This report consists of a compilation of a number of studies, which all of them have dealt with the effects of practice and instruction upon a person's test results. A literature review based on a dichotomy of practice and instruction under short-term and long-term conditions respectively, as well as on the denign of the investigations and certain characteristics of the individuals, resulted in general as well as in more specific statements on the arising of practice and instruction effects. In all the empirical investigations quantitative-numerical tests were used (KVR and NOG respectively), where one-of them (NOG) is a part of the so-called Swedish Scholastic Aptitude Test. The basic aim was to investigate if short-term practice (KTÖ) and short-term instruction (KTI) could have an effect upon the score of these tests. In this case practice was defined as taking a pre-test, and in the first three studies the instruction consisted of two phases: a first one concerning general testing strategies, and a second one concerning a specially adapted problem solving strategy. As a basis for the forming of this special problem solving strategy for NOG served a logically constructed and sequential problem solving strategy, which was formed as a direct influence on the individual via instruction in the first three studies, and as an indirect influence on the individual via the item format in the last three studies. From the achieved results can be mentioned that the literature paid attention to the importance of a person's knowledge of and familiarity with tests and examinations in a specific as well as in a more general sense. If a person has no or very little former experience of tests there is a certain probability that there will be a higher score due to KTÖ or KTI. This probability is reduced considerably if the person has some or relatively great former experience in this respect. This was also verified indirectly in the empirical studies, which were mainly carried out on individuals who could be classified as experienced in and familiar with tests. No practice or instruction effects could be found either in the first three studies or indirectly in the last three studies. In these ones a specially constructed item format had been used, which had been formed according to the sequential problem solving strategy. With the intention to obtain an interpretation frame for practice and instruction effects on tests, the basis in this report has been a theoretically constructed model, whereby the achieved effects are first of all related to the individual variable former experience, and secondly to whether a test consists of correctly or incorrectly constructed items. / digitalisering@umu

The effect on childrens' scores of different item formats used in standardized mathematics achievement tests was investigated. Second grade students were given a mathematics computation test using formats derived from five standardized achievement tests. Identical content was tested with each format . Differences in test scores between types r0f formats were statistically significant at p(.001 (F = 45.25). These results indicate that what a student appears to know is substantially influenced by the format of the particualar test used in measuring achievement. These differences are not accounted for by the rno~mative scaling of the different tests. Greater attention should be ~i,en to the effect of test item format in selecting and administering cac1 i evemen t tests.