Working on a ‘2d’ animation- basically moving and key framing individual 2d image mesh planes in the 3D viewport to make what I guess could be called a 2D puppet animation.
For one scene, I want the objects to be larger and rotated from their original positions, and for these changes to be applied across all existing key frames, without having to change the keyframes entirely. It’s a basic head bopping animation of a character and I want them to be closer to the viewer with their head bopping the same if that makes sense. Any way to do this without getting a headache? First time posting on here and not blender savvy so sorry if I mess something up.
Welcome! 
You can do it either in the graph editor by selecting all the keyframes you want to move, and dragging them up or down to change their values together the same amount. I’m not very good at using graph editor myself, maybe someone else will give you more detailed instructions on doing it. 
Another way is to parent the planes to an empty that you can then move, rotate and scale and keyframe (if you are not using a rig). If you are using a rig to control the planes you can add a new bone and parent the head bone to it, then parent the new bone like the head bone was parented before. You will then have a new bone for moving and scaling the head animation. If you use this parenting technique just make sure to first add a keyframe both before and after the frames you want to change with it, so you will have it in default position at other times. You must ofc take the interpolation into account with this technique.
This will be much easier if you get the Animation Layers addon. You can add a second layer on top set to Combine with a single key frame with your new values, and it will just work like magic
To add to what has been said.
In the graph editor, you can select all the key frames through time you want to change on a channel. For example scale.
And then use the move tool of the Graph Editor (y key I think) up or down this will change all the relative values.
This method of constraining movement on the vertical axis is how you can adjust all the values on a channel rather than their positions in time. And do it all at the same time to keep their relative values.
Same with the rotation value.
However this is more or less the brute force way to do it. It’s permanent and each time you want to change it, you’ll have to go through this process.
The benefit of this method is that it keeps the animation intact. And there is no added complexity.
The draw back as far as scaling is that it will scale the individual objects in their local space which means the spacial relationship will change. Less space between them.
To maintain the relationship you will need to scale them in world space with a parent.
For parenting, I would use a Child Of constraint which gives you more control.
If you are planning to make these kinds of changes often, building a rig with this kind of structure would be better from the start.
And if you are not using bones, you should be. With multiple objects, you have an action for each. With bones, the entire animation can be stored in one action. Much easier to edit.
Animation Layers that were mentioned is an essential tool for any animator.
“Don’t leave home without it”
Best of luck!