To be fair, the pure scientific impact of the robotic missions is probably much higher than human ones (besides advances in medical/bioscience). But, placing people on the moon probably brings more funding for NASA than robots.
To compare the scientific return of robots and people, we can compare the results of the only space rock both have visited: the Moon. The six manned Apollo missions brought back much much much more scientific return than all of the contemporary robotic missions. Apollo brought back 382 kg of moon rocks. Three Soviet probes (first one, Luna 16, between Apollo 12 and 14) brought back a total of 326g. While it's a bit facile to claim that's the whole difference, I would say that the difference in scientific return was at least one order of magnitude, if not quite as large as the moon rock numbers.
Now, the Apollo program also cost much more than the robotic missions. If you are willing to invest enough (e.g. Apollo was >1% of US GDP/year for most of the 60's) you can get an enormous amount of scientific return from a manned mission, but robots are useful for budgets that can't cover a manned mission.
Bear in mind that the Luna missions were done with 1960s Soviet robotics technology... And were only a side-show to their goal of a manned landing, which was hamstrung by repeated launcher problems. (And as soon as they lost the moon race, interest in this immediately dried up on the Soviet side - the sample return in the 70s was an afterthought.)
If your goal is to plant a flag and ship back ~400 Kg of moon rocks, you could do it today, using robots, for a tiny fraction of a manned mission's budget. The thing is, bringing back 400 Kg of moon rocks is not 400 times more valuable than bringing back 1 Kg of moon rocks.
Right, but the lunar science was not just limited to sample return, and here the J missions (Apollo 15-17) with their SIM bay cameras produced much better image quality than even Lunar Orbiter for much of the Moon's surface (Lunar Orbiter 5, in the polar orbit, was able to map parts of the Moon that the J missions never saw.) Similarly, the rover's traveled farther than the Lunkhod's did, showing us a much greater area of the surface. And the most sophisticated scientific instrument ever to go to the moon, even today, would be Harrison Schmitt, with his Harvard Geology Ph.D brain and hands.
As for "done with 1960's technology" so was Apollo: the ability to discover hydrogen (used to find the ice in the lunar crater shadows) wasn't possible with 1960's sensors that were light enough even for the much larger mass and power budgets of an Apollo spacecraft (vis a vis Lunar Orbiter or similar probe).
By the nature of the beast, a spacecraft with humans is going to have absolutely thunderous mass and energy budgets compared to robotic ones, so any instrument you can put on a robot you can put on a manned mission.
In a similar manner, I would expect any manned mission to Mars to employ a lot of robots: control is much easier when the human making the decisions is a few light seconds away versus 8 light minutes. And if you have the mass (and money) for a manned mission, you can tuck in a bunch of robots for very little extra. So a human mission will always be strictly greater than (in scientific return and cash budget) robot mission.
Coincidentally last night I re-watched one of my favorite episodes of "From the Earth to the Moon," "Galileo was Right," which focuses on the Apollo 15 crews getting field training in geology. Their instructor (along with backup LM pilot Jack Schmitt, a geologist, who then flew on Apollo 17) emphasized identifying and collecting the "right" rocks, not just "any" rocks, which led to some of the more interesting samples, including the "Genesis Rock."
Edit: Both are still extremely cool though