You will find below the question about D-Efficiency in English.   Hello,   I don't understand how SAS calculates D-efficiency. Indeed, when I manually calculate D-efficiency step by step, I don't get the same result as SAS. However, I'm using the same formula as SAS and the same design coding, namely orthogonal coding.   The article I'm referring to in my calculations is the following: https://documentation.sas.com/doc/da/pgmsascdc/v_040/qcug/qcug_optex_details31.htm which uses the exchange algorithm of Mitchell and Miller (1970) (EXCHANGE) or Mitchell (1974a) for DETMAX.   For the design coding, I follow the same approach as described in Warren F. Kuhfeld article (p. 74) (https://support.sas.com/techsup/technote/mr2010c.pdf) (p. 63 for D-Efficiency).   To illustrate this, in my message above you will find my design plan X and how I processed (the orthogonal coding of X, inverse of (X’X))). The determinant of (inv(X’X)) is equal to 3.05176E-06, the number of parameters to be estimated, which is equal to the number of columns of X, is 7 (=p), and the number of rows is 8 (=ND). When I apply the D-Efficiency formula, I have a D-Efficiency equal to 76.70637702. However, in SAS, I have a D-Efficiency equal to 59.4604 when I run an optex proc on this design with the method=EXCHANGE or DETMAX.   Why is my D-Efficiency value not equal to SAS result?   I thank you in advance for your answers. ... View more