Let me define my terms:
A standard exam is composed of a relatively large number (50 or so) of independent questions. It is not adaptive, and all items have the same value. Most are multiple-choice, though a small number may be 'select and place' or 'hotspot' questions (which often could be presented as multiple choice). Most current Microsoft exams fit this description.
A scenario question is common on standard format exams, and is simply a bit more realistic than a non-scenario question. For example, rather than ask "What is 2 + 2?", a question might say, "Maria has eleven items. Four are apples, two are oranges and the rest are vegetables. Bruno had six oranges, but gave one to each of his two friends and ate two more. How many oranges do Maria and Bruno have?". While it really boils down to the same question, you also have to spend extra time identifying exactly what the question is and what information presented is relevant.
A case study exam is composed of a small number of relatively lengthy business descriptions - possibly ten minutes or so of reading. Each of these case studies has a small number of related questions, usually five to ten. The entire exam may have fewer than thirty questions. The questions themselves can be much more complex than what you'd see on a standard exam - such as constructing a simple diagram for a database structure, flow of data through a system, or network design. The 'Solution Architecture' MCSD exams (70-100 and 70-300) and 'Design' MCSE exams (70-219, 70-220, 70-221, and 70-226) are in this format.
The case study format is not necessarily more difficult than standard format, but it is quite different, so it's worthwhile to know what to expect.
Microsoft has a FAQ page that includes questions about case study exams.
Examples
Case study exams are usually much longer than a typical standard exam. I had the pleasure of taking 70-219 during a brief window where the live version was FIVE hours long. It's back down to a merely exhausting four hours last I heard. I believe 70-100 has always been just under three hours. For me, this means just saying yes to some caffeinated assistance and on a related note, taking breaks during the exam. I highly recommend checking both of these needs out with your test center ahead of time. It doesn't seem to be a problem, but especially if you're not a 'regular', they may be uncomfortable with letting you leave the test room or bring a six-pack of carbonated energy in with you.
In my experience, running out of time is a rare problem on case study exams, however you can still make it a much easier experience with some good time management.
Read the questions first
Remember that Microsoft has a pool of questions for their exams, and you'll only see a subset each time you take the exam. It also seems that each case study has a pool of questions as well, however the case study 'story' doesn't change. This means that the case study presents lots of detailed information required to answer questions that you will not be asked. Rather than sweat through scattered details on globally available bandwidth to determine a good replication strategy, I think it makes sense to first peek ahead at the questions to see if you're even asked to do anything remotely related to available bandwidth. What I do is make a list of very brief summaries for each question, such as "1. global replication diagram". Usually I'll have no idea what the best answer is until I read the background information, but once I have an idea of what I'm looking for, I find reading the scenario goes much more quickly. Relevant details are easier to spot the first time through, and irrelevant information doesn't bog me down.
Do your rough drafts on paper
The exam software user interface for diagrams and even some of the other question types is a bit awkward, and most exam computers will have shamefully poor screen resolution. I find it very difficult to work with. I recommend using pencil & scratch paper for working out any diagrams, and when you're satisfied with that, just copy that work to the exam to minimize your interaction with it.
Some questions really are embarrassingly simple
It sounds condescending, but a lot of these exams boils down to reading comprehension. There are a few questions that I think a reasonably intelligent adult could answer without ever having seen a computer. These remind me of the 'quizzes' at the end of my high school Health class chapters that simply took a sentence straight out of the chapter, replaced one of the nouns with Who, What, or Where and added a question mark.
Some questions have more than one correct answer
So don't fret if you have two options that both seem just as correct (or less incorrect!) as each other. You may be exactly right. I think I've seen this come up when building an ordered list of items - there are many that clearly must precede or succeed other specific items, but some are not directly dependent on all other items. For these, I assume those items can go in any relative order. For example, if a question asks to order the following tasks:
The only firm requirements I'd see would be that the shoes have to go after socks. The case study might indicate if particular footwear is best for breakfast, but otherwise, I think breakfast could go anywhere in the list.
Most questions allow partial credit
Some case study questions are simple multiple-choice questions, and probably are all or none on being right or wrong. But the diagrams, trees, lists, and so forth are far more complicated. What if you have a list of fifteen items, and two are out of place? I've never seen specifics on how the scoring really works, but you should get credit for getting that question 'mostly' correct. See the Transcender demos for a feasible example of this partial credit scoring.
Not all questions count the same
As mentioned, questions can range from relatively simple multiple choice items to complex diagrams that can take ten minutes to work out. Again, I've never seen specifics on how questions are weighted, but it's clear that those more complex questions are worth more points than the simple ones. Again, the Transcender demos have one possible implementation of this. The important point is that unlike standard format exams, it is very unwise to give up on a difficult question to avoid spending too much time on it. Your time will be well spent if you work through each question carefully, even though you will spend far more time on some questions than on others.
Read the question carefully
I've been saved by the limitations of the test engine on a few diagram questions where I get an error message before I can finish diagramming the whole database. It has said I can only create N connections in the diagram, even though it's clearly impossible to build my diagram from A to Z with so few connections. I reread the question: "Create a diagram from A to G...". Whoops! On the 70-100 beta, I think the next question was "OK, now diagram the other half". If my oversight hadn't been caught, I'm afraid the 'extra' information in my answer would have counted against my score.
Read the instructions for each question format
For some of the complex question types, you don't have to use all options available. A data flow diagram may have multiple processes available for connecting, and multiple types of connections to make. I would assume most should be used, and wouldn't worry if I felt all should be used, but also don't worry if some don't belong in the part of the system the question covers, or are simply not appropriate anywhere. The same goes for list and tree building questions. But for these, items can sometimes be used more than once. Either way, I found it helpful to click the 'Instructions' button for these question types, and it was always specified whether or not items could be reused or unused.
NOTE - I think the following two points no longer apply to most MS case study exams. The 70-300 beta had one big chunk of time allotted for six case studies, and I understand the 70-219 iteration I saw with discrete time allocations for each study was only available briefly around the beginning of 2002. It seems that 70-100 is the only exam that still makes each case study a separate mini-exam with its own time allotment, and the others offer some kind of onscreen progress in terms of the entire exam.
Leave a trail of breadcrumbs
This varies with VUE and Prometric, and is certainly subject to change as they update their software. But some early deliveries of 70-100 did not offer any obvious way to figure out your overall progress in the exam. That is, whether you are working on the 3rd, 4th, or 5th case study. It may sound silly, but it's very easy to lose track of that after burying yourself in the vast information provided in a case study. It can be very discouraging to finish your last check of your answers on the last case study, decide you're ready to see your result, only to have one more case study pop up for you to work on. At the beginning of the exam, you should be told how much time will be allowed overall. I keep a brief list of each case study and the time allowed for each so I can gauge how close (or far) I am from finishing. It's been a while since I took a case study exam with VUE, but I was unable to find anywhere this information was available to me during the exam. With Prometric, double clicking the timer clock in the top right popped up a dialog that listed all of the cases and which ones still remained. Otherwise, I mostly preferred the VUE software for case study exams, so I wouldn't choose a test provider based on this alone.
Let the clock run
I understand this rule has varied for some exams. But on the case study exams I've taken and heard about from students, each case study had a separate time allotment for completion. So a three hour exam might have five case studies with anywhere from 25 to 50 minutes allowed for each one. If you finish a case study with time left over, you can't use that extra time on your next case study. So if you need to take a break, don't finish the case study yet. Close your eyes and relax, get up and go to the bathroom if you've cleared it with the proctor, stretch, etc. After two or three case studies, it can really help bring back your attention. NOTE that when each case study has its own time allotment, you cannot return to a case study once you start the next one. This should be very obvious when you finish a case study, with its own 'review' screen and an extra warning that you won't be able to go back to it when you move on.
Home Back to Certification Stuff
about this page
Last Update: January 18, 2003