Penelitian ini dilaksanakan pada tanggal 05 oktober sampai tanggal 05 november 2015 di smp negeri 3 tarakan. jenis penelitian ini adalah kuasi eksperimen dengan desain penelitian nonequivalent control group design. populasi dalam penelitian ini adalah siswa kelas viii dan sampel sebanyak 2 kelas, pengambilan sampel dilakukan dengan teknik purposive sampling. analisis data menggunakan bantuan spss 22.0 berdasarkan hasil analisis manova hasil uji test of between-subject effects menunjukkan bahwa nilai signifikansi variabel hasil belajar aspek pengetahuan kurang dari 0,05 yaitu sebesar 0,001 sehingga model pembelajaran problem based instruction berpengaruh terhadap hasil belajar aspek pengetahuan. hasil uji test of between-subject effects menunjukkan bahwa nilai signifikansi variabel hasil belajar aspek keterampilan kurang dari 0,05 yaitu 0,038 sehingga model pembelajaran problem based instruction berpengaruh terhadap hasil belajar aspek keterampilan. hasil uji test of between-subject effects menunjukkan bahwa nilai signifikansi variabel hasil belajar aspek sikap kurang dari 0,05 yaitu 0,027 sehingga model pembelajaran problem based instruction berpengaruh terhadap hasil belajar aspek sikap.

This study was conducted from october 5th 2015 to november 5th 2015 at smp negeri 3 tarakan. this study was quasi experiment of nonequivalent control group design. the population of this research was eight grade students and the samples were two classes, which were taken by purposive sampling. the data was analized by using spss 22.0. based on manova analysis, the result showed that the significancy of learning achievment in cognitive aspect was lower than 0,05 which was 0,001. therefore problem based instruction learning model was effective toward learning achievment of cognitive aspect. while the significancy of learning achievement of psychomotor aspect showed was also lower than 0,05 which was 0,038. therefore, problem based instruction learning model was effective for improving student achievement on learning biology in psychomotor aspect. for affective aspect, the result also showed that the significancy was lower than than 0,05, which was 0,027. therefore, problem based instruction learning model was effective to improve student achievment in affective aspect.