KOMPARASI ANTARADMINISTRASI ASSESSMENT OF LEARNING TERSTANDAR DENGAN WAKTU RESPON BUTIR YANG DIMODELKAN PADA COMPUTERIZED ADAPTIVE TESTING

HANDARU CATU BAGUS, . (2022) KOMPARASI ANTARADMINISTRASI ASSESSMENT OF LEARNING TERSTANDAR DENGAN WAKTU RESPON BUTIR YANG DIMODELKAN PADA COMPUTERIZED ADAPTIVE TESTING. Doktor thesis, UNIVERSITAS NEGERI JAKARTA.

	Text COVER.pdf Download (1MB)
	Text BAB 1.pdf Download (299kB)
	Text BAB 2.pdf Restricted to Registered users only Download (535kB) | Request a copy
	Text BAB 3.pdf Restricted to Registered users only Download (299kB) | Request a copy
	Text BAB 4.pdf Restricted to Registered users only Download (2MB) | Request a copy
	Text BAB 5.pdf Restricted to Registered users only Download (138kB) | Request a copy
	Text DAFTAR PUSTAKA.pdf Download (299kB)
	Text LAMPIRAN.pdf Restricted to Registered users only Download (729kB) | Request a copy

Abstract

Penelitian ini memodelkan penilaian tes prestasi belajar dengan Computerized Adaptive Testing (CAT) dengan setiap butir soal dibatasi waktu responnya. Tujuan dari penelitian ini yaitu Menganalisis lebih mendalam terhadap perbedaan efisiensi dan efektifitas, presisi dan akurasi terhadap Administrasi AoL terstandar yaitu antara administrasi model UNBK, model CAT yang mengkonstrain waktu respon butir dan model CAT tanpa mengkonstrain waktu respon butir. Metode penelitian ini yaitu kuantitatif dengan mencari correlational dengan merumuskan regresi linear antara waktu respon butir sebagai variabel terikat dan variable bebas yaitu tingkat kesukaran dan frekuensi kunjungan butir. Metode berikutnya yaitu causal-comparative antar model AoL terstandar yaitu administrasi model UNBK, model CAT yang mengkonstrain waktu respon butir dan model CAT tanpa mengkonstrain waktu respon butir. Uji efisiensi dan efektifitas dilakukan dengan membandingkan hasil akumulasi waktu respon butir soal dan membandingkan jumlah butir yang dibangkitkan. Uji presisi dilakukan dengan menghitung rerata dari mean absolute error (MAE) terhadap butir soal yang dibangkitkan disesuaikan hasil estimasi kemampuan penempuh tes. Uji akurasi dilakukan dengan membandingkan estimasi kemampuan penempuh dengan true score, menghitung Mean Square Error (MSE) dari selisih hasil estimasi kemampuan penempuh tes dan true score.Model CAT konstrain waktu respon butir lebih efisien, efektif dan presisi dibandingkan dengan model lainnya. Sebab penempuh dalam pengerjaan tesnya lebih cepat selesai yaitu rerata 37 menit dibandingkan model lainnya diatas 40 menit. Jumlah butir yang ditempuh oleh penempuh tesnya rerata kurang dari 12 butir sudah dapat mengestimasi kemampuan penempuh tes. Rerata MAE lebih kecil yaitu 0,41566 dibandingkan dengan model CAT tanpa konstrain waktu respon yaitu 0,51271. Model CAT tanpa waktu respon lebih akurat dengan hasil MAE lebih rendah yaitu 0.33626 dibandingkan dengan model CAT dengan waktu respon butir dengan MAE yaitu 0.4250. Jadi AoL terstandar yang dimodelkan CAT secara ideal akan tepat sasaran manakala bank soal memiliki jumlah butir banyak dan tersebar ke seluruh tingkat kesukaran butir soal agar butir soal yang dibangkitkan dapat menyesuaikan estimasi kemampuan penempuh tes. Kata Kunci : Computerized Adaptive Testing, Waktu respon Butir, Assessment of Learning ******** This study modeled the assessment of learning achievement tests with Computerized Adaptive Testing (CAT) with each item limited in response time. The purpose of this study was to analyze more deeply the differences in efficiency and effectiveness, precision and accuracy of standardized AoL administration, namely between the administration of the UNBK model, the CAT model that limitted in item response time, and the CAT model without limmited in item response time. This research method was quantitative by finding correlational by formulating a linear regression between the item response time as dependent variable and the independent variable, namely item difficulty and frequency of item visits. The next method was causal-comparative between administration of standardized AoL models, UNBK model the CAT model that limitted in item response time, and the CAT model without limmited in item response time. Efficiency and effectiveness tests were carried out by comparing the results of the accumulated items response time and comparing the number of items raised. The precision test was carried out by calculating the average of the mean absolute error (MAE) of the raised items appropriate with the estimated testee ability. The accuracy test was carried out by comparing the estimated testee ability with true score, calculating the Mean Square Error (MSE) from the difference between the results of the estimated testee ability and true score. CAT models constrained item response time more efficiently, effective, and precisely than other models. Because the testee in the test work was finish faster, which was an average of 37 minutes compared to other models above 40 minutes. The number of item was taken by the testee had an average of fewer than 12 items, which could already estimate the ability of the testee. The average MAE was more petite at 0.41566 compared to the CAT model without limmited in item response time of 0.51271. The CAT model without response time was more accurate with a lower MAE result of 0.33626 compared to the CAT model with an item response time with an MAE of 0.4250. Nevertheless, the standardized AoL modeled by CAT would ideally be right on target when the question bank has a large number of items and was spread throughout the difficulty level of the question items so that the raised items could be tailored the estimated testee ability. Keywords: Computerized Adaptive Testing, Item response time, Assessment of Learning

Item Type:	Thesis (Doktor)
Additional Information:	1). Prof. Dr. Awaluddin Tjalla, M.Pd. ; 2). Prof. Dr. Burhanuddin Tola, M.A.
Subjects:	Pendidikan > Teori, Penelitian Pendidikan
Divisions:	PASCASARJANA > S3 Penelitian Dan Evaluasi Pendidikan
Depositing User:	Users 15838 not found.
Date Deposited:	05 Sep 2022 05:40
Last Modified:	05 Sep 2022 05:40
URI:	http://repository.unj.ac.id/id/eprint/34645

Actions (login required)

View Item