It's kind of funny because two concepts are merged in that statement
Tests (logical rational statistically proven) and Jungian concepts (partially provable or unprovable concepts)
Tests can easily be analysed with factor analysis to show whether statements on a particular test are related, accurate and predictable using stats. I recall a Pearson psych text book claiming that some of the MBTI tests were highly reliable, yet most people still emphasised to go by the cognitive functions that Jung clearly pulled from his Ni.
I agree overall, to me the most simplest way to type is to read a description of each attribute - sensing is ...., intuition is ..... then narrow it down. Basically I find it funny that even though there are MBTI tests that are backed up by factor analysis people are basically saying... well just go with the Ni, what does it feel like to you? It makes sense, the unconscious is called unconscious for a reason it's not fully of rational logical thinking but feely, hunchy abstractions and mythology. Go with your gut I guess.
Anyway just a ramble. I guess I could see some INTJ's on here making a counterargument.