Bias in Matter Power Spectra ?

We review the constraints given by the linear matter power spectra data on cosmological and bias parameters, comparing the data from the PSCz survey (Hamilton et al., 2000) and from the matter power spectrum infered by the study of Lyman alpha spectra at z=2.72 (Croft et al., 2000). We consider flat--$\Lambda$ cosmologies, allowing $\Lambda$, $H_0$ and $n$ to vary, and we also let the two ratio factors $r_{pscz}$ and $r_{lyman}$ ($r^2_i = \frac{P_{i}(k)}{P_{CMB}(k)}$) vary independently. Using a simple $\chi^2$ minimisation technique, we find confidence intervals on our parameters for each dataset and for a combined analysis. Letting the 5 parameters vary freely gives almost no constraints on cosmology, but requirement of a universal ratio for both datasets implies unacceptably low values of $H_0$ and $\Lambda$. Adding some reasonable priors on the cosmological parameters demonstrates that the power derived by the PSCz survey is higher by a factor $\sim 1.75$ compared to the power from the Lyman $\alpha$ forest survey.