No one will ever see an isolated quark: the theory itself says quarks come only in bound groups. A physicist who thinks quarks are real and one who believes only that the quark model fits every observation run the same calculations, design the same experiments and get the same answers. So does the old argument over whether unobservable things "really exist" change anything?
On settled science, no. C. S. Peirce's pragmatist test is that "the essence of belief is the establishment of a habit": two beliefs differ only in the actions they lead to. His own example is the Eucharist. If the wine keeps every effect wine has on the senses, calling it blood in reality says nothing. The realist and the anti-realist about quarks have identical habits: one belief under two names. Arthur Fine's natural ontological attitude [1] draws the conclusion: accept the confirmed results of science the way you accept your senses, and drop both camps' philosophical additions as idle.
Atoms were once the frontier. Around 1900 no shared habits existed yet: Ernst Mach and Wilhelm Ostwald treated atoms as at most a calculating device; on that view "how big is an atom?" asks the size of a device, not of a thing. Einstein took atoms as real. He later wrote that the major aim of his statistical-physics work then was to find facts guaranteeing "atoms of definite finite size." His Brownian-motion paper predicted how far microscope-visible particles in water should jitter, and said observing it would allow an exact determination of atomic dimensions. Jean Perrin made the measurements from 1908, and Ostwald conceded that the atomic nature of matter had been proved by experiment. Einstein put the skeptics' earlier resistance down to their positivist philosophy.
So Peirce's test cuts both ways. Where results are settled, both stances have produced the same habits, and the debate is idle. At a live frontier the stance is what produces the habits: treating an entity as real turns its properties into research questions, and treating it as a fiction leaves them unasked. Fine's attitude starts from confirmed results, so its verdict holds only where the dispute has already stopped mattering.
References
The Shaky Game [HTML]
Fine, A., 1996. The University of Chicago Press.