170" - only for it to be reinstated two minutes later by a user called "Saintstephen000".  It was removed again at 01:15 on the 8th by a user named "LRV1007", and reinstated at 02:37, again by Saintstephen000.  Somewhat bizarrely IMO Saintstephen0000 then sent an incredibly patronising message to LRV1007 at 02:39 asking them, when making edits, to "make dam'd sure it's sourced" - which strikes me as being more than a little ironic in the context.  It wouldn't surprise me if LRV1007 had deleted their profile after receiving such an ill-judged put-down.

 

What the above shows is that, for all that may not be perfect about Wikipedia, one thing it does offer is a degree of traceability for the information presented in articles, together with the ability for people with better knowledge to correct inaccuracies in place (something sadly lacking in the FB/Twitter etc world), with changes being recorded and with oversight from both other informed contributors and the editorial staff**.  Dismissing it out of hand as being inherently untrustworthy is, to my mind, a classic case of the perfect being allowed to become the enemy of the good.

 

* I note that N1TH Music has contributed to a number of other locomotive and MU articles on Wikipedia.  One would hope that their 'research' behind those contributions has been rather more thoroughgoing then seems to have been the case for D2860's supposed UK-wide stravaigings.

 

** Not dissimilar in that respect to discussion forums like this one, one might argue.

 

A very useful guide on analysing Wikipedia, thank you.  I knew you could delve into the edit records but have never tried.

 

To respond to two of your comments (highlighted), I highlighted Wikipedia as it was where I have found an example of hilarious fiction.  If I'd found it on any other website, I'd have highlighted that.  As for the second comment, nowhere in the title or body of my original message have I said Wikipedia is inherently untrustworthy; in fact I frequently refer to it, but accept that if I want definitive and detailed information, I'll need to look elsewhere as well.

 

[Pacific231G]

On 07/09/2022 at 23:21, pH said:

I quite often use Wikipedia as a first source on a subject I know little about. The ‘References’ in an entry can be useful for further research. Or the contents of the entry itself can suggest other places to look for more information. 
 

(Never mind Wikipedia, there are several well-known authors of books and articles on railway subjects that I would never use as a single reference source. In fact, some I would just not use as a reference source at all!)

It can be a little alarming to find, as I have a couple of times, Wkipedia entries that seemed to contradict something I'd thought to be the case, only on checking the references to find that most of them were to articles I'd written several years before that had been misquoted or otherwise mangled, along with one or to others to sources that would come up high on a Google search but that I knew to be inaccurate. It turned out that there was a misguided but well meaning soul who'd taken it upon himself to simply precis or even take chunks from (with no regard for copyright) a large number of other people's published articles articles and put them on Wikipedia. Fortunately, I was able to edit them to a closer resemblance to my actual (and well researched from primary sources) articles.  

However and despite that I do find Wikipedia a very useful first or casual source for many subjects but, for any halfway serious research, you do need to delve a lot deeper and be very wary of content that is "common knowledge" so entered without checking. 

 

Edited September 9, 2022 by Pacific231G

[Nearholmer]

4 hours ago, melmerby said:

I would suggest, start with Wikipedia, then expand out to as many sources as possible to corroborate what's there and then the information gathered is correct as far as can be determined.

Agree. It’s no bad place to start, if you don’t have a solid book on whatever the subject is to hand, but it does contain some whopper errors in odd corners (and loads of ‘editing wars’ in controversial areas like politics and international relations).

 

Another worthwhile thing to do, if you can read other languages, is to look at Wikipedia in more than one. I can just about, very slowly, and checking lots in a dictionary, manage engineering history subjects in French and German, and I’ve several times found that they differ, with one being far better referenced, and more accurate, than the others.

 

Even solid books, well researched and referenced, need a bit of care though, not so much around errors as around omissions. The classic problem in engineering history is authors writing well about there own country, but completely overlooking what was happening in other countries, even though people at the time would have been aware of it. Iffy stuff, because it tends to underplay the importance of the transfer of ideas and knowledge between countries.

 

 

[BR60103]

A writer I know (Fantasy novels) tried to edit her own Wiki entry to correct a few misleading errors. It wasn't allowed as she didn't have any references to cite.

 

[jjb1970]

4 hours ago, BR60103 said:

A writer I know (Fantasy novels) tried to edit her own Wiki entry to correct a few misleading errors. It wasn't allowed as she didn't have any references to cite.

 

 

I know people in a similar position, also I've known people who have references which are not accepted for some reason. 

 

I'm not against Wikipedia but it's only a starting point on stuff I don't know much about, it's very useful for a quick primer and to look for a few starting points on further reading. If it's casual reading and entertainment it doesn't matter much anyway.

 

The key is to separate anything contentious from bread and butter encyclopaedia type stuff. For the bread and butter stuff I find it quite good most of the time and I think most errors are genuine. When it comes to the sort of subjects which elicit strong opinions or related to contemporary politics then I'd avoid Wikipedia like the plague. Then again, on contentious subjects it doesn't matter what you do, it comes down to balancing different sources, doing your own reading and applying your own judgement.

[Bernard Lamb]

16 hours ago, GDR said:

If I want to know anything about railways I dredge through here first. 

 

As for Wrongepedia I once, ok after a bit of a session, looked up the village I came from and added to the "famous residents" bit. So I made myself "a successful Formula 3 driver who never quite made it to F1" while several of my then friends became triallist for Liverpool/respected brain surgeon/convicted member of IRA. All that crap is still on there. 

 

So yes - always recheck anything on Wiki!!

There seem to be several places where people play that game.. My old school being one of them. I know several of the people who are included and they seem to remain on the favoured list, but other names are added and or removed for no apparent reason. This seems to happen frequently with TV presenters, of whom there have been rather a lot, who have supporters who play a game of maintaining the image of the person that they support.

 

I tend to use Wickipedia for a first look and then follow up with original documents. The danger with Wickipedia is it is only as good as the people who contribute. With military people for example it is very uneven as many people are not included when looking at a particular regiment or battle while others who played a lesser part are the subject of long articles.

Bernard 

[BR traction instructor]

Any reference material is only as accurate as the creator(s) make(s) it. Over time you get to know which are the most reliable sources that you can rely on, otherwise, cross checking various sources is the only way to validate information. This applies equally to reference books, websites and any other medium.

 

BeRTIe

Edited September 9, 2022 by BR traction instructor

[kevinlms]

1 hour ago, Bernard Lamb said:

 

 

I tend to use Wickipedia for a first look and then follow up with original documents.

 

So what do you do if the 'original documents' are unavailable for research purposes? Perhaps they either don't exist or are in fact locked away in museums and the like, where humble members of the public can't access them.

[jjb1970]

29 minutes ago, BR traction instructor said:

Any reference material is only as accurate as the creator(s) make it. Over time you get to know which are the most reliable sources that you can rely on, otherwise, cross checking various sources is the only way to validate information. This applies equally to reference books, websites and any other medium.

 

BeRTIe

 

The one people are often far too accepting of is technical papers. Far too many just assume that if something is in a technical paper it must be right. I am pretty much paid to read papers and some of them are rubbish. There's a very devious trick you see quite a lot in that the abstract is not consistent with the body content. A lot of authors know most people only read the abstract and are clever at hiding inconvenient stuff whilst still being able to point to it if challenged. Alternately, the media is very good at taking paragraphs completely out of context. Another devious trick is reporting modelled results without explaining the model, that amounts to 'trust me'. And my favourite (no, my biggest hate), conflating modelled values with measured data. If the subject of a paper is important then read the full paper. Rant off, sorry 

[Pacific231G]

1 hour ago, kevinlms said:

So what do you do if the 'original documents' are unavailable for research purposes? Perhaps they either don't exist or are in fact locked away in museums and the like, where humble members of the public can't access them.

It is surprising how much primary source material is available online but it often takes a lot more than a simple Google (or Bing etc) search to find them. For any kind of serious research,  museums and the like are generally be very cooperative: curators want the information in their collections to be used.  If you're writing something for publication, whether print or digital, then your research should of course be serious or if not, or where reliable original sources are not available, couched in suitably qualified terms such as "According to so and so....." rather than stated as fact. There are a lot of "well- known facts" that are anything but. 

 

Proper research does take time and effort and there is of course information that is lost to history forever. Having a works drawing of say a special wagon doesn't guarantee that it was actually built in exactly that way and, if no photographs are available, then there may never be a way of knowing.

 

Sometimes a little lateral thinking will help.  Accurate colours are an obvious example given that most photographs were monochrome until the 1960s. Even colour photos can't be relied upon as lighting, film stock and image ageing make a great deal of difference.    However, artists painting a scene from life may not get the details right but may well get the colours right so looking at a number of examples from different artists of say a livery may be a good guide. I would though never, for example, give much credibility to a colourised postcard as that will probably have been done in the studio with no reference to the actual scene. 

 

It's also worth looking for less obvious evidence. A few years ago I wrote some articles about the Gare de Vincennes (aka Gare de la Bastille) terminus in Paris. What I couldn't be sure of was whether when it first opened in 1859 it was set out, as in the original architect's drawings, with a separate departure and arrival platform with a couple of carriage sidings between them or with several platforms as every later plan or image showed. The documents that would have answered this had been lost in a fire during the Paris Commune so there seemed no way of finding out. However, a delve in the online site of the French National Library yielded a newspaper account of the station's official opening  in which the journalist who'd travelled on the inaugaral train described boarding it from the departure platform on the left hand side of the train shed with carriages stored on sidings between it and the arrival platform on the other side. 

Edited September 9, 2022 by Pacific231G

[jjb1970]

A hidden risk in some fields is the false consensus. As an example, about 12 years ago a government agency paid my then employer to do an assessment of a new device based on a completely different principle of operation for a mandatory safety critical application. I was given the project which was in two parts, replicate the laboratory experiments supporting the new heresy and to do a literature review of the existing orthodoxy which was supported by a mountain of papers and documented analysis. To cut a long story short, we replicated the laboratory experiments and got the same (well, equivalent, there was some variation as you'd expect) results. For the second part, what quite quickly became apparent was that for all that it looked like the existing theory was based on a mountain of evidence, every single paper we reviewed (literally, every one) was based on a set of experiments performed in the 1950's in Britain and never seriously questioned. Every one accepted a set of 60 year old lab findings at face value and then just reinterpreted them to arrive at the same conclusions as the original research team. So what appeared to be a wide consensus based on a large body of work was actually a lot of people regurgitating the same primary research on an assumption the findings were valid and remained so.

There was an interesting tail end to it. At an IMO meeting the national delegation of the manufacturer (not the one which engaged my employer) claimed the theory underpinning the device had been verified by class society x (i.e. me). I quickly e-mailed our own fearless bench and bless them they intervened to correct this false statement, class society x had replicated the experimental results submitted in support of the theory, it had made no comment on whether the theory itself was correct. The two concepts are completely different, a set of lab tests could support any number of theories and I had quite deliberately offered no opinion at all on that question, which was stated very clearly in the report.

Always read these things carefully.

[ejstubbs]

1 hour ago, Pacific231G said:

For any kind of serious research,  museums and the like are generally be very cooperative: curators want the information in their collections to be used.

 

I can remember way back when I was a kid, the staff at the museum in the castle at Haverfordwest being extremely helpful digging through their archives for documents relating to Brunel's plans for railways in Pembrokeshire, including the 1845 planned line through Treffgarne Gorge to Goodwick, and an initial survey for the alternative route to Abermawr proposed in 1847, the latter complete with IKB's signature (work on both routes eventually being abandoned due to the Irish potato famine).  Potentially priceless stuff to which they were happy to provide supervised access simply because I was interested in the subject.

 

These days I am a member of the National Library of Scotland, which gives me access to a lot of online information sources, many of which would otherwise require a subscription of some kind (one of my regular haunts is the complete OED), and the ability to request access in their reading room to items from its copyright library  (which reminds me, I do need to renew my reading room ticket).  They have an online system for requesting material from their catalogue, and for booking reading room timeslots to view it.  You turn up at the scheduled time and the book, or whatever, is at the desk ready for you to peruse.  You're even allowed to take a "reasonable" number of photos with your mobile phone of the information you're after.  It's all free of charge, and I don't have to provide any credentials or justification for this, just "private research".

 

When this stuff is owned by public institutions, it's difficult to deny the public reasonable access.

[MyRule1]

The finest sources of misinformation can be family histories. Every family has their own cherished stories about their forebears and fairly often these stories turn out not to be 100% accurate. In both my, and my wife's, case research into our respective family trees has found primary sources that disput long held beliefs. We also have found one example of the opposite where the official primary records don't reflect the reality. In both cases we have found family trees on the main genealogical websites that reflect the untruths.

 

As thhis is RMweb my last example is railway related, I can guarantee that at least once a year I will be told by a visitor to our museum that "my father/grandfather drove the Flying Scotsman".