The purposes of this study are: (1) to know if students' mathematical problem-solving ability taught by contextual learning model is higher than students taught by expository learning, (2) to know if students' selfconfidence taught by contextual learning model is higher than students taught by expository learning, (3) to know if there is interaction between learning model and students' early mathematical ability to improve students' mathematical problem-solving ability, (4) to know if there is interaction between learning model and students' early mathematical to improve students' self-confidence. This study is a quasi-experimental research. The population in this study consists of 180 students in grade VIII SMP Muhammadiyah 11 Pangkalan Brandan. Two classes (60 students) are taken as sample. Data were analyzed by two way Anova. The results of this study indicate that (1) students' capability of solving mathematical problems taught with contextual learning model is higher than students taught by expository, (2) students' self-confidence taught by contextual learning model is higher than students taught by expository, (3) there is interaction between learning model and students' early mathematical ability to improve students' mathematical problem-solving ability, (4) there is interaction between learning model and students' early mathematical to improve students' self-confidence.

This study aims to improve the ability of high-level thinking by developing learning models based on problems in senior high school students. The type study is research development. The subject of dissemination consists in 3 distric/city in North Sumatera, namely: SMK Negeri 6 Medan, MAN Deli Serdang Distric and SMA Yapim Taruna Langkat Distric, SMA YPK Medan City. Instrument of Collecting data used are observation guidelines, questionnaires, interview guides, students mathematics text books for 11 th grade in Senior high school, Teacher's guide book, instrument of pre-test and post-test. Development of model is adopted from Thiagarajan's model and Semmel & Semmel's model. This study has compiled teaching materials in the form of textbooks for students' and teacher's guide book that includes the structured steps of solving mathematical problems based on problem solving which can construct high-level thinking. Results of dissemination showed a significant improvement of students' problem solving ability in three districts/cities in North Sumatra at four schools.

The students' difficulty which was found is in the problem of understanding, drawing diagrams, reading the charts correctly, conceptual formal mathematical understanding, and mathematical problem solving. The appropriate problem representation is the basic way in order to understand the problem itself and make a plan to solve it. This research was the experimental classroom design with a pretestposttest control in order to increase the representation of visual thinking ability on mathematical problem solving approach with contextual learning. The research instrument was a test, observation and interviews. Contextual approach increases of mathematical representations ability increases in students with high initial category, medium, and low compared to conventional approaches.

Abstract-This research is a qualitative descriptive research. Data analysis has been used is Miles and Huberman Methods.The result of the research shows that: the level of ability of mathematical creative thinking is at the level of 'medium' creative thinking ability; at the preparatory stage, students have prepared themselves to solve problems by learning to think, seek answers; at the incubation stage, searching and gathering data/information; at the stage of illumination, raises new ideas; until at the verification/evaluation stage; the analysis of the difficulty of the mathematical creative thinking process of this research is the difficulty in applying the principles and problem solving. Then, there is the inability to detail problem-solving characterized by the difficulty of principles and procedures that include the inability to plan the settlement, the inability to engage in the discovery, the inability to abstract patterns, the inability to express the meaning and not apply the principle. In addition, the inability to give many ideas, the inability to solve problems from different perspectives, the inability to solve problems in one's own way, and the inability to develop or detail in a situation; students activity are effective so that learning by applying Problem Based Learning Model provides an opportunity for students to explore various answers and ways of completion with attention to aspects of fluency, flexibility, originality, and elaboration.

This research study is research and development learning devices. This study aimed to describe how the validity, practically, and effectiveness of learning devices oriented of problem based learning which is developed and knowing about increase students' combinatorial thinking in mathematical problem solving ability after using learning devices which is developed. The product that produce in this study is lesson plan, handbook's teacher, student's book, and worksheet. Learning devices development using 4D model which developed by Thiagarajan, Semmel and Semmel with four step, that is define, design, develop and disseminate. This study was conducted in two trials in two different class. Data was analyzed descriptive and result of research showed the devices was categorized valid both in terms of content and construct, practical to use, and effective and students' combinatorial thinking in mathematical problem solving ability increased by using learning devices which developed. The increasing of students' ability seen from the result of test in trial I and trial II were increase about 6,67%.