A wavelet or time-scale approach to analyzing a single time series and two time series, in which the fluctuating quantities are statistically nonstationary, is presented. The time scale and scale “power spectra” are introduced and utilized to analyze transient potential fluctuations measured at the core of sawtoothing TEXT-U plasmas. The results show features that have not been previously observed using any Fourier techniques. In addition, the linear time-scale “coherence spectrum” is developed to quantify the degree of linear relationship between two nonstationary fluctuating quantities in the time-scale domain. Such a spectrum is also useful in tracking the time-varying phase difference. A numerical example is provided to demonstrate the efficacy of the time-scale spectra.