Another way to posit the stem is -
inspite of rejecting 500 of the 2,000 bulbs produced why will the plan of keeping quality constant not work?
Note the final aim is to keep quality constant not increase or decrease the production of bulbs.
Let us look at the answer choices:
A - How does this help keep quality constant? No link, this is out of scope.
B - Note we are NOT reducing the number of light bulbs produced the plan suggests
rejecting 500 of the 2,000 produced. Hence this choice does not have the facts right, hence can not dent the plan.
C - So what? We review each bulb, not related to the plan.
D - this is an interesting choice, if it is difficult to measure quality one would think how can one keep it constant. Needs analysis so lets make it a contender for the time being.
E - This is certainly a flaw, what if, as the choice says the quality of the 1,500 bulbs that are accepted every week is different. For example the 1,500 produced in Week 1 are of quality X and in Week 2 are of quality Y and we know X does not equal Y, well this makes the manager's plan flawed for sure. Hence, this is a strong contender.
Note: Choice E is about quality difference of the 1,500 bulbs across weeks and not about the quality difference of the 1,500 bulbs of any one week. Assume, 1,500 bulbs of week 1 are batch 1, week 2 are batch 2 ...... and so on, the issue at hand is - whether the quality across batches 1, 2, 3......taken as a whole is constant not whether the quality of the 1,500 bulbs within batch 1 or any other batch is same.
Choice between D and E:
D - Even though, we find it difficult to measure the quality of light bulbs but hypothetically we could set up processes such that exactly 500 bad bulbs are produced for every 2,000 bulbs produced. Furthermore, the 1,500 bulbs are produced of the same quality. This is far fetched but possible. Hence this is a weak choice as in some conditions it could actually not be a flaw in the logic of the manager.
E - as explained above.
If I were to pick a choice, I would go with E
always, the reason being that E always weakens the plan whereas D may not always weaken it.
Hence Pick E
But wait a second we have not fully milked the question

We learn a few key things:
1.
If you are unclear about the choice and can not find a clear reason to eliminate it fast, then make it a contender and move on. There are two benefits of this - Say you are unsure about choice B. If you unequivocally eliminate all the other choices, you can safely mark B without mulling over it. Secondly, it saves time, chances are high that if you make an unclear choice a contender and move on, at first what seemed nebulous will more clear when you visit the choice the second time. So essentially the point is rather than spending 20 secs on the choice in one go if you can split it into two chunks of 10 secs each it will benefit you more.
2. In the above case in choice D, the chances of a process being set up the way it is described may be 0.001% but it is possible, if that happens then the choice does not weaken the manager's plan. Whereas if choice E does not happen every week, it is 100% sure that the plan will not work. This is a subtle but HUGE difference! GMAT will give questions that are water tight hence always pick choices like E for answers which make the answer unequivocally correct, it is almost like the Quant section in a way!
Hope this was helpful.