DEVELOPMENT & STANDARDIZATION OF ACHIEVEMENT TEST IN GEOGRAPHY FOR SENIOR SECONDARY SCHOOLS IN BENUE STATE

Abstract

          The present study was motivated by the paucity of standardized assessment tools and serious need for such instrument in the senior secondary schools.  The purpose of this study, was therefore, to develop and standardize a geography achievement test capable of obtaining valid and reliable data in respect of senior secondary school geography achievement.  The study involved 48 selected secondary schools out of which 1,781 senior secondary class three geography students where used for the study.  Random and stratified sampling techniques were employed in drawing up the standardization sample for the study.  A 200-item test was developed (after the first and second trial test) based on the National geography curriculum to answer the seven research questions and five hypotheses postulated for the study.  Data obtained from the test (inform of students’ scores) were analyzed using item analysis, variance, correlational analysis, mean and standard deviations.  The first four hypotheses of the study were tested using multiple classification of Analysis of variance (ANOVA) while the fifth hypothesis was tested using the t-test of significance of correlation coefficient.  All the hypotheses were tested at the .05 level of significance.  The study among other things reviewed that 99% of the item of the standardize Achievement Test in Geography (SATG) where of satisfactory facility and discrimination capacity.  Again that the Average Test facility (0.41) was close to the optimum value (0.50) for an achievement test and Total Test Variance of (73.98) which was sufficiently high.  Also the subscales correlated highly and positively among themselves and individually with SATG.  The reliability of the SATG was 0.94 while those of the subscales range from 0.70 to 0.91.  These values are admittedly high.  The standard error of measurement of the SATG was 6.72 while those of the subscale range from 2.15 to 4.71.  These values are admittedly low.  In addition both grade, gender and location norms obtained for SATG were high.  Gender and school location were significant factors in students geography achievement while content area were found not to be significant.  Gender X content area, and gender X location interaction effect were found to be significant while content area X location; were found not to be significant.  Furthermore, the intercorrelations between the SATG and its subscales were found significant.  These findings have implication for geography teachers and school administrators, recommendations were made limitation of the study were identified and suggestions for further study were made.

target of obtaining a reliability coefficient of at least 0.75 will be granted in the present study.

          Having reviewed the concepts of validity and reliability, it would be necessary to mention some statistical and psychometric measures, which are used to ensure the quality of items to be selected.  Some of the measures which have been used by some researchers include:-

          Item test correlation (Carver, 1974).  Item analysis (Ebel, 1985: Nworgu, 1985; Russel, 1982) and domain specifications (Roid and Haladya, 1980).  The present study made use of item analysis.

Item Analysis

          The application of statistical item analysis techniques constitute an important step in the development of standardized tests.  Anastasi (1961) noted that item analysis makes it possible to shorten a test, and at the same time increase its validity and reliability.  Item analysis is frequently conducted against total score on the test itself.  It is also based on comparison using two contrasting criterion groups composed from the upper and lower 27 percent of the sample on which the items were trial tested.  Nworgu (1985), added that the upper and lower 27 percent groups were chosen because it is the optimum point at which balance is obtained between the sensitivity of the results for normal distribution.  According to Anastasi (1961) experiments have proved that in a normally distributed sample, the optimum group consisted of the upper and lower 27 percent of the cases.  She pointed out however that, if the distribution was flatter than the normal curve, the optimum percentage was slightly greater than 27 percent and approaches 33 percent.

          It is pertinent to note that recommended percentage cut off points for the upper and lower group of achievers vary from one authority to another.  For instance, Ohuche and Akeju (1977) proposed the use of the upper and lower 25 percent of students tested.  Harbor Peters (1999) recommended ¹/₃ of the population of the upper and lower groups of the achievers (i.e 33%).  Anastasi (1961: 433) summed up by positing as follows: “any convenient number between 25 percent and 33 percent will serve satisfactorily”.  The 33 percent of the population of upper and lower group of the achievers as recommended by Anastasi (1961) and Harbor Peters (1999) is applicable to the present study.

          Thus, the qualities of the items that make up a test determine the qualities of the test as a whole, and the assessment of these essential qualities of items in a test constitute item analysis.  These essential qualities are discussed briefly in below.