Group: IPERNITY - Ambassador Images
The assessment procedure (Tools)
By
Bergfex
club
|
1. Specification
The assessment of the proposed pictures should make as little work as possible for the administrators as well as for the jury. Nevertheless, it should deliver a well-differentiated, solid result - even with only a few jury members.
2. Proprietary solution
Ipernity offers the possibility to rate images with asterisks. This simple solution would do the least work. Unfortunately, it is useless because the proposed pictures are already pre-rated by other members. Other tools are not offered by ipernity. One would have to painstakingly improvise them, which would be labour-intensive for all involved.
3. Third-party services
3a. An attempt was made to perform the voting with the help of Discord. In terms of preparing, this is acceptable because the pictures can be linked directly from ipernity. However, counting is labour-intensive and error-prone. Moreover, DISCORD - like ipernity - only allows simple yes/no ratings. This leads to little differentiated results with few judges. More differentiated solutions would have to be laboriously improvised, which would not be adequate.
3b. Therefore, we were looking for a service that specialised in online photo competitions. With Poll Unit we found a service that fully meets our requirements. You can find a detailed description of the options here:
English: pollunit.com/en/photo-contest
French: pollunit.com/fr/concours-photographique
German: pollunit.com/de/fotowettbewerb
The following possibilities are particularly interesting for us:
● Execution of more differentiated votings (exactly tailored to our needs).
● Anonymisation of the image suggestions (avoidance of prejudice).
● Randomisation of the image proposals (avoidance of systematic errors).
● Anonymisation of the evaluation (no mutual influence of the jurors).
● Automated evaluation (saving of work).
● Easy documentation of the results (enables transparency for the proposers too).
The PRO version of this service costs 3 euros per month by subscription. Tests were successful. It was subscribed for 1 year in order to gain further practical experience.
Disadvantage: The preparations are more time-consuming because no direct linking of the pictures is possible. The pictures must therefore first be downloaded from the user accounts and then uploaded to Poll Unit. The latter can be done very easily as a batch process via Poll Unit's upload wizard. But the download from ipernity causes problems. Many members have blocked this option. This is perfectly correct. It is not to be influenced.
Solution: A general technical release of the download for the group administrators would require an intervention in the ipernity software, as well as a change in the terms of use. However, taking screenshots is permitted because this is publicly accessible content. This method will therefore be used until further notice. The HD image quality is completely sufficient for the application purpose.
4. Further procedure
For the time being, jury voting should be done with Poll Unit. The ima team has agreed. Bergfex is to develop a workflow for taking the screenshots that saves as much work as possible. He should document it here in due course.
5. Assessment method
Until further notice, the jury will vote with up to 5 stars ⭐⭐⭐⭐⭐ per picture. This results in an average number of stars per picture. Abstentions will not be taken into account in determining this value.
Example: 9 jurors score (with one abstention and two absences due to holidays) as follows:
┃ ⭐⭐⭐ ┃ ⭐⭐ ┃ ⭐⭐⭐⭐ ┃▬ ┃ ⭐⭐⭐ ┃ ⭐⭐ ┃ ⭐⭐⭐ ┃ ▬ ┃ ▬ ┃
The result is: 17 / 6 = 2.83
After the end of a voting, the scores will be listed in rank order by Poll Unit. This list can be made public for the information of the submitters. This might enhance transparency and acceptance. Of course, the voting of the jurors should remain secret.
....................................................................................................................................
edit:
5a. The assessment method has been improved in the meantime. See: comment below
5b. Members of the editorial team can get a large display of the photos to be able to assess them better. To do so, simply click on the first picture. You can even get a full-screen view. There is a special icon for this at the bottom right of the display. You can use the arrow keys on the keyboard to scroll forwards or backwards. This way you can comfortably move through all the photos and vote without leaving the full screen mode.
5c. NOT awarding a picture is equivalent to abstaining.
5d. Members of the editorial team can sort the pictures according to their personal scoring. This way they can check their vote at any time and readjust it if necessary until the end of the assessment period.
The assessment of the proposed pictures should make as little work as possible for the administrators as well as for the jury. Nevertheless, it should deliver a well-differentiated, solid result - even with only a few jury members.
2. Proprietary solution
Ipernity offers the possibility to rate images with asterisks. This simple solution would do the least work. Unfortunately, it is useless because the proposed pictures are already pre-rated by other members. Other tools are not offered by ipernity. One would have to painstakingly improvise them, which would be labour-intensive for all involved.
3. Third-party services
3a. An attempt was made to perform the voting with the help of Discord. In terms of preparing, this is acceptable because the pictures can be linked directly from ipernity. However, counting is labour-intensive and error-prone. Moreover, DISCORD - like ipernity - only allows simple yes/no ratings. This leads to little differentiated results with few judges. More differentiated solutions would have to be laboriously improvised, which would not be adequate.
3b. Therefore, we were looking for a service that specialised in online photo competitions. With Poll Unit we found a service that fully meets our requirements. You can find a detailed description of the options here:
English: pollunit.com/en/photo-contest
French: pollunit.com/fr/concours-photographique
German: pollunit.com/de/fotowettbewerb
The following possibilities are particularly interesting for us:
● Execution of more differentiated votings (exactly tailored to our needs).
● Anonymisation of the image suggestions (avoidance of prejudice).
● Randomisation of the image proposals (avoidance of systematic errors).
● Anonymisation of the evaluation (no mutual influence of the jurors).
● Automated evaluation (saving of work).
● Easy documentation of the results (enables transparency for the proposers too).
The PRO version of this service costs 3 euros per month by subscription. Tests were successful. It was subscribed for 1 year in order to gain further practical experience.
4. Further procedure
For the time being, jury voting should be done with Poll Unit. The ima team has agreed. Bergfex is to develop a workflow for taking the screenshots that saves as much work as possible. He should document it here in due course.
5. Assessment method
Example: 9 jurors score (with one abstention and two absences due to holidays) as follows:
┃ ⭐⭐⭐ ┃ ⭐⭐ ┃ ⭐⭐⭐⭐ ┃
The result is: 17 / 6 = 2.83
After the end of a voting, the scores will be listed in rank order by Poll Unit. This list can be made public for the information of the submitters. This might enhance transparency and acceptance. Of course, the voting of the jurors should remain secret.
....................................................................................................................................
5a. The assessment method has been improved in the meantime. See: comment below
5b. Members of the editorial team can get a large display of the photos to be able to assess them better. To do so, simply click on the first picture. You can even get a full-screen view. There is a special icon for this at the bottom right of the display. You can use the arrow keys on the keyboard to scroll forwards or backwards. This way you can comfortably move through all the photos and vote without leaving the full screen mode.
5c. NOT awarding a picture is equivalent to abstaining.
5d. Members of the editorial team can sort the pictures according to their personal scoring. This way they can check their vote at any time and readjust it if necessary until the end of the assessment period.
The topic of this discussion has been edited by Bergfex 2 years ago.
You must be a member of this group to reply to this topic. (Join?)
Jump to top
RSS feed- Latest comments - Subscribe to the comment feed for this topic
- ipernity © 2007-2025
- Help & Contact
|
Club news
|
About ipernity
|
History |
ipernity Club & Prices |
Guide of good conduct
Donate | Group guidelines | Privacy policy | Terms of use | Statutes | In memoria -
Facebook
Twitter
Bergfex clubWhy this is so has never been found out. One assumption is that a picture slips through if two users suggest the same picture for the group at the same time. Since this effect is rare, it was not investigated further.
raingirl club has replied to Bergfex clubraingirl clubI did, however, like the annonymity of the voting! I don't care how other people vote (but can see how that might influence someone else in their voting), but not knowing the photographer's name was very appreciated.
The reasons I didn't like using it:
1) It was not obvious how to get to a full sized image,
2) The full sized image always has a rectangle over it at the bottom middle of the image - this was very distracting from the voting process, and
3) Not knowing how the star system was going to be used to make a decision, I erroneously left stars off of all the images that I though should not be in the group. I now realize that I should have placed a single star so that my vote would be counted.
Possible solutions
1) Notify jurors how to get to full sized image and that keyboard arrow key works even though the screen doesn't show an arrow to move through the images. [I did notify the current jurors via message.]
2) Use the one sized smaller image to make decision. Not the best solution as it's harder to check for focus etc, but the only one I can think of using PollUnit.
3) Decision should be made about how the stars (or lack of stars) will be used in voting, then all jury members should be notified before the next vote. Maybe a word per star could be noted, such as: 1 star = completely dislike (a thumbs down vote), 3 stars = could go either way, 5 stars = 100% think the photo should be included.
Bergfex club has replied to raingirl club1) I myself did not know about the possibility of displaying the image in full screen until I received your private e-mail a few days ago. I would be happy to add your comments to the discussion statement. (see edit)
2) Unfortunately, the smaller pictures cannot be displayed larger. The reason is the limitation by the photographers in their account settings. If one were to enlarge such size-limited images, one would not see a sharper image.
The only remedy would be to ask the photographers concerned to allow a larger display format in their account settings.
3) I will gladly take up the hint that NOT awarding a star is equivalent to abstaining. Furthermore, we can also choose any other form of voting method. Poll Uint offers a rich palette of possibilities: pollunit.com/en/photo-contest
Karin has already made suggestions. What are your ideas?
raingirl club has replied to Bergfex clubBergfex club has replied to raingirl clubBergfex club has addedSee also: 2022-11-04 Club News
Bergfex clubTo our pleasure, there is a selection method in the Poll Unit voting tool that works without the usual competition symbols (stars, thumbs up, etc.). The editorial team thinks that it is advantageous to avoid associations with competitions. It is a kind of slider that can be moved from left to right:
By reducing the increments to 4, the indifferent middle position is avoided. In this way, our vote would become even more meaningful.
This selection method has been successfully tested twice in the meantime. In future, it will replace the method described in paragraph 5 of the introductory article.