Koperasi Pegawai Republik Indonesia (KPRI) Jelita memiliki anggota dengan kemampuan membayar angsuran yang beragam, sehingga status risiko kreditnya perlu diidentifikasi. Penelitian ini bertujuan menganalisis data anggota, memberi label status risiko berdasarkan Debt Service Ratio (DSR), serta memetakan keterkaitan karakteristik sosial-ekonomi dengan status risiko kredit menggunakan Firth penalized logistic regression. Data yang digunakan berjumlah 121 anggota yang dihimpun dari kuesioner serta catatan administrasi simpan pinjam koperasi. Pelabelan memakai ambang DSR sebesar 30%, sehingga diperoleh 63 anggota (52,07%) berkategori Berisiko dan 58 anggota (47,93%) Tidak Berisiko. Model dibangun dengan delapan variabel bebas dan rasio events per variable sebesar 7,25. Dari kedelapan variabel tersebut, hanya biaya rumah tangga yang berhubungan signifikan pada taraf 5% dengan odds ratio sebesar 0,7020 dan arah negatif, sedangkan biaya transportasi serta skor status rumah menunjukkan kecenderungan positif meskipun belum signifikan. Model memiliki kalibrasi yang memadai menurut uji Hosmer-Lemeshow, tetapi daya pisahnya masih terbatas dengan akurasi sebesar 0,6446 dan area under curve sebesar 0,6762. Validasi internal bootstrap memperlihatkan penurunan kinerja setelah koreksi optimisme, dengan AUC turun menjadi 0,5926. Secara keseluruhan, penelitian ini memberikan gambaran arah keterkaitan karakteristik sosial-ekonomi dengan status risiko kredit sekaligus menghasilkan data status risiko berbasis DSR pada anggota KPRI Jelita sesuai ruang lingkup penelitian.

The Republic of Indonesia Employees Cooperative (KPRI) Jelita has members with varying capacities to repay loan installments, so their credit risk status needs to be identified. This study aims to analyze member data, assign credit risk status based on the Debt Service Ratio (DSR), and map the association between members' socio-economic characteristics and credit risk status using Firth penalized logistic regression. The data comprised 121 members compiled from questionnaire responses together with savings-and-loan administrative records. Labeling used a DSR threshold of 30%, yielding 63 members (52.07%) classified as At Risk and 58 members (47.93%) as Not At Risk. The model was built with eight predictor variables and an events-per-variable ratio of 7.25. Of these eight variables, only household expenditure was significantly associated at the 5% level, with an odds ratio of 0.7020 in a negative direction, while transportation cost and home-ownership score showed positive tendencies that were not yet significant. The model was adequately calibrated according to the Hosmer-Lemeshow test, but its discriminative ability remained limited, with an accuracy of 0.6446 and an area under the curve of 0.6762. Internal bootstrap validation showed a performance decline after optimism correction, with the AUC dropping to 0.5926. Overall, this study describes the direction of the association between socio-economic characteristics and credit risk status and produces a DSR-based credit risk status dataset for KPRI Jelita members within the scope of the study.