There are passages here copied verbatim from source. Don't do that. Don't copy from sources and then "scuff up" the text, which doesn't avoid accusations of plagiarism. You're looking to use information from the sources by with entirely original presentation, from article organization all the way down through sentence structure to peculiar word choices and distinctive phrases. See the advice at WN:PILLARS#own, and the remarks at WN:Plagiarism.
Attributing information, correctly, to where information came from is key for both accuracy and neutrality. If we say, for example, that seven people were injured, rather than attributing that number to who said it, we are both giving the reader less information and treating that number as uncontroversial objective fact — unattributed information that turns out to be wrong then requires a {{correction}}, something we try very hard to not have reason to do (though it's also a point of pride that when we do make that sort of mistake we own up to it). The lede should attribute the numbers. A more troublesome accuracy problem arose in the second copy-from-source passage I happened on: the start of paragraph 3 ("Armenia's Health Minister, ..."; eighteen words long) differs in one spot from the source, by changing the number of injured from six to seven. Afaict, that change makes the thing unverified at the same time that it's substantially copied from the source: if there's any evidence in the sources that Armen Muradyan said the seven-injured thing in combination with the rest of that sentence, I didn't find it.
A stylistic principle to keep in mind: It's generally advisable to find ways to avoid building into the headline particular numbers that are apt to change very soon — such as the number of injured in a case like this, or the number of dead in a bigger disaster. For technical reasons we don't like to renaming an article after publication if we can avoid it (post-publish rename can produce double entries in some feeds).
The only links in the body of an article should be links either to Wikinews or to one of its sister projects ("wikilinks"), and those should use {{w}}. If you wish to help the reviewer by indicating which sources to look in for verification of particular details in the text, a good way to do that is an embedded html comment <!-- like this -->.
If possible, please address the above issues then resubmit the article for another review (by replacing {{tasks}} in the article with {{review}}). This talk page will be updated with subsequent reviews.
There are passages here copied verbatim from source. Don't do that. Don't copy from sources and then "scuff up" the text, which doesn't avoid accusations of plagiarism. You're looking to use information from the sources by with entirely original presentation, from article organization all the way down through sentence structure to peculiar word choices and distinctive phrases. See the advice at WN:PILLARS#own, and the remarks at WN:Plagiarism.
Attributing information, correctly, to where information came from is key for both accuracy and neutrality. If we say, for example, that seven people were injured, rather than attributing that number to who said it, we are both giving the reader less information and treating that number as uncontroversial objective fact — unattributed information that turns out to be wrong then requires a {{correction}}, something we try very hard to not have reason to do (though it's also a point of pride that when we do make that sort of mistake we own up to it). The lede should attribute the numbers. A more troublesome accuracy problem arose in the second copy-from-source passage I happened on: the start of paragraph 3 ("Armenia's Health Minister, ..."; eighteen words long) differs in one spot from the source, by changing the number of injured from six to seven. Afaict, that change makes the thing unverified at the same time that it's substantially copied from the source: if there's any evidence in the sources that Armen Muradyan said the seven-injured thing in combination with the rest of that sentence, I didn't find it.
A stylistic principle to keep in mind: It's generally advisable to find ways to avoid building into the headline particular numbers that are apt to change very soon — such as the number of injured in a case like this, or the number of dead in a bigger disaster. For technical reasons we don't like to renaming an article after publication if we can avoid it (post-publish rename can produce double entries in some feeds).
The only links in the body of an article should be links either to Wikinews or to one of its sister projects ("wikilinks"), and those should use {{w}}. If you wish to help the reviewer by indicating which sources to look in for verification of particular details in the text, a good way to do that is an embedded html comment <!-- like this -->.
If possible, please address the above issues then resubmit the article for another review (by replacing {{tasks}} in the article with {{review}}). This talk page will be updated with subsequent reviews.
Thanks for your help. I have edited my article and attributed all of my sources. Sorry, this is my first time using Wikinews and it's taking me some time to get used to it. I put everything in my own words and put it through a plagiarism checker, so hopefully it all works out! Please let me know if you need me to do anything else. Thanks!Tia UOW (talk) 15:28, 26 April 2016 (UTC)Reply
Ironically, when I did a preliminary scan for passages too-close-to-source, the one thing that turned up was the one passage I'd specifically mentioned in my review comments — unchanged, except for a bit of "scuffing up", just what I was saying does not help. I'll see what I can do about that during review; provided there's nothing much else problematic, it may be something I can handle without compromising my independence for review. --Pi zero (talk) 14:05, 27 April 2016 (UTC)Reply
Two other potential snags I've noticed, so far:
There appears to have been a significant further development in the story; as an independent reviewer, I basically can't add stuff, so the question becomes whether a credibly publishable article can be produced by cutting stuff out, without adding in the new development. Don't know yet.
Numbeo content is, I gather, crowdsourced. What that implies for its useability for this purpose, I'm considering.
I've had to edit this extensively enough that I really had to think carefully about whether I'd need to disqualify myself from publishing (which could easily result in the article not being published). Major categories of edits made:
Distance from source; you need to work on this — a lot of material wasn't verbatim from source but was still the same text with some scuffing up. Substituting synonyms and changing verb tenses doesn't avoid accusations of plagiarism (at least, not when there's enough text where that's all that was changed). Note that I don't rely on software to make calls on distance-from-source; I don't think it's possible to really detect this stuff without actually understanding what the text means (which makes it similar to translation between natural languages: really can't be done by a computer). I generally do a first-cut check for distance problems using a program that finds verbatim consecutive sequences (toollabs:dupdet), but those sequences often only suggest where to look for longer similar-but-not-verbatim sequences, and sometimes — as here — I'll find sequences during the detailed source check that I never found a hint of when using the tool. Note, the presence of these sorts of problems greatly increases difficulty of review; and it can also cause reviewer edits to add up unexpectedly.
Attributing to a news service should only be done if either it's being done to give them credit for an exclusive, or the news service is the weak point at which the reliability of the information is questionable. It's almost never necessary to say "According to the BBC, David Cameron said ...", because it's usually not plausible that BBC would report David Cameron had said something that he didn't actually say; though, again, if he said it specifically to BBC that's an exclusive and we should acknowledge their exclusive by mentioning that they're the ones he said it to. A further complication here was that the CTV source wasn't really CTV; it was syndicated from AP, so nothing in it would be attributed to CTV.
There was a bunch of background for which specialized sources had been introduced, that I cut because I was unable to verify the claimed background from those sources. Btw, the Assembly of Turkish American Associations thing wasn't about terrorism in Armenia, it was about terrorism allegedly by Armenians committed elsewhere in the world, as part of a Turkish revisionist Armenian-genocide-denial movement.
On consideration, I decided Numbeo could be cited legitimately so long as the reader is made aware that it's crowdsourced; we always want to empower readers by letting them know where information comes from, and also help them to be better information consumers by encouraging them to think about where information comes from.
The reviewed revision should automatically have been edited by removing {{Review}} and adding {{Publish}} at the bottom, and the edit sighted; if this did not happen, it may be done manually by a reviewer.
I've had to edit this extensively enough that I really had to think carefully about whether I'd need to disqualify myself from publishing (which could easily result in the article not being published). Major categories of edits made:
Distance from source; you need to work on this — a lot of material wasn't verbatim from source but was still the same text with some scuffing up. Substituting synonyms and changing verb tenses doesn't avoid accusations of plagiarism (at least, not when there's enough text where that's all that was changed). Note that I don't rely on software to make calls on distance-from-source; I don't think it's possible to really detect this stuff without actually understanding what the text means (which makes it similar to translation between natural languages: really can't be done by a computer). I generally do a first-cut check for distance problems using a program that finds verbatim consecutive sequences (toollabs:dupdet), but those sequences often only suggest where to look for longer similar-but-not-verbatim sequences, and sometimes — as here — I'll find sequences during the detailed source check that I never found a hint of when using the tool. Note, the presence of these sorts of problems greatly increases difficulty of review; and it can also cause reviewer edits to add up unexpectedly.
Attributing to a news service should only be done if either it's being done to give them credit for an exclusive, or the news service is the weak point at which the reliability of the information is questionable. It's almost never necessary to say "According to the BBC, David Cameron said ...", because it's usually not plausible that BBC would report David Cameron had said something that he didn't actually say; though, again, if he said it specifically to BBC that's an exclusive and we should acknowledge their exclusive by mentioning that they're the ones he said it to. A further complication here was that the CTV source wasn't really CTV; it was syndicated from AP, so nothing in it would be attributed to CTV.
There was a bunch of background for which specialized sources had been introduced, that I cut because I was unable to verify the claimed background from those sources. Btw, the Assembly of Turkish American Associations thing wasn't about terrorism in Armenia, it was about terrorism allegedly by Armenians committed elsewhere in the world, as part of a Turkish revisionist Armenian-genocide-denial movement.
On consideration, I decided Numbeo could be cited legitimately so long as the reader is made aware that it's crowdsourced; we always want to empower readers by letting them know where information comes from, and also help them to be better information consumers by encouraging them to think about where information comes from.
The reviewed revision should automatically have been edited by removing {{Review}} and adding {{Publish}} at the bottom, and the edit sighted; if this did not happen, it may be done manually by a reviewer.
