Quality Counts
Evidence Based Materials
The validity, reliability and effectiveness of the Gallistel Ellis Test of Coding Skills has been extensively studied by the University of Minnesota, where more than 3000 testings were done on more than 2000 pupils. The pilot and developmental versions of the GE Test were used in several major evaluation and prediction studies in the public schools.
The reliability of the test is very high. (Reliability coefficients are estimates of the extent to which the same score would be obtained if the test were given again to the same pupil on another day before new learning had taken place. They indicate the confidence that can be placed in a test score.)
Split-half reliability coefficients for the total score on the GE Test ranged from .940 to .992 in four separate samples of pupils in low reading groups in grades 1-6 in two public schools in Minneapolis.
The test/retest reliability coefficients were .985 for the Reading subtest and .981 for Spelling.
Since the GE Test is a criterion referenced test, its ultimate validity rests upon its ability to substantiate such statements as “Can read any word with closed syllables and single consonants.” Or “Can spell any word with vowel-R combinations.” It provides the teacher with concrete information concerning specific skills which have been learned, or need to be learned.
The validity of the GE Test is also supported by its high correlations with the Wide Range Achievement Test (WRAT) which measures word recognition of the words usually contained in leveled readers in the elementary grades. Studies indicated that the correlation between the GE Test and the WRAT ranged from .707 to .936 on the reading subtest and from .636 to .884 on the spelling subtest, with different samples of students in the low reading groups at the end of first grade.