I suspect you have been reading a bit f Ken Rockwell on the topic of the D3200.. and his comments that the 24Mpix 'resolution' of that camera is 'wasted' by the Kit Lens that can't resolve the level of detail that the camera might capture..... (am close?)... but we'll return to that.
A crop sensor is 18x24mm, a full frame sensor 24x36mm; to all extents and purposes its 'Half-Frame', as in half the sensor area. So.. if you have the same sensor resolution, the individual receptors on the bigger sensor will be, err... bigger... which means that they can catch more photons of light, during an exposure, meaning that they can grab a much 'stronger' signal, before processing, which in turn means that the original signal should be 'cleaner' and need less amplification, before being sampled to get the 'digits' a digital camera records. Next up the lens, putting the image on a bigger sensor, does't have to have as much (reverse) magnification or bend light as much, so there's less opportunity for distortion, and from the larger scale, any there is, will tend to be a smaller % of the sensor size... so you get a number of complimentary effects from a bigger sensor, that are 'good' for IQ; This is basic reasons large sensor 'Full-Frame' cameras can deliver better IQ than Crop-Sensor ones, and where they can still maintain that IQ advantage even when they have Pixel counts much lower than on a smaller sensor.
What you asked, indirectly was whether, shooting at 'Medium' 4512x3000 instead of 'Large' 6016x4000,on the D3200, would have the same effect on IQ by effectively reducing receptor density on the sensor, as going to a larger sensor.... And the short answer is
NO, its still a small sensor camera, and the receptors are still the same size, capturing just the same amount of light, provided by a lens with just as much (reverse) magnification...
Elaborating on Phil's comment (While I was compiling this!) Cherry picking some graphics from tutorial
Understanding; Mega Pixels, Mega Bytes; tif, jpg, bmp, nef, & raw?
There is a pixel array, in this case, 60 x 40, the 'grid' that the the camera fills in 'painting by numbers' to make a picture (usually with a lot more squares though!)
And, easy to think, that the camera sensor has a similar grid, with a receptor for each pixel it will make in the photo... Trouble is, to make a colour photo, you need to get three numbers, for each pixel; one for the level of red, one for the level of green, one for the level of blue. Number of ways you could do this, and the simplest would be to have three receptors for each square... which is awkward and reduces your receptor density three fold.. so they don't... they tend to use a honey-comb array, rather than a grid array, a bit like this;-
Shown to a larger scale; but, you can see how you have a honey comb array of alternate red, green and blue receptors, that give pretty much the maximum receptor density; taking the levels from each of these you can then 'interpolate' the three colour values for the square image 'pixels' over lapping the hexagons; one hexagonal colour receptor providing a colour value for perhaps four square picture pixels over-lapping it.. SO.. your 24 Million picture pixels might have bee created from just 18 million receptors, 6 million of each colour, depending on how much 'interpolation' the camera does to make the image file from the sensor readings. If you switch to a lower image resolution, all the camera will do is change the 'interpolation' calculation to make less image pixels from the same recorded sensor values, it wont 'couple' receptors to grab a larger signal like having a bigger receptor.
Back to Ken Rockwell.... He suggested that the 24MPix sensor count on the D3200 was a bit of a gimmick for a entry level camera, to wow bigger is better is more 'consumers', who want bragging rights printed on the box..... and there may be more than a grain of truth in that.... he then went on to say that the kit 18-55 lens it ship with is not really sharp enough to resolve the detail that the sensor can at highest res setting.. which also possibly has more than a little truth behind it.... before he went on to compare to the D3100, and suggest that not trying to process so many pixels, that was the better 'balanced' package, and a D3200 at 'medium' resolution, delivered as much IQ as the lens could provide, while managing the file sizes much more easily.. which is probably also not unreasonable. Its HORRIBLE to actually agree with the chap, if only in part given his reputation... especially as a D3200 owner..... however...
Weak link in the chain, as far as image quality goes, on this camera IS, that kit lens. Its not a BAD lens... it's just not 'great'. I've shot my daughter's 35mm f1.8 on it ad that IS a verry sweet little lens and far sharper than the kit zoom; I have also shot most of my 'Legacy' M42 screw lenses from film cameras on it via adaptor, and again, they show me, that that kit 18-55 and even my 55-300 are no where near as good as they could be.. but for what £70 for the 18-55 and £150 for the 55-300, they are certainly pretty good for the money, and who am I trying to impress here? They are certainly 'adequate' for my purposes at least but, if your looking for IQ gains, that's first place I'd suggest you look, and from discussion on your 'portfolio' post, of where you are at and what you are about, that 35mm f1.8, could be a good way to go on that front, but, that's another topic.
There is good reason NOT to shoot at 24Mpix on this camera; it does make for cumbersome large file sizes, especially in NEF. (about double Jpeg Lage + fine) If you are machine gunning an action sequence, as my daughter was the other week, lying on the floor whilst her brother rode a BMX at and over her (eeeek! with MY camera in her hand!!! EEEk EEK!!).. can hit the buffers pretty quick, and you start dropping shots.. camera also gets rather more than a 'bit' warm. So circumstance dependent, if you can live with a lower resolution, maybe times it cold be useful, but NOT for IQ.
As for the upload speed clearing down the SD card? I rarely find that much of a chore, even using a cheap memory-stick card reader on a USB front port. More significant is keeping the card clean, not having a chunk of already up-loaded pictures on it for the PC to 'read' before moving the new ones, and having a decent, 45Mb/s or quicker card to start with. Smaller cards can actually be a bit of a boon here, demanding you keep-em clean, and not giving PC so much to read before it displays it or shifts it... b-u-t... even a 'dirty' slow 32Gb card on that plug-in reader... plug it in, select all, drag ad drop, leave it to run while you put the kettle on and put the batteries on charge! Its not a real 'problem'.
Will using a lower resolution be detrimental to your image quality? Err... probably not... as said IQ's effected by a whole bag of stuff, not just pixel count... which you will probably be reducing for most display purposes anyway. Daughter does well enough with her D3100 with its mere 16Mpix resolution as its 'best' setting, printing up to A2 size for her school-work..
Reason I bought the D3200 over that, though is 'cropping', where you might chuck away a lot of them pixies before display; and I have a full-round 4.5 fisheye with 180 degree field of view on both axis, making a round image in the middle of the frame, on barely half the available pixels. And with that amount of 'masking' before I might take a square out of the circle the extra resolution is VERY worth while;
But most of the time? Does't make that much odds, and as 99.9999% of everything I ever shoot only ever gets looked at on a monitor, with a 1024x780 ish screen resolution, its ALWAYS going to be down-sized, if not by me, to something around 1000px on the long side for upload, then by the graphic card when it sends the signal to the screen.