oooohhhh!  a poetry contest!

the things I could say, the rhymes I could make
but I think I'll decline for my account safety's sake

while I do get upset and I do want to scream
when I see the unfairness that is awfully extreme

sometimes I snarl, and at times want to swear
when CF is unhelpful and I wonder if they care

I try to be calm, I try to keep hope
and if CF remains stubborn, I'll just have to cope

so I'll keep doing tasks and will keep up the good work
and my fingers are crossed that I don't get a TA that's a jerk

the end

copyrighted by me Mar 31/2016

Read the new CF blog post today.  A little overdone on the being nicey nicey stuff, but what the hey.

After reading that, I decided to peruse the Terms of Service.

OMG peeps - the space bar is your friend!

Starting at #7 under the "You WILL NOT"

Things like that is like nails on a chalkboard!

a.please fix it

b. thank you

It left me wondering what the heck is fengage and ginsert

Am I:

a) a wee bit annoyed
b) slightly annoyed
c) a bit more annoyed
d) annoyed
e) more annoyed
f) even more annoyed
g) extremely annoyed



Just trying to illustrate the insanity of splitting hairs when it comes to those types of tasks.

If a TA really needs that kind of data, fine.  But we all come from different walks of life, have different frames of reference, different cultures.  These types of tasks should not count for Global Accuracy.

the correct answer was g) 

if you got it wrong, I'll flag you  

Me, after attempting a subjective task where I was expected to be a mind reader:

The flogging flagging will continue until morale improves

And morale is pretty low right now

If I didn't know better (but do I?) I'd be thinking that TAs get a discount directly related to the number of flags they hand out.

To quote another worker:  "is like crushing us contributors"

2016 - The Year of the Flag

No, I didn't come up with that - other workers did.  It really does give an indication of how the workers are feeling these days.

As for me, I'm leaning more towards

2016 - The Year of the Broken Promises

Prove me wrong

Just saw a story about Uber in my facebook feed today.

Now I'm sad all over again.

#IWantToAuditTickets

#You'reJustCruel

Can you give me one good reason why I should try a brand new task which could result in putting my levels at risk?  Sometimes I see a new task and think to myself "hmmm...that looks interesting".  But I've read too many horror stories lately, so I check the forums and sure enough I see posts like "xyz new task, major flag!!"  with many "me too" responses.  So then I think "whew, dodged a bullet there!"

I've had level 3 pretty much from the getgo when levels were first introduced.  I've never received a flag nor have I ever lost a level.  My accuracy is in the high 90s.

You'd think by now I would have established myself as a good and honest worker.  So why should I live in fear of one of Those Types of Task Authors?  If I give them the benefit of the doubt and complete their task and do end up with a flag, will I have to jump through hoops in getting it removed?  Why should I have to?  Why doesn't my history speak for itself?  And WHY would you allow a new and unproven TA to trump someone's excellent history?

Enquiring minds want to know

It's been a few weeks since my first post regarding unfair Crowdflower policies and unfair and downright crooked task authors.

Are we any closer to a resolution?  Sure, CF and their reps are making noises and asking opinions, and even posting a few tasks that ask for feedback from the workers, but are they just placating us by throwing water on the flames in the hopes that this will blow over?  Will things remain unchanged after giving empty promises and false hope?

I know CF has it tough when dealing with the scam workers.  I remember reading in the old CF forum of a worker complaining of being banned.  Turns out he had copied and pasted gibberish into text boxes in a task. His excuse?  He was in a rush and didn't have time to type in the correct answers.  After my eyes finished rolling right out of my head, I thought to myself that this moron should not only be banned from Crowdflower, but from the internet as well.  Seriously.  In what world would that excuse be ok?  Was he really stupid enough to think that CF would give him a second chance?

I'm sure this wasn't an isolated incident, and that CF has to deal with that kind of crap on a daily basis.  What concerns me is that good workers are being lumped in with the bad.  Flags are handed out like cookies, CF denies responsibility and the workers have no recourse and no way to contact the TA.  mTurk may have its faults (not paying Canadians in real money is one of them, but I digress, and that's a topic for another blog) but at least they offer a way to contact a requester.

While I don't have the answers, I do think that things need to change.  There needs to be better training for the TAs, tasks need to be double and triple checked before going live, there needs to be a mechanism in place to dispute flags, there needs to be stricter guidelines for handing out flags.  Really, is submitting a task 5 seconds before some Top Secret Time Limit actually worthy of a flag?

If a worker is deliberately trying to scam like the example given above, that worker should be banned permanently.  But any other type of flag should have an expiry date.  As I mentioned a few weeks ago, even a bank robber can get parole - he isn't in prison for life.  Why should CF workers get life without parole? 

They shouldn't.

I have an idea for a new task

Title:  Please Evaluate this Task Before It Goes Live

1)  Are the instructions thorough and easy to understand?

2)  Are the answers to the test questions correct?  If not, tell me which ones are incorrect.  What do you think are the right answers?

3)  Does the task submit properly?

Optional:  Would you recommend this task to other workers or would you tell them to avoid at all costs?

Would you jump into a snake pit if someone told you that the snakes MIGHT bite you, or they might not.?  And they MIGHT be poisonous and they might not.  And if a poisonous snake bites you, we MIGHT give you the antidote right away, or we might make you wait and wait and wait...or we just might not give it to you at all.

Sounds like fun, huh?

Sure would be helpful to have the ability to identify the biting snakes from the non biting ones, and the poisonous snakes from the non poisonous ones.

Is Crowdflower trying to pull a fast one with this recent development of double quiz mode or is it a glitch?  One can make a mistake on the first page but won't know WHAT got corrected and WHY, which is imperative if one is to learn how to do the task properly, especially when it isn't a true learning mode like yesteryear.  Then there's the possibility it was an unfair correction that needs to be disputed. Though that almost never happens.  /snark

I want to give them the benefit of the doubt but I'm not sure they've earned that yet.

-----

Got a real bad Uber-craving goin' on - throw in some wine and I'll be one happy camper

Evidently even CF created tasks aren't immune to bad test questions

good. grief.

If you CF Peeps were as strict with the TAs and their task creation abilities, or lack thereof/bad test questions as you are with the workers the Tasking World would be a much better place.

Wishing we could give flags to TAs, or at least give them a time out for bad behaviour.


P.S. I'm talking about the fraudulent "click on my banner" one, and the knucklehead whose corrections say "you answered xyz and the correct answer was xyz"

**all references to answers are fictitious and the names have been changed to protect the innocent.  (me)**

If Crowdflower is thinking of adding a new level in addition to Levels 1, 2 and 3, I have a suggestion - a level for the Certified Mind Reader.  So many tasks seem to be aimed at those that have that ability.

Really appreciative of the tasks you've been posting asking what features we would like to see, and what our favourite task is and what we want to tell the TA.  It's giving me a little bit of encouragement.

And now a request - on the job history page we can search by job id.  We'd have to keep our own record of IDs in order to use that.  How about adding a keyword search?   That would be more useful. ThankYouVeryMuch.

Per the CF rep in the forums, the higher level workers should be held to a higher standard and judged more strictly in their work.  Is this called CrowdLogic? 

We can only work with what we are given - broken tasks, bad TAs and bad test questions will not result in high quality work.

And even IF the tasks weren't broken, the TAs were fair and the test questions were correct, (I can dream) if a "lower level" worker and a "higher level" worker are each going to get 2 cents for doing the same task, why should the "higher level" worker knock themselves out for the same pay as the "lower level" worker?  Higher quality comes at a premium.

Nice double speak, btw.  Did you have a big pow-wow to try and come up with counter-arguments?  If you can't dazzle them with brilliance...

The flag system still needs to be fixed.  The TAs have too much power and some of them abuse it.

It should be mandatory for TAs to fix bad test questions if the task affects global accuracy.  If they don't want to be bothered, then the task should simply not count against the GA.  Though even if it doesn't, it is very disrespectful to not fix the bad questions because it would still affect our earnings should we be expelled unjustly.  However, that is more preferable than a hit to our GA.