 |
Asia-Pacific Forum on Science Learning and Teaching, Volume 17, Issue 1, Article 2 (Jun., 2016) |
This research is descriptive research utilizing survey method. The data were collected using two instruments, test and questionnaire. Test were used to measure divergent thinking of basic skills of sciences process skills of life aspects on natural sciences subject in elementary school students. Meanwhile, the questionnaire is used to collect information about the characteristics of students and schools.
The tests were developed in three stages. The first stage was divided into two phases. In the first phase, the researcher developed the blue prints of Science Process Skills (SPS). The formulation of SPS blue prints based on SPS blue prints developed by Subali (2009) with reference to Rezba et al. (2007), Bryce et. al. (1990), and Cox (1956). The blue prints were used to construct the test items to measure divergent thinking ability of SPS in Biology subjects for Senior High School students.
The aspects of basic skills which are measured include (a) observing, (b) recording the data/information, (c) following the instructions, (d) classifying, (e) measuring, (f) manipulating the movement, and (g) implementing procedures/techniques/using equipment. With reference to the blue print of SPS, the researchers developed divergent thinking tests for SPS. All items have been judged by experts.
The divergent test scoring techniques were based on Diakidoy & Constantinou model (Kind and Kind (2007). In this model, there categories scale are used, which include category-1 if a testee gets score of 2 (giving two correct answers), category-2 if a testee gets score of 1 (giving one correct answer ), and category-3 if a testee gets score of 0 (giving no correct answer).
The second stage of the research was trying-out the tests, which voluntarily involved 3070 students from 60 elementary schools employing purposive sampling techniques. This sample was taken by considering the minimum size of the sample. The number of ideal samples is at least 500 testees in order that a set of test can be used operationally (Muraki & Bock, 1998). The testees included students of grades 4 and 5. They were tested using 4 sets of test completed with anchor items in order that the data could be placed in a single scale. 783 testees did test I, 764 did test II, 753 did test III, and 760 did test IV.
The third stage was measurement stage. In this stage, the measurement was conducted to 84 elementary schools students selected using a purposive sampling technique by considering the distribution of the schools in Yogyakarta Special Province i.e. the schools in the province, regencies, and districts. There were 14 districts. Each district was sampled 4 public elementary schools and 2 private schools. The number of students from 84 elementary schools was 7867 testees consisting of 2563 testee of grade 4, 2685 testee of grade 5, and 2619 testee of grade 6.
The results of the tests were analyzed by using items analysis which utilizes Quest program (Adam & Kho, 1996) and scaled with polytomous of 3 categories to get information about the student’s ability, the difficulty of items, and validity of each item. The testing of the fitted items to the Partial Credit Model is based on the score of Infit Mean Square (Infit MNSQ), with scores ranging from 0.77 to 1,30, therefore each item can be declared "valid" (Wright and Masters, 1982). Quest program can also be used to gain the test reliability based on the value of measurement errors and internal consistency. The measurement errors show the reliability of estimate from persons estimate (Adam & Kho, 1996).
