Precision measurements of higher order angular galaxy correlations using 11 million SDSS galaxies

We present estimates of the N-point galaxy area-averaged angular correlation functions (omega) over bar (N)(theta) for N = 2,..., 7 from the third data release of the Sloan Digital Sky Survey (SDSS). The sample was selected from galaxies with 18 <= r <= 21 and is the largest ever used to study higher order correlations. The measured (omega) over bar (N)(theta) are used to calculate the projected, s(N), and real-space, S-N, hierarchical amplitudes. This produces highly precise measurements over 0.2-10 h(-1) Mpc, which are consistent with Gaussian primordial density fluctuations. The measurements suggest that higher order galaxy bias is non-negligible, as defining b(1) = 1 yields c(2) = 0.24 +/- 0.08. We report the first SDSS measurement of marginally significant third-order bias, c(3) = 0.98 +/- 0.89, which suggests that bias terms may be significant to even higher order. Previous measurements of c(2) have yielded inconsistent results. Inconsistencies would be expected if different data sets sample different galaxy types, especially if different galaxy types exhibit different higher order bias. We find that early-type galaxies exhibit significantly different behavior than late types at both small and large scales. At large scales (r > 1 h(-1) Mpc), we find that the SN for late-type galaxies are lower than for early types, implying a significant difference between their higher order bias. We find b(1, early) = 1.36 +/- 0.04, c(2, early) = 0.30 +/- 0.10, b(1, late) = 0.81 +/- 0.03, and c(2. late) = 0.70 +/- 0.08. Our results are robust against the systematic effects of reddening and seeing. The latter introduces minor structure in (omega) over bar (N)(theta).