Data quality and how it presents represents your research. What is the high quality in data? What is the good presentation figure? I sounds philosophical, but blaming the vagueness of the quality standard does not solve anything. When I was young and wild and free, I was thinking that the standard is mere someone's taste or opinion, not generalized things. Now I, old and experienced and disciplined, know that to have a standard is helpful, which I want to share a bit with you.
Data Quality
This is one of the areas I am most confident about. Please look at my TEM and SAXS data—these are examples I take pride in. Clean, precise data carries a kind of craftsmanship. So just like a sushi chef is in a zen state of mind and cut fish at a precise angle to make his sushi exquisite, you will adjust the sample condition of your samples.
I believe that exceptionally clean data has value on its own. Even when the novelty of the work is not very high, world-class data quality can still be persuasive, especially for high-impact journals. My favorite example is a quasicrystal single crystal that Professor Tsai at Tohoku Univ Japan fabricated. People knew the existence of quasicrystals, but Professor Tsai refined and refined the quasicrystal creation and finally he made a single crystal quasicrystal. Scientifically new? Maybe not much. However, the impact and appeal to the field are huge. Even Prof. Steinhardt highly acclaim the work in his book.
When your data quality is maximized and your analysis is done carefully, you often notice things that have not been reported before. For me, this pattern has led to new and exciting findings many times.The opposite is also true: poor-quality data is one of the most common reasons for rejection. And personally, I simply don’t want to do research with messy data. That’s why I often say: in academia it’s “publish or perish,” so the work we do is to “polish.”
What Is “Clean Data”?
In order to take clean data, you need to understand what is noise and what is signal, and eliminate the noise thoroughly.
For TEM, for example:
-
Remove stigmation completely
(check the Live FFT every time). -
Make sure the focus is perfect, and take multiple images at different magnifications.
-
Spend time searching for the best sample area.
-
If necessary, take ED or EDS data to strengthen the overall dataset.
-
For ED, choosing the correct area of interest is extremely important.
The goal is to collect data that is as clean and pure as possible. This requires attention, effort, and a sense of craftsmanship.
Know your enemy
Also, it happens that some unexpected instrument trouble or software glitch cause problems. Thus, keep in mind your targeted data so that you can modify your measurement when you are doing the measurement. Sometime (or often time), it is just a stupid mistake that you can solve right away. If you are not prepared for it, you will have a consequence. e.g., You bring the data back to your laptop, which was not checked 2 hours before the group meeting and you find something is wrong in your data and you just said "something is wrong" in the meeting, which make your audience bored as hell.
Just know your enemy that you shoot. Check published papers and know what the result from your sample should look like. If something is wrong or unexpected, take a note and be mindful. If you can solve the problem right away, do it so. If you find something unexpected and interesting, chase it now. Finding something unexpected and interesting is the best fuel to drive your research well and it is the best part of science. So for you, please get prepared well.
Do extra saves your time
You satisfy the quality. Now just take one more measurements. Having two or three more, for example, TEM images does not hurt. Indeed, in many cases, it saves your time. When preparing your manuscript, more data always helpful.
Be mindful that the data will be used in the end, and what you are doing at the measurement table should be useful in the process. It is more like a personal management task, rather than science task, since measuring same data does not advance science. However, in order to make your work efficiently in the end, doing extra save your time always.
