What jumps out at me is Mom's hand which is bigger than baby's face and her bare arm. In any people shot any brightly lit skin competes for attention with the faces. In a candid situation like that grab the snapshot then look for distractions like those and refine it if possible by coaching the subject to move the hand, look at camera or down at baby, etc.
Using what I call "Inside-Out" cropping I identified what was most important (faces), cropped in tight on it, then expanded the frame outwards until I found distractions entering the frame. To make the striped shirt less distracting I desaturated it and edit the lighting on the faces a bit. Not perhaps the best crop to tell the complete story but one I think that better delivers it's "punchline" more effectively.
A tip for posing babies without the hand getting in the way is to have the parent grab the shirt at the back, which supports the child and controls where the are oriented. In family group shots it minimizes the # of hands seen vs. hands on shoulders, etc. If you click my WWW button it will link to my tutorial site.
