Full Idea
According to anti-realists, scientific theories which posit unobservable entities are underdetermined by the empirical data - there will always be a number of competing theories which can account for the data equally well.
Gist of Idea
Theories with unobservables are underdetermined by the evidence
Source
Samir Okasha (Philosophy of Science: Very Short Intro (2nd ed) [2016], 4)
Book Reference
Okasha,Samir: 'Philosophy of Science: very short intro (2nd ed)' [OUP 2016], p.67
A Reaction
The fancy version is Putnam's model theoretic argument, explored by Tim Button. The reply, apparently, is that there are other criteria for theory choice, apart from the data. And we don't have to actually observe everything in a theory.