Instrumental variable estimation is a statistical technique used to determine causal relationships when controlled experiments are not feasible. It helps address issues of endogeneity, where an independent variable is correlated with the error term, potentially biasing results. An instrumental variable is a third variable that influences the independent variable but does not directly affect the dependent variable. To use this method, researchers identify a suitable instrument that meets two criteria: it must be correlated with the independent variable and must not be related to the dependent variable except through the independent variable. This approach allows for more accurate estimation of causal effects in observational studies.