Regarding how to improve in Data Sufficiency questions, you may need to adjust how you approach those questions versus how you attack problem-solving questions. For example, you need to understand that Data Sufficiency questions do not have to be solved out to the “bitter end.” Take a look at
example 10 here. Notice that the goal of the problem is to determine the mean grade for the left-handed students. If you look at the solution for statement two, it’s clear that we can determine the value of L well before we actually do. So, taking that mentality when solving DS questions should greatly help your accuracy and your timing.
There are definitely certain nuances or traps to be aware of when solving Data Sufficiency questions. One of those traps is the dreaded “C Trap.” In C-trap questions, you’ll be baited into choosing answer choice C because C so clearly and obviously seems to be the correct answer; DON’T TAKE THE BAIT!! For instance, take a look at
example 9 here. Upon first glance, choice C seems like a logical answer, right? Yet, if you properly attack the problem, you’ll see that you can determine the value of x using statement one alone.
Once you further develop your general Data Sufficiency skills, whether you correctly solve DS questions will be based more on your topical knowledge than on anything else. For example, if you are given a DS question testing you on “units digits,” such as
example 1 here, and you are not skilled in working with units digits and do not know that the base of 7 has a units digit of 1 when raised to an exponent that is a multiple of 4, then how can you expect to correctly answer that particular DS question, right? Thus, to improve, you would have to spend time reviewing not Data Sufficiency concepts but concepts related units digits patterns.
In summary, yes, there are some general Data Sufficiency skills that you could use to avoid being trapped or wasting time doing unnecessary math. At the same time, if you thoroughly master each GMAT quant topic, then you should not have any major issues with Data Sufficiency questions.