چکیده (انگلیسی):

The present essay was intended to compare between two methods of estimation, namely, Maximum
Likelihood (ML) and Diagonally Weighted Least Squares (DWLS) in determination of construct validity of
achievement goals. The statistical population of this investigation includes all high school third grade students in
the fields of empirical sciences and math- physics from Isfahan City during academic year 2005-6, that they were
totally 7278. 400 students were chosen by means of stratified sampling technique with proportional assignment.
To measure achievement goals, two subscales of mastery goals and performance goals were used that have been
exclusively prepared to utilize in the field of mathematics by Middleton and Midgley. The Cronbach alpha
coefficients for variables of mastery goals were obtained as 0.76 and 0.82 respectively. The results showed that if
the observed variables are ordinal and their categories are lesser than 5, then it is important to select DWLS
method as the best choice for estimation before estimating parameters and review of data distribution with
respect to type of data. If researcher does not checkup assumptions and adapts default estimation method (ML) so
the results may be biased and this may lead to making the erroneous decisions.