PSYC 3250 Chapter Notes - Chapter 7: Standard Error
Document Summary
The standard error of measurement: the reliability coeicient does not reveal how much variability should be expected on the basis of measurement error, standard error of measurement (sem) is a funcion of 2 factors, sem = Not all test scores have an equal sem. Correcion for atenuaion provides an esimate of the correlaion between perfectly reliable measures of x and y. Correcion almost always overesimates the correlaion between x and y. 4 factors afect the reliability of a test: characterisics of people taking the test, characterisics of the test. Spearman-brown formula: used to predict the efect of lengthening the test o new r xx= n old r xx. 1+( n 1) old r xx n = raio increase (ex. double the test, n=2: intended use of test scores, methods used to esimate reliability. Proper methods of esimaing and interpreing test reliability may depend on what the test measures. Reliability of diference scores (pre and post-tests) can pose an issue.