WebAIM - Web Accessibility In Mind

E-mail List Archives

Thread: Error prevention

for

Number of posts in this thread: 5 (In chronological order)

From: Isabel Holdsworth
Date: Mon, Apr 30 2018 5:12AM
Subject: Error prevention
No previous message | Next message →

Hi guys,

I have a bit of an anomaly here, which I'm sure some of you will have
come across in the past.

WCAG2 success criterion 3.3.4 Error Prevention (Legal, Financial,
Data) requires checking, confirmation or reversal of data in order to
pass.

We are dealing with test questions that get easier or more difficult
depending on each successive answer. So two students giving different
answers to question 1 could be given to almost totally different
tests.

If we were to offer a review screen, and the user changed their answer
to a question early in the test, this would likely invalidate the rest
of the test.

Would it be enough in this instance to offer decent client-side form
validation, given that the tests are delivered on devices that
definitely have JavaScript enabled?

Cheers, Lynn

From: JP Jamous
Date: Mon, Apr 30 2018 5:18AM
Subject: Re: Error prevention
← Previous message | Next message →

Hi Lynn,

I am having a hard time understanding your question. Can you give us an example?

Based on what I was able to follow, you should not be combining students if each has a different types of questions. You would need a consistent logic that should work across any test. As you add students and different tests to the mix, you are complicating your life and the project.

Secondly, I am not sure how this success criteria would apply to your situation. That was why I needed more clarifications.



--------------------
JP Jamous
Senior Digital Accessibility Engineer
E-Mail Me |Join My LinkedIn Network
--------------------


From: Mallory
Date: Mon, Apr 30 2018 5:28AM
Subject: Re: Error prevention
← Previous message | Next message →

I would think this falls under Testing. This isn't students' personal data (PI) is it? It's test questions? And it's not them doing banking or agreeing to contracts or the such? So I suspect you fall under a different WCAG guideline with notes on "testing". Tests have some special considerations with timing, error prevention etc because of the fact that they are school or certification tests.

cheers,
_mallory

On Mon, Apr 30, 2018, at 1:18 PM, JP Jamous wrote:
> Hi Lynn,
>
> I am having a hard time understanding your question. Can you give us an example?
>
> Based on what I was able to follow, you should not be combining students
> if each has a different types of questions. You would need a consistent
> logic that should work across any test. As you add students and
> different tests to the mix, you are complicating your life and the
> project.
>
> Secondly, I am not sure how this success criteria would apply to your
> situation. That was why I needed more clarifications.
>
>
>
> --------------------
> JP Jamous
> Senior Digital Accessibility Engineer
> E-Mail Me |Join My LinkedIn Network
> --------------------
>
>
>

From: Patrick H. Lauke
Date: Mon, Apr 30 2018 5:32AM
Subject: Re: Error prevention
← Previous message | Next message →

On 30/04/2018 12:28, Mallory wrote:
> I would think this falls under Testing. This isn't students' personal data (PI) is it? It's test questions? And it's not them doing banking or agreeing to contracts or the such? So I suspect you fall under a different WCAG guideline with notes on "testing".

3.3.4 does include "or that submit user test responses", so it's
applicable I'd say.

I'm wondering if offering the user a way to bail out and start again
(with a different set of randomly generated but dependent questions)
would be an acceptable workaround here. Failing that, I'd document the
decision not to allow users to go back and edit individual responses on
the grounds that it would invalidate the purpose of the test.

P
--
Patrick H. Lauke

www.splintered.co.uk | https://github.com/patrickhlauke
http://flickr.com/photos/redux/ | http://redux.deviantart.com
twitter: @patrick_h_lauke | skype: patrick_h_lauke

From: Isabel Holdsworth
Date: Mon, Apr 30 2018 5:53AM
Subject: Re: Error prevention
← Previous message | No next message

Thanks everyone. Yes, testing is included in 3.3.4.

So here's how it works:

A student answers question 1. Depending on his answer, he either gets
an easier or a harder question next time around. And so it continues
all the way through the test. There are 1,500 questions, of which the
student must answer 26. So it would be very rare for two students to
answer exactly the same questions.

The tests are timed, so offering the student a whole new test isn't a
feasible option.

I guess we're going to have to document the fail. I'd rather have
found a work-around, but I don't think there is one.

Cheers, Lynn

On 30/04/2018, Patrick H. Lauke < = EMAIL ADDRESS REMOVED = > wrote:
> On 30/04/2018 12:28, Mallory wrote:
>> I would think this falls under Testing. This isn't students' personal data
>> (PI) is it? It's test questions? And it's not them doing banking or
>> agreeing to contracts or the such? So I suspect you fall under a different
>> WCAG guideline with notes on "testing".
>
> 3.3.4 does include "or that submit user test responses", so it's
> applicable I'd say.
>
> I'm wondering if offering the user a way to bail out and start again
> (with a different set of randomly generated but dependent questions)
> would be an acceptable workaround here. Failing that, I'd document the
> decision not to allow users to go back and edit individual responses on
> the grounds that it would invalidate the purpose of the test.
>
> P
> --
> Patrick H. Lauke
>
> www.splintered.co.uk | https://github.com/patrickhlauke
> http://flickr.com/photos/redux/ | http://redux.deviantart.com
> twitter: @patrick_h_lauke | skype: patrick_h_lauke
> > > > >