Tree growth is an imperfect recorder of past climate, because trees are also affected by non-climatic factors including wind and insect damage and competition from nearby trees. These effects are minimized by collecting samples from numerous trees and averaging them together. The resulting reconstructions are validated with rigorous statistical analysis, and not used unless statistically valid. There has never been any manipulation of data to skew results.