Pearson failure today

Are folks having good experiences with Pearson Vue during the retake? Anyone heard anything? (Not asking for exam specific details. Just wanting to know if the system issues are resolved.)

Ongoing system issue: Bring F4 anchor functionality to the exams! Haha

2 Likes

Guy on my team who had to retake seemed to go fine

Nice!

I saw the CAS made a comment on the May 1st section of their site today stating they are still getting a ton of emails and calls. Hopefully things are working out.

Oh trust me that has not changed when the CAS makes the exams too hard. Fall 2018, 2021, and 2022 all come to mind when this continued to happen.

The only time they take the opposite pass everyone approach is when the issues are caused by technical failures like Spring 2018 with the botched TBE where they blamed TrueAbility/PSI or whatever it was called, and now this sitting where they blame Pearson.

Honestly as a candidate, it’s actually extremely advantageous to be sitting when these technical issues arise since you’ll see the highest pass ratios ever, as opposed to ‘normal’ sittings where they do what you outlined.

1 Like

The CAS released a full throated apology for the May 1 failure-from Pearson VUE, laying the blame squarely on them.

The CAS, throwing someone else under the bus? No. Way.

If Pearson VUE turns around with a statement pointing the finger of blame back at the CAS, complete with receipts, @Breadmaker will have to post all the popcorn GIFs.

I am glad to see that the CAS is re-evaluating its exam administration process and discussing ways to make the entire testing process better so as to avoid problems with … :rofl: :rofl: :rofl:

The question to me is the memory issue due to the way the CAS has now changed the tests to. Previously there was no scrap section and there was only 20 questions. Now they added a scrap section and changed the tests from 20 questions with multiple parts to 40 questions requiring you to go back to the prior question.

Candidates have expressed how the new format is worse and makes it harder to reference previous answers.

If it were “My fingertips brushed with the proctor’s when I handed over my ID”, or “I saw another candidate’s bare ankle!!”, then you know I would have a cheeky popcorn gif.

But this some ******in’ :ox: :poop: that the candidates endured!!!

1 Like

As @Rastiln noted, the CAS didn’t even bother to spell check their statement! :joy:
https://community-new.goactuary.com/t/2024-exam-7-progress-thread/8597/74?u=joejam

That statement is from a PearsonVUE employee, not the CAS.

1 Like

2 things:

  1. Would it kill you to proofread something that you are publishing on your website and email blasting?
  2. Was there a CAS statement that we missed in which they accepted some responsibility for their handling of the aftermath of the debacle, such as making candidates wait a week for any updates?
1 Like

I don’t know what they told candidates, but they sent several conciliatory emails to members about how difficult it was and they would do their best and here’s where to look for the latest info.

It’s not as if they have a contingency plan ready to publish for every conceivable problem. Waiting a week for the CAS to fully understand what happened and agree on a mitigation seems reasonable to me.

I thought their communication with candidates after the fact was pretty good.

1 Like

Unless it is your contention that CAS has falsified the statement from Pearson, this is a pretty unfair characterization. Pearson said “this was our fault, so sorry” and CAS published Pearson’s statement/apology that appears to have been intended for CAS to share in exactly this manner.

I’m not CAS, not taking or proctoring exams and am only dimly aware of what happened, but everything I do know points to this being Pearson’s shortcoming. CAS is only to blame for using them, and possibly for not insisting on more rigorous advanced testing of the system after what seems to be a format change.

Well, and we can certainly pick nits over how the CAS handled the aftermath with the candidates. But the initial problem appears to have been Pearson’s fault.

The CAS could have seen something like this happening. There have been issues in the past with exams freezing on candidates that has been brought up both here and on Reddit. In addition, in the last two sittings there have been a paucity of seats in the test window concentrating most candidates in the same testing window.

Well, what happens when you have a lot of people all trying to run a memory intensive process on a server at the same time?

(I am directing my comments toward the CAS not at the twig)

Yes, I suggested that perhaps CAS should have required more robust system-testing by Pearson.

fair enough

I’m sure it’s a one-off thing, it will never happen again.

I assume you’re out of red font?