It's worse than this -- it's often very difficult (expensive, time consuming, etc) to generate the sample being run on the gel, the impact of using a bad or misrepresented antibody are significantly higher than the cost of the antibodies alone.
There's no "collective delusion" here. There is a long-established tradition that formal scientific writing should avoid use of first-person pronouns in general because it makes findings sound more subjective. It's taught this way from early on.
This is slowly starting to change, but it's still pretty much the rule.
Not to be rude, but given current daily attacks on science and the scientific method, I can't let this stand - I think your meta intuition represents a fundamental misunderstanding of how science works.
It is simply wrong to think that scientific questions can never be definitively settled. Clearly there are some hypotheses that have been difficult (and may be impossible) to prove, for example, Darwin's idea that natural selection is the basis of evolution. There's ample correlative evidence in support of natural selection, but little of the causal data necessary for "proof" (until perhaps recently). In the case of evolution the experiments required to prove that natural selection could lead to systematic genetic change were technically challenging for a variety of reasons.
In the case of climate change, the problem again is that the evidence is correlative and not causal. Demonstrating a causal link between human behavior or CO2 levels and climate change (the gold standard for "proof") is technically challenging, so we are forced to rely on correlations, which is the next best thing. But, you are right, it is not "proof".
Establishing causality can be difficult but not impossible - the standard is "necessary and sufficient". You must show necessity: CO2 increase (for example) is necessary for global warming; if CO2 remains constant, no matter what else happens to the system global temperatures remain constant. And you must also demonstrate sufficiency: temperatures will increase if you increase CO2 while holding everything else constant. Those are experiments that can't be done. As a result, we are forced to rely on correlation - the historical correlation between CO2 and temperature change is compelling evidence that CO2 increases cause global warming, but it is not proof. It then becomes a statistical argument, giving room for some to argue the question remains "unsettled".
My point is that there are plenty of examples in science where things have been proven -- DNA carries genetic information, DNA (usually) has a double stranded helical structure, V=IR, F=Ma, etc. And there are things that are highly likely, but not "proven", e.g., human activity causes of climate change.
While some of the issues you bring are remain unproven, what's really absurd is to think that no scientific questions can be settled.
Sounds nerdy and specialized, but Abelson & Sussman's SICP; 1st read an early draft in a CS class ~1985 and it changed the way I thought about computers, programing, science and problem solving forever...