Background (Introduction)
 
It may be many factors that effect students’ math and science learning. Teaching strategies and methods and students’ learning preferences also are the most important factors of students’ math and science learning. Therefore, responses for the following questions has been searched in this research;
 
Question 1. Are these scores which were taken by primary school students from the factors of the scale towards students’ science learning preferences showing a significant difference according to their grade level? Question 2. Are these scores which were taken by primary school students from the factors of the scale towards students’ math learning preferences showing a significant difference according to their grade level? Question 3. Is there a relationship between primary school students’ science and math learning preferences?
 
Purpose
 
This research attempted to examine primary school students’ math and science learning preferences based on their grade level. Thus, the study was adopted quantitative research approach, where it will then be employed to test hypotheses of the study.
 
Sample
 
Subjects of the research is 679 primary school students who were selected by random and studied at 6th, 7th and 8th grades of primary schools in Sivas in 2004-2005 spring semester.
 
Design and methods
 
The study was used a five-point Likert-type scale which was developed by the researchers. The scale were primary based on the work conducted by Taylor, Fraser and White (1994). Furthermore, there were also benefit from the works done by Knochs and Riggs (1990), Koul and Rubba  (1999), and Riggs and Knochs (1990). It was administered to primary school students for determining primary school students’ math and science learning preferences based on their grade level variable. Exploratory factor analysis showed that the scale included two factors. It contained items measuring primary school students’ individual learning preferences (factor-1) and teacher-centered learning and working with group (factor-2). The results indicated analysis of variance for the entire scale was 42.90%; for each factor, analysis of variance ranged from 26.05% to 16.85%. Factor loading of items in the scale also ranged from .572 to .712. Furthermore, Cronbach Alpha Coefficient is also calculated as .81 for reliability of the entire scale; for each factor, alpha ranged from .891 to .847. Based on the reliability and validity analysis, it showed that there were satisfactory factor structure and reliability of the scale. For determining primary school students’ opinions according to the scale factors were used descriptive statistics methods. In addition, multivariate analysis of variance (MANOVA) and simple correlation analysis were used for the statistical analyses.

Results

At the end of analyzing data, it was found that there was no meaningful relation between grade level and the entire scale towards students’ math and science learning preferences. Similarly, it was also found that there was no meaningful relation between grade level and the factor mentioned above towards students’ science learning preferences. However, it was found that there was meaningful relation between grade level and individual learning factor towards students’ math learning preferences. Furthermore, it was also determined that teacher-centered learning and working with group was preferred more than individual learning in the 6th, 7th and 8th grades students in both the lessons. In addition, the results indicated that there was significant correlation between primary school students’ science and math learning preferences (r =.356, p <.01). In other words, primary school students’ science learning preferences was related to their math learning preferences. 
 
Conclusions
 
Results of the research showed that teacher-centered learning and working with group was preferred more than individual learning in the 6th, 7th and 8th grades students in both the lessons. Furthermore, the results indicated that primary school students’ science learning preferences was related to their math learning preferences. Therefore, in the light of findings of the present research was suggested that there may be common studies and similar students’ science and math learning preferences among national and intermational boundaries according to ages, class levels, and gender and so on. So, differences and similarities should be discovered if we want to understand about students’ science and math learning preferences both national and intermational.