MODEL REGRESI TOBIT KONSUMSI SUSU CAIR PABRIK (Studi Kasus Rumah Tangga di Provinsi Bali)

Abstract

Regression analysis is used to study the relationship between dependent (response) variable with one or more independent (causal) variables. While response data were censored, then Tobit regression model could be applied.  According to Greene (2003), censored data were data with incomplete observation or the dependent variable has a value of zero, while for the other observations have particular value.  This research aimed to model dairy milk’s consumption from households at Bali Province.  By using data from Survey SosialEkonomiNasional (SUSENAS) or Social Economy’s National Survey (SENS) for year 2012, 615 households were selected as sampling unit using simple random sampling technique, and found 123 households who consumed dairy milk.  The independent variables in our model were last education level completed by head of household’s (X₁), head of household’s work (X₂), age of head of household’s (X₃),  amount of expenditure for food consumption’s (X₄), number of household members (X₅), and household income (X₆), the response variable was budget for buying dairy milk (Y).  From six independent variables, is found only last education level by head household and amount of expenditure for food consumption had siginficant effect on Y’s.  The final Tobit regression model were obtained using AIC (Akaike Information Criterion) method is Y = -3314724 + 565429,7 X₁ + 0,014278 X₄ with pseudo R² as much as 16.79 per cent.