本論文提出一具有測量誤差的自我選擇模型(self-selection model with measurement errors),將由問卷所得到的個人資料(truncated sample)及普查(census)資料合併成為一非完全截斷性的樣本(censored sample)。在此非完全截斷性的樣本中,普查的資料被視為未回答問卷者的實際資料加上測量誤差。在少許假設下,本論文所提出的具有測量誤差的自我選擇模型可用最大概似法(maximum likelihood)加以估測而得到具備一致性的參數估測值。 經由蒙地卡羅實驗法(Monte Carlo experiments)的驗證,本論文所提出的具有測量誤差的自我選擇模型確實帶來相當好的估測結果。
A self-selection model which combines micro survey data with subaggreg atedcensus data is proposed in this study. In the model, a truncated sample is transferred into a censored sample by combining survey micro data with sub-aggregated census data which provides information for non-respondents' neighborhoods (e.g. census blocks). Based on the censored sample, two ML estimators are derived where census data are treated as if they are the true values plus errors. Results from Monte Carlo experiments suggest that the self-selection model which combines micro survey data with sub-aggregated census data performs well. The results also indicate that if the self-selection model is correctly specified, adoption of the proposed model will not contaminate the original truncated sample.