人口遷移與死亡：經濟發展的角色

Abstract

目的：本研究旨在分析臺灣鄉鎮市區人口遷移與死亡率的關係，並探討經濟發展在其關係間所扮演的角色。 方法：本研究資料來自內政部戶政司與財政部財政資訊中心，分析年代為2000-2014年，分析單位為鄉鎮市區。研究自變項為人口遷移，包括遷入率、遷出率、淨遷移率、總移動率與人口變化率等五項指標；依變項為標準化死亡率；第三變項則為經濟發展，包括未報稅家戶比例、低收入戶比例以及綜合所得總額平均數等三項指標。在統計分析上，本研究首先針對標準化死亡率與人口遷移、經濟發展各指標計算皮爾森相關係數。而後選取人口遷移與經濟發展指標中，皮爾森相關係數最大之變項，進行多元階層迴歸分析，探討經濟發展指標加入迴歸模式後，人口遷移指標與標準化死亡率關係的變化。 結果：在皮爾森相關分析中，五個人口遷移指標皆與標準化死亡率呈顯著負相關，其中以總移動率與標準化死亡率的皮爾森相關係數最大，達-0.462（p<0.001）。經濟發展指標中，未報稅家戶比例、低收入戶比例與標準化死亡率為顯著正相關，而綜合所得總額平均數與標準化死亡率則為顯著負相關，以綜合所得總額平均數與標準化死亡率的皮爾森相關係數最大，達-0.662（p<0.001）。其次，在多元階層迴歸分析中，總移動率與標準化死亡率的迴歸模式加入綜合所得總額平均數後，總移動率的迴歸係數由原本的-0.092（p<0.001）改變為-0.016（p>0.05），未達統計顯著水準。 結論：由於加入經濟發展變項後，人口遷移與標準化死亡率的相關性即隨之消失，代表人口遷移與標準化死亡率的關係可能為假相關。

Objective: This study analysed the relationship between population migration and mortality rate in Taiwanese townships, cities, and districts as well as the role of economic development in said relationship. Methods: This study obtained data from the Department of Household Registration (of the Ministry of the Interior) and Fiscal Information Agency (of the Ministry of Finance). The research data comprised data of Taiwanese townships, cities, and districts from 2000 to 2014. In this study, population migration (measured using indicators in-migration rate, out-migration rate, net migration rate, gross migration rate, and population percent change) was set as the independent variable, standardised mortality rate was set as the dependent variable; and economic development (measured using indicators untaxed household percentage, low-income household percentage, and average income) was set as the third variable. Subsequently, this study performed a statistical analysis to calculate the Pearson correlation coefficients between standardised mortality rate and the population migration indicators as well as those between standardised mortality rate and the economic development indicators. Next, population migration and economic development indicators with the highest Pearson correlation coefficients were subject to a hierarchical multiple regression analysis to explore changes in the relationship between standardised mortality rate and population migration indicators after economic development indicators were included the regression model. Results: According to the Pearson correlation analysis, all five population migration indicators were significantly and negatively correlated with standardised mortality rate, among which gross migration rate had the highest coefficient at −0.462 (p < 0.001). Among economic development indicators, untaxed household percentage and low-income household percentage were significantly and positively correlated with standardised mortality rate, whereas average income was significantly and negatively correlated with standardised mortality rate. Among the three indicators, average income exhibited the highest correlation with standardised mortality rate, at −0.662 (p < 0.001). According to the hierarchical multiple regression analysis, the inclusion of average income in the regression model of gross migration rate and standardised mortality rate changed the regression coefficient of gross migration rate from −0.092 (p < 0.001) to −0.016 (p > 0.05), indicating statistical nonsignificance. Conclusion: The inclusion of economic development indicators erased the correlation between population migration and standardised mortality rate, signifying that this correlation could be spurious.