Research talk:Characterizing Wikipedia Reader Behaviour/Robustness across languages

From Meta, a Wikimedia project coordination wiki

Final results[edit]

@Strainu: @Tgr: @Amire80: @AWossink: @Antanana: @Lyzzy: @Shangkuanlc: @Whym: @Kaganer: @عباد ديرانية: @Satdeep Gill: @Racso: @Hasive: Please review my recent post to wikimedia-l and let me know if you have any question or comments: https://lists.wikimedia.org/pipermail/wikimedia-l/2018-September/091004.html

As I mentioned in the post, I expect all documentation to be completed by the end of October and at that time you can take the learning to your communities if you choose to. You should also feel free to sign up for a presentation by one of the researchers if that's of interest to you. Please consider all of these as an offer. Please don't feel obligated to act on them if you're busy with other things. --LZia (WMF) (talk) 00:48, 11 September 2018 (UTC)[reply]

Results[edit]

Hi all. Thanks a lot for your patience! We are ready to share with you the results of the first part of the analysis. We have put the plots with all languages included at Research:Characterizing Wikipedia Reader Behaviour/Robustness across languages#Results. Here is what we'd like to do next:

  • Create a one-page description of the survey specific to your language. You can take this one page to your community and translate it to your language if you wish.
  • Continue with the second part of the analysis where for some languages we can create more insights about the characteristics of the readers with specific information need, prior knowledge (familiarity), and motivation.

On your end, it would be great if you help us see any patterns in the data that are interesting to you or raise questions for you. This is a lot of data and the more eyes on it, and the more questions asked (even if we can't respond to all of them), we will be able to better make sense of it.

If you have any question or comment, please let me know. Otherwise, you will hear back from us hopefully soon with the one page information, and later, for the second part of the study. (We don't have a good sense as how long this second part will take.) --LZia (WMF) (talk) 19:00, 12 March 2018 (UTC)[reply]

@Strainu: @Tgr: @Amire80: @AWossink: @Antanana: @Lyzzy: @Shangkuanlc: @Whym: @Kaganer: @عباد ديرانية: @Satdeep Gill: @Racso: @Hasive: Please see above. Thanks! --LZia (WMF) (talk) 20:10, 12 March 2018 (UTC)[reply]

@LZia (WMF): "You can take this one page ..." - you mean entire page Research:Characterizing Wikipedia Reader Behaviour/Robustness across languages or "#Results" section only? Maybe will be better to extract this page into separate page and mark this for translation in Meta-wiki? And then propose to local communities links to (maybe partially) translated versions? --Kaganer (talk) 23:29, 12 March 2018 (UTC)[reply]
@Kaganer: I created the first draft of the one page I had in mind at Research:Characterizing Wikipedia Reader Behaviour/Robustness across languages/Communications. Can you review it and see if it can make sense for your audience? I don't know if this is the best way to transfer the information to your community, for example, and I don't know if a template is useful for you at all. I also just realized that the plots labels are all in English, so there will be more work for you all to translate them to your language. :/ I'm trying to figure out how I can help more to reduce the work for all of you. --LZia (WMF) (talk) 01:19, 15 March 2018 (UTC)[reply]
@LZia (WMF): thank you for providing us the template. I have created an announcement for the result of part one and post it on the village pump of Chinese Wikipedia. --Liang(WMTW) (talk) 17:21, 29 March 2018 (UTC)[reply]
@Shangkuanlc: Great. Thanks for doing it and letting me know. :) --LZia (WMF) (talk) 17:31, 29 March 2018 (UTC)[reply]

What languages will the survey run in?[edit]

Language Point of Contact Feasible? Documents to translate Comments
Arabic Yes Done
English LZia_(WMF) Yes Done
Hindi Atgomez Yes Done
Japanese Whym Yes Done
Spanish Racso Yes Done

Spanish[edit]

@3BRBS: I now know that you're quite passionate about research and results in non-English languages, thanks to WMCON and our conversations there. :) We want to do research in multiple languages, including in Spanish, to understand Wikipedia readers. Can you check my announcement to wikimedia-l and let me know if you're interested to help with it? A few people have helped with a good chunk of the translation work that is needed, but I'd like for us to have someone in the Spanish speaking community to be our point of contact for this research, someone I can go to if I have questions about Spanish language, someone who's excited to see the results for Spanish and spend some time with us digging deeper in the future if needed or can connect us to the right people. Are you that person? :) --LZia (WMF) (talk) 22:40, 27 April 2017 (UTC)[reply]

Hindi[edit]

@Atgomez: Can you find someone in the Hindi language community who is willing to be our point of contact for this survey? Having someone directly in the editor community would be really helpful. :) --LZia (WMF) (talk) 18:54, 3 May 2017 (UTC)[reply]

Japanese[edit]

@Whym: Can you help me with translating one more item? Please feel free to translate it directly on the page. --LZia (WMF) (talk) 01:28, 10 May 2017 (UTC)[reply]

Questions[edit]

Alice Wiegand. Good question and something the Legal team should give advice on. I've pinged them. If we don't get a response in time, let's go with Wikimedia. Otherwise, I'll ping you to ask for update. --LZia (WMF) (talk) 11:43, 26 April 2017 (UTC)[reply]
  • Hi Leila, nice to meet you at Wikimania. Any project of implementing the survey in French? Also, I am wondering if a small portion of sampling of articles to survey could take into account a thematic parameter. For example, I am currently working on a pilot research project about the franco-ontarian community of Canada (how they are represented on Wikimedia projects, who are the editors working on their portal) and it could be interesting to add insights on readers, by sampling a few articles dealing with this community to learn more about the motivation and provenance of readers. Seeris (talk)
Hi Seeris. Nice meeting you, too, and thanks for following up. Regarding running the survey in French: At this point, the focus is on doing the analysis for the languages we ran the survey in, in June 2016. This being said, there may be other opportunities for running similar surveys in the future, and I'd love to somehow keep you informed about those. In the absence of any other central place to communicate, I suggest keeping an eye on wikimedia-l as we will announce this as a service on that list if we run more surveys. If you have a list of articles or topics you're interested in, please share a link to them and we can see if we can dig deeper in them as part of the analysis to provide some insights to you. I must say that I'm not very optimisitc that we've had enough requests from Canada (let's say to enwiki) who have participated in the survey and have gone to very specific pages/topics. Usually when you slice and dice the data with a few dimensions like this, you get very few data to be able to say something statistically significant, but we can always check. :) I also must say: the analysis will take in the order of couple of months on our end, and I don't want us to block you from your research. Please consider this as parallel information that can help you at some point, if we have enough data. :) Thank you again for your input during Wikimania and the follow-up. --LZia (WMF) (talk) 23:39, 25 August 2017 (UTC)[reply]

What other languages to run the surveys in?[edit]

Language Point of Contact First part? Second part? Documents to translate Comments
Bengali Hasive Yes No Done
Chinese Liang (WMTW) Yes No Done
Dutch AWossink Yes Likely No Done
German Lyzzy Yes Yes
  • the first item directly translated in the survey
  • survey
  • "Beantworten Sie drei Fragen und helfen Sie so mit, Wikipedia zu verbessern."
  • privacy statement.
 Done
Hebrew Amire80 Yes Maybe
  • הסקר "למה אתה קורא את הערך הזה היום?" אינו מקבל תשובות עכשיו. אפשר לקרוא עוד על המחקר מאחורי הסקר ועל התוצאות של הסקר ושל סקרים קודמים, כשהן תהיינה זמינות, בכתובת: ...
  • survey
  • אנא ענה על שלוש שאלות כדי לעזור לנו לשפר את ויקיפדיה.
  • privacy statement.
 Done
Hungarian Tgr Yes Maybe Done
Konkani Fredericknoronha No No
Ukrainian アンタナナ Yes Likely No
  • 'Опитування «Чому Ви читаєте цю статтю сьогодні?» не проводиться зараз. Ви можете прочитати більше про дослідження у рамках цього опитування, а також результати цього та інших опитувань, коли вони будуть доступними, за: ...'
  • Опитування
  • "Дайте відповідь на три запитання і допоможіть нам поліпшити Вікіпедію"
  • Заява про конфіденційність
 Done
Romanian Strainu Yes Maybe Done
Russian Amire80 Yes Likely Yes
  • 'Сбор ответов в рамках опроса «Почему вы читаете эту статью сегодня?» сейчас не ведётся. Об исследовании, в рамках которого проводится этот опрос, а также о результатах этого и других исследований, когда они будут доступны, вы можете прочесть здесь: ...'
  • Опрос
  • «Ответьте на три вопроса и помогите нам улучшить Википедию.»
  • Заявление о конфиденциальности
 Done

Hebrew[edit]

@Amire80: Hebrew and Russian are ready for you. For the privacy statement, feel free to use /language_code and translate there directly. The other three bullet points should ideally be translated by one person, and I will have to have a 20-30 min voice-chat with this person to make sure the meanings are conserved in the way the research intends them to be. For the two bullet points where the text to be translated is directly in the table, please replace the English text with the corresponding local language. If you have questions, just ping. I won't be constantly available but will definitely check here 1-2 times a day over the weekend. Thanks! :) --LZia (WMF) (talk) 18:09, 22 April 2017 (UTC)[reply]

@Amire80: Thanks for the translations for Hebrew and walking me through them. I've just marked Hebrew as done. There are some things we need to do on our end to make it happen in Hebrew (sampling rates, testing IDs, etc.) but we've got all the help we could wish for from you. Thanks for that. :) I will ping you a few days before the survey goes live to ask you to communicate it to your community. --LZia (WMF) (talk) 20:52, 25 April 2017 (UTC)[reply]

German[edit]

@Lyzzy: German Wikipedia has enough traffic for us to be able to do full analysis. As you may know, some of our collaborators are in GESIS and I'm sure they'd really like to contribute to the understanding of German Wikipedia readers. :) I've kept you as a point of contact for this request. If another username should be replaced, please go ahead and edit the table. In this specific case, and given that we have expertise on board, I'd like to ask that you only have the privacy statement page translated. One of the researchers on the team who is a native German speaker can translate the rest of the materials. Please let me know if this works. --LZia (WMF) (talk) 18:23, 22 April 2017 (UTC)[reply]

@LZia (WMF):, that sounds great, thanks. Will send you a link to the translation soon. 19:05, 23 April 2017 (UTC)[reply]
Sounds good. @Lyzzy:, please feel free to put the translation of the privacy statement directly at https://wikimediafoundation.org/wiki/Survey_Privacy_Statement_for_Schema_Revision_15266417/de . --LZia (WMF) (talk) 22:48, 23 April 2017 (UTC)[reply]
Done. Alice Wiegand (talk) 09:57, 26 April 2017 (UTC)[reply]
Thanks, Alice Wiegand. Just a note that we now know that "Wikimedia" should be replaced by "Wikimedia Foundation" in the part you had question about. That will be fixed in the English version. Please update it in the German one if you can as well. Also, I will get back to you a few days before the surveys go out to ask for your help to letting your community know about the survey launch date and any other information available and useful for them. Thanks! :) --LZia (WMF) (talk) 17:14, 26 April 2017 (UTC)[reply]

Ukrainian[edit]

@Antanana: The unique device counts for Ukrainian Wikipedia (to both mobile and desktop sites) is high enough that we can at least do the first part of the analysis (We will be able to generate Figure 2 in the paper) but in order to do the second part of the research which is subgroup analysis (which enables us to have results such as the ones in Table 3 of the paper that tell us more about groups such as those bored or those who use Wikipedia for work/school) we need to set the sampling rate quite high (the more exact sampling rate I can share with you once we fix all languages). This means a lot of the readers of your language will be prompted to participate in the survey and I want to make sure you are comfortable with this. My recommendation is that we go ahead and give it a try given that the kind of insights it can create for your language are potentially important enough to justify the high sampling rate, and given that realistically we have one shot to do this kind of comparative research in the near future. Let me know what you decide to do and we will take it from there. :) --LZia (WMF) (talk) 18:51, 22 April 2017 (UTC)[reply]

@LZia (WMF): thank you! yeah, let's do the first part of the analysis. I'll start the discussion with the Ukrainian community about the rest of it, is it okay with you? --アンタナナ 17:53, 23 April 2017 (UTC)[reply]
Sounds great, アンタナナ. :) I just updated the above table with the items which will need translation. Ideally, we want one person to translate everything, if this is not possible, at least everything other than the privacy statement should be translated by one person. Please let me know who that person is and I will give them access to the survey file. Also, I will need to have a 20-30 min voice-chat with this person in the coming week, after they do the translations to make sure research intentions are preserved in translation. :) Thanks for your help. --LZia (WMF) (talk) 18:18, 23 April 2017 (UTC)[reply]
@Antanana: we checked the sampling rates today. We are making bunch of assumptions for computing sampling rates (for example, that the response rate to the survey in Ukrainian will be similar to English, which may or may not be true but that's the best data we have). Based on our estimates, no matter how high the sampling rate, if we run the survey in Ukrainian Wikipedia for a week, we won't have enough responses to be able to do the second part of the research. However, we will be able to do the first part of the research. All I need from you at this point is to complete all translations and then let me know a 45-min time slot that we can walk over the survey translations together to make sure everything is correct. :) --LZia (WMF) (talk) 15:23, 26 April 2017 (UTC)[reply]
@Antanana: we're moving fast to finalize all surveys that will go live (ideally by Wednesday this week but if we miss the train on Monday, May 8. Can you give us an update about Ukrainian? Are you still interested and if yes, when do you think you will have the translations ready? We still need a 45-min call with whoever translates the survey and myself, so let's make sure we give it enough time. Thanks! :) --LZia (WMF) (talk) 19:51, 30 April 2017 (UTC)[reply]

Bengali[edit]

@Hasive: Thanks for the ping re Bengali. I looked up some numbers and I need some time to talk to the rest of the researchers on Monday/Tuesday to see if we can make it work in Bengali. I think we can run it and at least get results similar to Figure 2 in the paper. Doing more than that may not be feasible. Again, I'll update you once I know more. --LZia (WMF) (talk) 18:56, 22 April 2017 (UTC)[reply]

Hi LZia (WMF) Thanks! Wait your update. -Hasive • talk • 10:17, 23 April 2017 (UTC)[reply]
@Hasive: We just had a meeting and went over the sampling rates. Bengali is border-line in the sense that there is a chance that we run it for a week and we can't get enough data even for the first part of the study. This being said, given that it's borderline, I'm happy for us to run the survey for you and give it a try, maybe we get lucky and enough people respond. Of course, whether this is worth your time given the effort needed for translation is your call. So, please don't feel obligated but if you decide to do it, we're here to help you make it happen. --LZia (WMF) (talk) 14:26, 26 April 2017 (UTC)[reply]
@Hasive: Can you let me know if you are still interested to run the surveys in Bengali anyway? If yes, we should move fast and I'm happy to open up time to help you for that. If not, please let me know so we don't wait for Bengali. Thanks! :) --LZia (WMF) (talk) 19:48, 30 April 2017 (UTC)[reply]
Hi @LZia (WMF): Sorry to late! Yap, I'm interested to run the surveys for Bengali. Please let me know the details. -Hasive • talk • 05:14, 1 May 2017 (UTC)[reply]
@Hasive: no worries. I've added the items that need translation as bullet points to the table above. If you can translate them today in your time-zone and we can have a meeting on Tuesday to go over them together, that'd be great. Timezones will be tricky but I'll do my best to be as flexible as possible. One note about the survey in the Google Form: try to keep the translated text as close as possible to the English version. We want to be able to compare the results across languages, and the less differences in the translations themselves, the easier for us to do the analysis and infer results later. :) Thanks for your help in advance. --LZia (WMF) (talk) 05:54, 1 May 2017 (UTC)[reply]
@Hasive: is the privacy statement for Bengali translated? How about the survey itself, linked from the table above? I don't seem to find the translations. Once you update the translations and they are completed on your end, please email me with a couple of time slots that work for you for a 45-min meeting to walk over the translations together. We are aiming to send the surveys out on Monday, May 15, fyi. :) --LZia (WMF) (talk) 16:03, 10 May 2017 (UTC)[reply]
@Hasive: this is the final ping that we are planning to launch the surveys this week and I need all the translations in by Tuesday end of the day UTC to be able to include Bengali in the survey. Please let me know if you can make it. Thanks. --LZia (WMF) (talk) 10:20, 19 June 2017 (UTC)[reply]
@LZia (WMF): I already completed all translation. -Hasive • talk • 10:11, 13 March 2018 (UTC)[reply]

Konkani[edit]

@Fredericknoronha: Thanks for pinging us about Konkani. :) I looked into the unique device count to Konkani Wikipedia and it is very low, even if we show the survey to all readers for a week, it won't give us enough data to be able to have statistically significant results. This being said, let me talk to the rest of the team on Monday/Tuesday and update you here with a final response. --LZia (WMF) (talk) 20:05, 22 April 2017 (UTC)[reply]

@Fredericknoronha: the team just met and I can not confirm that we won't have enough data points in Konkani if we run the survey for one week and event if we show it to the all readers of the language. I'm sorry that I don't have a better news now :(, but I do look forward to studying Konkani and share results with you when there is more traffic to the language. --LZia (WMF) (talk) 14:12, 26 April 2017 (UTC)[reply]

Dutch[edit]

@AWossink: can you go back to the survey and also translate "Confirmation Page" in the bottom of the survey? The rest looks good to me. Thanks! :) --LZia (WMF) (talk) 19:59, 24 April 2017 (UTC)[reply]

@AWossink: and @Effeietsanders: thanks a lot for your help with the translations and for going with me over them. :) I will get back to you no later than 48 hours before the launch to ask for your help with informing your community about the launch of the survey via Village pump or other appropriate channels. --LZia (WMF) (talk) 11:36, 26 April 2017 (UTC)[reply]

Chinese[edit]

Hi@LZia (WMF): thank you for putting the yes in Chinese, I am starting the translation process. I am wondering because Chinese Wikipedia is serving people using Tradtional Chinese and also Simplified Chinese, do you think this survey should turn on the geo located site-notice? In that case, we can redirect the people who use Taiwan and Hong Kong IP to Traditional Survey version and Mainland China IP to the simplified one.

@LZia (WMF):, I have completed the translation of survey in Traditional Chinese. However when I take a look at the link of the privacy statement in the section 2 of this survey, I can only find out the 2008 version on meta, the Privacy_policy/zh-hant seems to be deleted. It seems like a minor issue but just want to let you know it may cause some confusion when people look up the translation of the policy. Also since I don't have any right to edit foundation wiki, can I use my user page sandbox on meta to work on the Survey Privacy Statement for Schema Revision 15266417? Thanks. --Liang(WMTW) (talk) 18:07, 25 April 2017 (UTC)[reply]
Thanks, Liang(WMTW). Are you suggesting that you link this version from the survey Footer [1]? Re the translation of the privacy statement, sure, sandbox is fine. I can then copy/paste from there to /zh-hant if you don't mind. (I understand that it's not ideal that you cannot edit that page directly. sorry about that. :/) Also, I need to set up a time to walk with you through your translations. From the Dutch and Hebrew experience I can tell that my earlier estimate of 20-30 min was too low. It's more around 45-min. Can you send me some availabilities on your end via email and I'll schedule a Hangout/Skype call? I'll try to be as flexible as possible time-wise. :) --LZia (WMF) (talk) 00:24, 26 April 2017 (UTC)[reply]
@LZia (WMF):, yes, I am suggesting that we link this version to the survey Footer [1]. I agree with all your decision on other issues above, and my time can be seen here in this doodle. I have squeezed as many options as possible, so please come and select one with me, thanks. :) --Liang(WMTW) (talk) 08:05, 26 April 2017 (UTC)[reply]
Sounds good, re replacing footer [1] with the link you suggested. Please go ahead with it. I also noticed that there is one translation missing from the survey itself. Please translate "This is you survey ID. Please do not modify.". Thanks a lot for the availabilities as well. I'll send you a calendar invitation in the next few minutes. --LZia (WMF) (talk) 11:23, 26 April 2017 (UTC)[reply]
@Shangkuanlc: Thank you very much for the call today and working on the translations. It takes couple of iterations to get all the translations right and I'm grateful that you spent the extra time with me on it. Chinese (zh-hant) is now ready and I've marked it as done in the table. A few days before the surveys go live, I will contact you to ask you to notify your community about the launch of the survey. Looking forward to collecting the data for you and analyzing and sharing the results. :) --LZia (WMF) (talk) 02:19, 27 April 2017 (UTC)[reply]
Shangkuanlc, LZia (WMF): Small note about Privacy policy localizations: correct link to 'zh-hant' is //meta.wikimedia.org/w/index.php?title=Privacy_policy/zh&variant=zh-hant (maybe this "autoconverted" text is unclear and not completely accurate, but this is the technology that is working currently). All links to translated meta-pages was changed to using 'Special:MyLanguage'.--Kaganer (talk) 10:44, 27 April 2017 (UTC)[reply]
@Kaganer: Thank you! That explains my confusion. I just updated the link in footer [1] in the survey. --LZia (WMF) (talk) 20:53, 27 April 2017 (UTC)[reply]

@LZia (WMF):For Chinese Wikizpedia, please don't use google form. Google is blocked in China. From our past experience, please use [1] instead. You may ask User:AVasanth (WMF) for help if needed. Also, for simplified/traditional chinese problem, you may add a question at first, then split the questionaire into two versions, or display both simplified/traditional chinese at the same time. --Temp3600 (talk) 11:14, 25 April 2017 (UTC)[reply]

@Temp3600: thanks for bringing these two points up. Regarding simplified Chinese, my understanding is that the community will need around 2 weeks for translations and we cannot wait for that, unfortunately. Regarding the survey form itself, this is an important point but something that changing it will require non-negligible work on our end. Here is why: In section 3.5 of the paper we describe how we use the connection between survey responses and webrequest logs to remove some of the bias in the responses. For this to happen, we have to have a way to pass a unique ID from the survey form back to our servers. Whether the service you have offered has such a feature is something we should look into. Also, our privacy statement will need to be updated to reflect this change of service (minor change, as far as I can tell though). I'm looking at the traffic to zh.m.wikipeida and zh.wikipedia, and most of the traffic is coming from countries other than China. It's of course problematic to not receive responses from China at all, but given the amount of work that will be created and the short time we have left to launch the surveys (we are aiming for Monday next week), I'd tell you that we will look into the service you suggested but I can't promise we will be able to switch to it for the Chinese survey. I hope we can find a solution but if we can't, we should go ahead as planned. --LZia (WMF) (talk) 16:24, 25 April 2017 (UTC)[reply]

Russian[edit]

Hi @LZia (WMF):,

For Russian, I completed the sentences in the table, and the survey in Google Forms (although we haven't verified it together yet). @Kaganer: completed the privacy policy, and also set up the language selector there. Thanks, Kaganer! --Amir E. Aharoni (talk) 14:08, 26 April 2017 (UTC)[reply]

Hi Amir E. Aharoni and Kaganer. Thank you so much for doing it and the ping. Amir, you have an invitation from me for a one-on-one meeting for us to go over Russian together tonight your time. It would be great if Kaganer can also attend (please feel free to add them to the meeting or let me know and I'll add them). Looking forward to finalizing Russian language with you in the next few hours. :) --LZia (WMF) (talk) 14:16, 26 April 2017 (UTC)[reply]
LZia (WMF), Russian text in table above is fixed now. --Kaganer (talk) 15:13, 26 April 2017 (UTC)[reply]
Great! Thanks, Kaganer. --LZia (WMF) (talk) 15:16, 26 April 2017 (UTC)[reply]
Amir E. Aharoni and Kaganer, thank you so much for the meeting and all the care you put for updating the survey. I really appreciate it as it makes us more confident on our analysis and the results of the survey once this research is done. Russian is ready to go, I've marked it as done in the table, and will ping you a few days before the launch of the survey to ask you to notify your community. --LZia (WMF) (talk) 21:30, 26 April 2017 (UTC)[reply]

@Kaganer: Amire80 mentioned that is better if someone from the Russian community be the point of contact between us and the Russian Wikipedia community. Can you help us with it? I expect no more than 3 hours time commitment in the next couple of months and the role will be around making sure our notifications get to ruwiki folks (one of them we've asked you to kindly consider doing in the phab task linked above), and also to make sure that the conversations that happen in ruwiki that we will need to be aware of are transferred to us. Please let me know. Thanks! --LZia (WMF) (talk) 20:05, 20 June 2017 (UTC)[reply]

@LZia (WMF): No problem. I don't expect any specifical discussions or conversations. --Kaganer (talk) 11:38, 21 June 2017 (UTC)[reply]
@Kaganer: thank you! I added you as a PoC then. :) --LZia (WMF) (talk) 20:44, 21 June 2017 (UTC)[reply]

Hungarian[edit]

@Tgr: when you're done with the translations for Hungarian, can you ping me for us to have 45-min call to go over your translations together? This can help us assure that the intents of the research are preserved in translation (as much as possible) which makes the work for comparing results across languages more reliable. :) --LZia (WMF) (talk) 02:23, 27 April 2017 (UTC)[reply]

@Tgr: thank you for the translations and for walking me through them. Hungarian is ready now. :) I will ping you a few days before the survey goes live to ask for your help for notifying your community. If there is anything in the mean time I can help with, please ping. --LZia (WMF) (talk) 17:04, 28 April 2017 (UTC)[reply]

Romanian[edit]

@Strainu: Thank you so much for doing the translations and also walking me through them. I appreciate the time and the care you've put in making this happen. :) I will get back to you with the next steps in the coming days, as I mentioned: there is a chance we can do some light tests in Romanian on Monday/Tuesday to test the infrastructure, adjust sampling rates if needed, etc. We are hoping that all survey can go out of the door no later than Monday, May 15. I will keep you posted about the community notices. Thanks again, and if there is anything I can do in the mean time, just ping! --LZia (WMF) (talk) 18:49, 3 May 2017 (UTC)[reply]

@Strainu: sorry, I forgot to ask for one more translation. Can you translate the message that links the user to the privacy statement from QuickSurvey itself as well? Feel free to do it directly on that page. thank you! --LZia (WMF) (talk) 17:03, 8 May 2017 (UTC)[reply]
@LZia (WMF): done.--Strainu (talk) 18:57, 8 May 2017 (UTC)[reply]
great. thanks! --LZia (WMF) (talk) 01:02, 9 May 2017 (UTC)[reply]

QuickSurvey widget on mobile[edit]

@Jdlrobson: I'm wondering: For the purposes of this study (and I'm guessing it's also relevant to some of the other work your team does with QuickSurveys) is it possible to show the QuickSurvey widget before the infobox on mobile? Quite a few languages that we will be studying in the upcoming reader research expansion have very high traffic to the mobile site (for some of them more than desktop) and the location of the widget on the page can create bias in the results that we cannot control for unless we move the widget before the infobox (more consistent with the Desktop experience). Can you let me know if this is possible and how much work it entails for you? --LZia (WMF) (talk) 15:55, 25 April 2017 (UTC)[reply]

Nope. I'm afraid currently we are very limited about where we can place a QuickSurvey. There are a couple of problems with the existing positioning as well - see https://phabricator.wikimedia.org/T115058 and https://phabricator.wikimedia.org/T118485 Jdlrobson (talk) 17:46, 26 April 2017 (UTC)[reply]
@Jdlrobson: hmm. This can be quite problematic in terms of our ability to collect enough data from these languages (and inconsistencies of course create other issues with the data). Let me call a meeting with you and Adam to see what can be done. --LZia (WMF) (talk) 18:37, 26 April 2017 (UTC)[reply]

Testing the sampling rates and data collection[edit]

In order to assess the accuracy of the sampling rates estimated for the surveys in each language as well as to test the data collection, we need to run tests in a few of the languages that have signed up to participate. All tests are scheduled for start time between 13:00-14:00 UTC on May 15 and stopping in 18:00-19:00 UTC on May 15. Below you can find the help we need from each of the volunteers who help with these languages. We ask all volunteers to translate the following message and share it with their communities.

"Hi all. We are preparing to run a survey in your language to learn more about Wikipedia readers across languages. The first part of this study on English Wikipedia was completed last year and we are expanding the research to other languages. More than 10 languages will participate as part of this second round of research. At this point, we are scheduling a ~6 hour test in your language at a sampling rate of 1 out of take_the_denominator_from_below to assess the accuracy of our estimated sampling rates as well as to make sure our data collection infrastructure is functioning correctly. The surveys are expected to launch on May 15. The task corresponding to this test is phab_task_number. Thanks! --LZia_(WMF)"

Japanese[edit]

@Whym: The sampling rate in Japanese Wikipedia will be 1 out of 10 requests. Can you translate the above message and share it with your community? Also, can you let me know if you want me to put the message in English anywhere in jawiki? We track this work at task T164894. Thanks! --LZia (WMF) (talk) 15:43, 10 May 2017 (UTC)[reply]

For archive happiness, this is done. Thanks, Whym. --LZia (WMF) (talk) 18:08, 14 May 2017 (UTC)[reply]

Romanian[edit]

@Strainu: The sampling rate in Romanian Wikipedia will be 1 out of 2 requests. Can you translate the above message and share it with your community? Also, can you let me know if you want me to put the message in English anywhere in rowiki? We track this work at task T164769. Thanks! --LZia (WMF) (talk) 15:43, 10 May 2017 (UTC)[reply]

Done.--Strainu (talk) 09:10, 11 May 2017 (UTC)[reply]
Thank you, Strainu. I tried to use the translation option to read the comment to your post and from what I could understand, it seems this study can be helpful for your community. :) --LZia (WMF) (talk) 00:32, 12 May 2017 (UTC)[reply]

German[edit]

@Lyzzy: The sampling rate in German Wikipedia will be 1 out of 10 requests. Can you translate the above message and share it with your community? Also, can you let me know if you want me to put the message in English anywhere in dewiki? We track this work at task T164960. Thanks! --LZia (WMF) (talk) 17:23, 10 May 2017 (UTC)[reply]

Hebrew[edit]

@Amire80: Please read the entry paragraph to this section. The sampling rate in Hebrew Wikipedia will be 1 out of 2 requests. Can you translate the above message and share it with your community? Also, can you let me know if you want me to put the message in English anywhere in hewiki? We track this work at task T164963. Thanks! --LZia (WMF) (talk) 17:46, 10 May 2017 (UTC)[reply]

Retrieved from "https://meta.wikimedia.org/w/index.php?title=Research_talk:Characterizing_Wikipedia_Reader_Behaviour/Robustness_across_languages&oldid=18371584"