An explanation of the statistical problems encountered in economics education research with using ordinary least squares (OLS) rather than two-stage least squares (TSLS) in the context of simultaneous structural equations is provided. When endogeneous variables are simultaneously related in a system of equations, alternative statistical estimation procedures, such as TSLS versus the traditional use of OLS, are required. A demonstration of the OLS and TSLS estimation procedures is offered, using as an example the possible simultaneous relationship between economics achievement and attitudes towards economics which has been extensively discussed in the economic education literature. Also reviewed are alternative models to a simultaneous specification suggesting that the achievement-attitude relationship is one requiring joint product rather than simultaneous modeling and estimation procedures. (RM)