nosepot,
You asked whether bodily length-contraction is a real physical effect.
Now, I could have just said that the moving space "that the moving pencil is at rest in" length-contracts, and so so too does the pencil ... because its atoms contract right along with the space they are at rest with. There is space between atoms, beyond atoms, and inside atoms. The size of the atoms are defined by a plot of locations in their spacetime system, and that spacetime system length-contracts when in motion, and so so too do the atoms and hence the body length. However, this response is somewhat insufficient IMO. Time desynchronisation cannot be ignored in any good explanation of the realness of bodily length-contraction, otherwise the meaning of it all is missed. Considering both time desynchronisation in unison with bodily length contraction allows one to undertand why the moving body REALLY IS length contracted per the stationary observer (because it's measurable, and the math requires it so), while at the same time possesses an understanding as to how the moving body is also always its original proper length per itself (because it's measureable per a ruler at rest with the body).
No body ever changes in and of itself simply because it changes in its own state of motion, or becomes gazed upon by relatively moving others. Relative velocity produces the relativistic effect of length contraction of moving bodies, as recorded by the inertial spectator. However, how does it contract if it also never changes in and of itself? ...
All relativistic effects arise and vanish in unison, when relative motion arises and vanishes. Bodily length-contraction is not the only relativistic effect. We may envision the moving body to have clocks affixed eveywhere inside synchronised with each other. Per the stationary observer, the body's fwd clocks will lag in time readout wrt aftward clocks. That's called time desycnrhonisation, another relativistic effect. Each atom along its length exists in a different time era of the moving body. The desynchronisation arose from v>0 just as length contraction did, and as it turns out, the length contraction cannot be fully explained without it.
In analogy, we might image a pencil at rest in a spacetime system S. An observer off the side of the pencil holds a ruler parallel to the pencil and records its length, L. Next, the pencil is rotated angularly thru (say) 60 deg, and the pencil appears shorter per the observer. He's told he cannot rotate his ruler to align it with the pencil-axis to verify if the pencil's length really changed or not. Deep down, from everyday experience, he knows the pencil length now spans a depth dimension it did not span before, and that the pencil still has its proper length of L. All micro wonder clocks affixed everywhere to the pencil always remain synchronised per both the pencil and the observer. However ...
In the case of relativity, the spacetime systems of the moving pencil and the observer (who measures it) possesses an angular orientation differential in the fused 4 dimensional spacetime continuum. That is, the 2 frames are angularly rotated wrt one another, called frame rotation, and that drives all the relativistic effects. The pencil & observer's systems are angularly rotated (in spacetime) wrt each other, because a relative velocity exists between them. Therefore, the observer cannot just turn his ruler to realign it parallel (in 4d) with the pencil, but instead must accelerate to the pencil's state of motion to become at rest with it, and then his ruler is again parallel with the ruler's length axis in 4d spacetime. OK, so the moving pencil is angularly rotated in 4d spacetime wrt the observer (and his ruler), and so the ruler measures a shorter pencil length, ie a length contracted body. Now then, is the ruler truly contracted in length, or is it an illusionary effect of sort? ...
The math requires the contracted length be real, just as time dilation, time desynchronisation, and all other relativistic effects are real. Light signals (theoretically) could verify this, technology permitting. If it's measureable, its a real physical effect. If the math requires it to exist as such, it must be real. As such, the moving contracted body length is a real physical effect ... even though said body never changes in and of itself as it goes, as it accelerates, or if moving others happen to gaze upon it. It's proper length never changes, and that's just as real and physical. Bottom line, it is not required that a moving contracted length be non-physical simply because it also "always holds itself" at its original proper length. All POVs are equally correct, and each their own measurements are equally real "to them". It's a (real physical) relativistic effect, not an illusionary (or optical) effect.
Hope that helps.