I present an analysis of techniques used to determine the mass power spectrum from observations of the intergalactic medium. Simple Monte Carlo simulations are presented which clarify some the problems which must be overcome when attempting any inversion process. The 1 dimensional flux power spectrum is dependent on both the clustering of the absorption features and on the scaling relations existing between the dark matter and baryon distributions. More realistic simulated spectra are used to illustrate differnces in the mean flux power spectrum for a range of cosmologies and the ratio of the 1 dimensional linear mass and flux power spectra. The mass distribution is much more dependent on cosmological parameters than the flux power spectrum, highlighting the very accurate measurements of the latter quantity required for an accurate recovery of the mass density power spectrum. These spectra are further analysed by deconvolution into Voigt profiles. This techniques is shown to be an excellent approximation, in spite of the current cosmic web paradigm of the Ly-α forest. However the power spectrum of the positions of these fitted lines is shown to be a very poor indicator of the underlying mass density field. Inferring the 3 dimensional forms of power spectra from 1 dimensional data is shown to be problematic. This, coupled with analysis of the covariance matrix of the flux power spectrum, casts doubts on the reliability of the error analysis presented in the literature.