The next generation and cheating

I'm curious what MS is actively doing to curtail cheating with the next
generation of certs?

My suggestions are:

1. A minimum question bank of one thousand questions per exam, adding an
additional 100 questions every two months. These don't have to be killer
difficult questions. In fact I would divide questions into two groups. The
first group would be "gimme" questions of which this thousand would belong.
This makes it brain dead easy to keep the size of the question bank high.
Then have a secondary set of questions of which you maybe have 100 which are
tough questions, rotating as often as possible. Each exam should be 70%
gimmes and 30% toughies. Scores would be weighted as a percentage of gimmes
to toughies.

If you do this, you remove the cheating incentive, because it would be
easier to get an honest understanding of the topics than to memorize 1000
questions.

2. Remove the cheating incentive by minimizing the amount of esoteric stuff
in the tests. I failed the SQL exam the first time, but even after
extensive studying, I only made a 700. Does this mean I do not have a high
experience level with SQL Server? I think not. I think it means that
whether I can come up with the correct DBCC command without referencing
books online is meaningless.

3. Monitor braindump sites extensively. Flood them with incorrect
questions and answers. Make the information unreliable.

4. Make people buy into honesty. As part of the testing process, make
people sign a secondary agreement of honesty or take a verbal oath - or
anything to get them to treat the tests with respect.

5. Use technology - we are technologists, right? Craft dynamic questions,
where the computer can randomly pick from 10 or so options to inject in a
problem, and have an algorithm that translates options to correct answers.
This in itself would increase the complexity of the exam by a factorial.

I have one test left for my MCSD, and a thousand dollars and hundreds of
hours later, I question the value of it because I keep running into people
with certs that have no business having them.

Do this generation of certs right MS.

Justin Weinberg

Echoed forth from the dank caverns of microsoft.public.cert.exam.mcsd,
the plaintive wail of Justin Weinberg:

> I'm curious what MS is actively doing to curtail cheating with the next
> generation of certs?

Fsck-all.

<rant at the wind snipped>

Sux, don't it? They could not care less. It's nothing more than a money
farm for Bill, and a bloody genius bit of marketing. If they discouraged
cheats, they'd sell less tests. We have reported cheats to MS, we have
notified them of cheat sites, I have even gone so far as to e-mail and
ask whether they condoned or condemned certain cheat sites.

Do you know what response I (or anybody I know) has gotten from
Microsoft?

Care to guess?

Anyone? Anyone? Class? Anyone?

Nada, Zip, Zero, in short; fsck-all.

--
JaR
MCNGP 10110
Remove hat to reply
The only worthwhile doors in life that will open for you are the ones
you kick down or make for yourself. ~Rob Lind man

JaR

Hello.

Try brainbench.com
.... they have a better solution: 40 out of about 100-150 questions (I
guess), differentiated difficulty on the questions, the better you answer,
the harder the questions get, questions are invisibly classified something
like "easy", "not easy" and "hard", where "hard" questions often are
confusing or based on niche-knowledge (much like this post); ca. 50% correct
is enough to pass, ca. 75% gives you something called the master-level. The
system is not perfect, some of the tests are so hard that none has made the
master-level and others so easy that too many can get the master-level. I
think it is expected that about 40% shall pass, 20-25% of these with master
level.

Suggestion
1) Questions should be updated regulary, without notice.
2) You could require at least one master-level pass to get any title.
3) There might be something called Master Professional with additional
benefits (there are fewer and fewer jobs in the western world and such ways
to differentiate on self could be a selling-point for the specialists)
4) You should not get your score reported, only "failed", "passed" or
"master" (with the possible addition of "don't go anywere; you need to come
work for us")
Braindump-business should need to take more tests to get all the knowledge
and not as much basis on how much of a test material they understand.
It would be challenging to create braindumps helpful to pass any exam at
master level, meaning that even if people used braindumps, they would need
to master at least one area to get a title (see 2).

Expected results
This would, in my opinion, not only kill the braindump-business, but also
help to give additional (automatic) general advice to test-takers once they
get their, say MCSD, for instance:
- "BizTalk speciallity adviced" (because of strong integration and process
skills)
- "Go for Certified Architect" (because of strong
business/scenarion-understanding)
- "Add DBA-skill to you resume" (because of strong database-skills)
- "Add security/network-skill to you resume" (because of what ever is needed
here)
- "..." (because of ...)

Cost
This would cost money. Money must come from somewhere. Preferably earnings.
The additional cost of this way of doing things could be covered by
selling/feeding information to reqruiters and sharing prospects with
internal MS and partners.
Helping in the process of matching people against job requirements.
Also, if it could help msdn and technet to develop documentation on areas no
one understands (say XQvery for instance, or maybe I'm the only one that
can't even spell it...), it might be seen as an investment in the
documentation.

Gorm Braarvig

"Justin Weinberg" <jweinberg@_spammyoff_mrgsoft.com> wrote in message
news:...
> I'm curious what MS is actively doing to curtail cheating with the next
> generation of certs?
>
> My suggestions are:
>
> 1. A minimum question bank of one thousand questions per exam, adding an
> additional 100 questions every two months. These don't have to be killer
> difficult questions. In fact I would divide questions into two groups.
> The first group would be "gimme" questions of which this thousand would
> belong. This makes it brain dead easy to keep the size of the question
> bank high. Then have a secondary set of questions of which you maybe have
> 100 which are tough questions, rotating as often as possible. Each exam
> should be 70% gimmes and 30% toughies. Scores would be weighted as a
> percentage of gimmes to toughies.
>
> If you do this, you remove the cheating incentive, because it would be
> easier to get an honest understanding of the topics than to memorize 1000
> questions.
>
> 2. Remove the cheating incentive by minimizing the amount of esoteric
> stuff in the tests. I failed the SQL exam the first time, but even after
> extensive studying, I only made a 700. Does this mean I do not have a
> high experience level with SQL Server? I think not. I think it means
> that whether I can come up with the correct DBCC command without
> referencing books online is meaningless.
>
> 3. Monitor braindump sites extensively. Flood them with incorrect
> questions and answers. Make the information unreliable.
>
> 4. Make people buy into honesty. As part of the testing process, make
> people sign a secondary agreement of honesty or take a verbal oath - or
> anything to get them to treat the tests with respect.
>
> 5. Use technology - we are technologists, right? Craft dynamic
> questions, where the computer can randomly pick from 10 or so options to
> inject in a problem, and have an algorithm that translates options to
> correct answers. This in itself would increase the complexity of the exam
> by a factorial.
>
> I have one test left for my MCSD, and a thousand dollars and hundreds of
> hours later, I question the value of it because I keep running into people
> with certs that have no business having them.
>
> Do this generation of certs right MS.
>

Gorm Braarvig

"Gorm Braarvig" <> wrote in message
news:uH%23hs%...
> Hello.
>
> Try brainbench.com

I don't know whether to laugh or cry.

Frisbee®

Echoed forth from the dank caverns of microsoft.public.cert.exam.mcsd,
the plaintive wail of Gorm Braarvig:

> Helping in the process of matching people against job requirements.
> Also, if it could help msdn and technet to develop documentation on
> areas no one understands (say XQvery for instance, or maybe I'm the
> only one that can't even spell it...), it might be seen as an
> investment in the documentation.

In the immortal words of an old friend:

"You just don't understand! It doesn't scaaaaaale!"~$.50 to L.A.R.

--
JaR
MCNGP 10110
Remove hat to reply
Have you noticed that a slight tax increase costs you two hundred dollars,
and a substantial tax cut saves you thirty cents?

JaR

>> Helping in the process of matching people against job requirements.
>> Also, if it could help msdn and technet to develop documentation on
>> areas no one understands (say XQvery for instance, or maybe I'm the
>> only one that can't even spell it...), it might be seen as an
>> investment in the documentation.
>
> In the immortal words of an old friend:
>
> "You just don't understand! It doesn't scaaaaaale!"~$.50 to L.A.R.

What does not scale?

- XQuery
Agreed, but XQuery against a field of a record of a recordset in SQL2005
does not necessary need to scale.

- Matching people against job requirements
No. Does not scale. There is room for smart solutions here. Maybe you are
the guy to invent something?

- develop documentation on areas no one understands
Scales?

BTW: do this scale: http://62.89.36.69/g/g.html (Suspicios site)?
Scalability is a special interest of mine, if you find something that does
not scale, it might be something I can fix.

> You just don't understand!
You are right. I might be a bit slow. Please explain, I am into learning.

Gorm Braarvig

"Frisbee®" <> wrote in message
news:%...
> "Gorm Braarvig" <> wrote in message
> news:uH%23hs%...
>> Hello.
>>
>> Try brainbench.com
>
> I don't know whether to laugh or cry.
>
>

Knowledge is over-rated.

Gorm Braarvig

Echoed forth from the dank caverns of microsoft.public.cert.exam.mcsd,
the plaintive wail of Gorm Braarvig:

>> You just don't understand!
> You are right. I might be a bit slow. Please explain, I am into
> learning.

It's a quote from a discussion we had long ago with one who allegedly had
inside info on the exam prep process. Her constant reply to various
suggestions about the testing procedure was always that we did not
understand the arcane machinations of the test architects, and that our
suggestions were invalid, as they just would not scale.

The one to explain has not been around for a while, and probably
fortunately, as she would be just as unable to explain why $Solution does
not scale, but would be repeatedly telling us all what dolts we are for
not understanding in the first place.

I leave it to others more masochistic than I to invoke L.A.R. if they
feel the need for further clarification.

--
JaR
MCNGP 10110
Remove hat to reply
If you're playing a poker game and you look around the table and can't
tell who the sucker is, it's you.

JaR

"JaR" <> wrote in message
news:Xns9765744022213Talleywhacker@207.46.248.16.. .
> Echoed forth from the dank caverns of microsoft.public.cert.exam.mcsd,
> the plaintive wail of Gorm Braarvig:
>
>>> You just don't understand!
>> You are right. I might be a bit slow. Please explain, I am into
>> learning.
>
> It's a quote from a discussion we had long ago with one who allegedly had
> inside info on the exam prep process. Her constant reply to various
> suggestions about the testing procedure was always that we did not
> understand the arcane machinations of the test architects, and that our
> suggestions were invalid, as they just would not scale.
>
> The one to explain has not been around for a while, and probably
> fortunately, as she would be just as unable to explain why $Solution does
> not scale, but would be repeatedly telling us all what dolts we are for
> not understanding in the first place.
>
> I leave it to others more masochistic than I to invoke L.A.R. if they
> feel the need for further clarification.
>
> --
> JaR
> MCNGP 10110
> Remove hat to reply
> If you're playing a poker game and you look around the table and can't
> tell who the sucker is, it's you.

I understand.

Most people I have met with the "will not scale"-argument
a) does not know what they are talking about
b) have no low-level experience
c) have no math-background
d) have low problem-solving skills
e) talks mostly in general terms
f) has abstract reasons ready for anything and specific knowledge of nothing
....the list goes on...

the point I want to make is that ANYTHING is scalable. Many things are hard
to scale, "arcane machinations of" quote on quote on quote on quote "test
architects" can't possibly qualify to be hard to scale. This discussion
would leave an impression on me too, just sad I could not be part of it,
sounds out of this world...

BTW:
Only exp I have on these tests are on betas on the new TS and PRO-exams, and
I don't think they are too bad (the questions, the composition I know
nothing of). I quite liked the scenarios with 5-10 questions bound to real
world scenarios (on some DB-exams). I might change opinion when I get the
results, of course (I am a soar looser).

Gorm Braarvig

"JaR" <> wrote in message
news:Xns9765744022213Talleywhacker@207.46.248.16.. .
> Echoed forth from the dank caverns of microsoft.public.cert.exam.mcsd,
> the plaintive wail of Gorm Braarvig:
>
>>> You just don't understand!
>> You are right. I might be a bit slow. Please explain, I am into
>> learning.
>
> It's a quote from a discussion we had long ago with one who allegedly had
> inside info on the exam prep process. Her constant reply to various
> suggestions about the testing procedure was always that we did not
> understand the arcane machinations of the test architects, and that our
> suggestions were invalid, as they just would not scale.
>
> The one to explain has not been around for a while, and probably
> fortunately, as she would be just as unable to explain why $Solution does
> not scale, but would be repeatedly telling us all what dolts we are for
> not understanding in the first place.
>
> I leave it to others more masochistic than I to invoke L.A.R. if they
> feel the need for further clarification.

I think she might have just been on the rag or something.

I never understood it either.

But I was askeered to ask her anymore.

Frisbee®