Is there any evidence for the observer effect, and could that be used to support some form of Idealism over a solely materialist view?
There is a huge diversity of opinion inside quantum mechanics as to what constitutes an observer. But there are standard examples of where it matters. Someone in comments already noted this one http://theobservereffect.wordpress.com/the-most-beautiful-experiment/. A continuous version that gives more of an idea of what 'partially observing' might be is described here http://www.sciencedaily.com/releases/1998/02/980227055013.htm.
More bizarre combinations also occur. For instance, you can split light into polarized beams that then combine in ways you would think their polarization should rule out http://www.colgate.edu/portaldata/imagegallerywww/98c178dc-7e5b-4a04-b0a1-a73abf7f13d5/ImageGallery/photon-quantum-mechanics.pdf)
Many people, including advanced physicists like Roger Penrose [The Road to Reality] and many other interesting thinkers [I loved "What the Bleep"], argue from this for something not reducible to mechanics. This is a stretch to begin with, but [sorry to harp on a pet point] if you adopt a view that allows for reversible time, even the ghost of the implication goes away.
If the particles are not so much 'put in their place' by observation, as 'tied to history' by having traceable effects later in time, then reversible mechanical actions can explain what is going on.
If there is a perpendicular dimension of time which our own time can move independently across, we can imagine a point of view from which our timeline is a complicated curve in space which occasionally reverses for short periods, but always eventually makes progress in one direction.
Boltzmann suggests there is an entropy low-point in 'real time' and we rely on accumulating entropy because of our proximity to that. (Hard to find non-paper references -- See something like http://www.eoht.info/page/Boltzmann+brain+problem?t=anon.) This would suggest our phenomenological time that requires entropy for memory encoding might vary somewhat independently of some more basic 'realer' time along which entropy tends to increase, with minor variations. Our phenomenological time then can be seen as one dimension of time, and these variations happen along a perpendicular across it.
As long as the information is lost, any mechanical process is free to flow back and forth, retracing many possible paths to the future. But if the information about this past is referenced in the future, then only paths that actually get to that future go through this past. Many repetitions and false-starts, that get reversed out before flowing that far forward, are ruled out by choosing to use that information later.
This was proposed seriously by Richard Feynman, but he entertained it rather lightly. In particular, he proposed simplifying our particle model by modelling antiparticles as real particles "caught in the middle of" travelling backward in time http://philsci-archive.pitt.edu/4601/1/antiparticles_v7_classical_archive.pdf.
To my mind, it gives us a less mysterious and less 'ontologically wasteful' approach to explaining quantum indeterminacy than "Many-Worlds" or presumptions that 'observing' has a time-bound meaning, and immediate significance. Instead, observation is simply the future use of information by any surviving process. It does not require an infinite degree of redundancy in representations, only a finite-but-arbitrary one, modeled easily by a second time dimension (in imaginary time http://en.wikipedia.org/wiki/Imaginary_time), or a model of time that is focused directly on entropy, and not explicitly dimensional, and therefore has many more independent dimensions. (Basically, if time had hundreds of dimensions, how could we know? We could only measure the magnitude of the part of time along our experience, and declare the rest independent of that. So macroscopically, multidimensional time looks like complex time.)
So there is at least one interpretation of quantum indeterminacy that can retain materialism. (Even if that is a materialism with multi-dimensional time.)