We investigate the scaling properties of the long-range temporal evolution and intermittency of SDO/AIA intensity observations in four solar environments: an active region core, a weak emission region, and two core loops. We use two approaches: the probability distribution function (PDF) of time series increments, and multifractal detrended fluctuation analysis (MF-DFA). Noise taints the results, so we focus on the 171 Angstrom waveband , which has the highest signal-to-noise ratio. The lags between pairs of wavebands distinguish between coronal versus transition region (TR) emission. In all physical regions studied, scaling in the range 15-45 min is multifractal, and the time series are anti-persistent on the average. The degree of anti-correlation in the TR time series is greater than for coronal emission. The multifractality stems from long term correlations in the data rather than the wide distribution of intensities. Observations in the 335 Angstrom waveband can be described in terms of a multifractal with added noise. The multiscaling of the EUV data agrees qualitatively with the radiance from a phenomenological model of impulsive bursts plus noise, and also from ohmic dissipation in a reduced magnetohydrodynamics (RMHD) model for coronal loop heating. The parameter space must be further explored to seek quantitative agreement. Thus, the observational signatures obtained by the combined tests of the PDF of increments and the MF-DFA offer strong constraints which can systematically discriminate among models for coronal heating.