使用者的聲音你聽見了嗎？以敘述性選擇實驗探討都市設計元素與行人路線選擇偏好關係

Abstract

所謂「新都市主義」、「理智型成長」等以改變都市形式、引導民眾轉變運輸行為之政策之有效與否在學界仍未尋得一致之共識，然而此類創造適宜使用非機動運具與大眾運輸工具之環境以鼓勵民眾使用的政策卻正以「人本交通」之名於台灣逐漸風行。有鑑於此，本研究透過分析行人路線選擇之行為，試圖增進吾人對個體與步行環境間互動關係之理解，並為此類政策之有效性提供運輸行為上之證據。本研究首先探討個別都市設計元素（與其背後所代表的都市形式改善政策）影響步行路線選擇的相對重要性。其次透過實驗設計手法避免因子間多元共線性的發生，進一步探討都市設計元素間之交互作用對於行人路線選擇偏好之影響，以釐清是否有特定政策間之組合能夠對行人之路線選擇偏好發揮更大的影響力。 由於實際環境中都市設計元素種類繁多且以混雜的方式呈現，測量與界定上均有其困難，變數間之多元共線性更可能影響模式校估之結果，本研究以蒐集敘述性偏好資料校估多項羅吉特模式，觀察受訪者面對簡化假設情境之路線選擇反應。所建構之敘述性選擇實驗之假設情境屬於「未標記」實驗，由兩條路線組成，包含路線、路段與路口等三類之九個變數。由於假設情境之構成以名目變數為主，在水準值個數之選取受限下本研究改以電腦輔助之D-optimal實驗設計安排實驗計畫。 敘述性偏好資料於九十七年五月間以街訪方式於台北國道客運總站蒐集，共完成538份，經剔除無效問卷後共1,508個敘述性選擇實驗回應（每份問卷最多提供三個）。多項羅吉特模式校估之結果顯示都市設計元素之改善確實提供了受訪者改變路線選擇的誘因，其中提供寬敞的人行道最為重要，其次則是路口行人輔助設施的提供與人行道-車道間分隔方式的改善，步行路線沿途穿越路口數的多寡同樣也會影響受訪者選擇步行路線。相形之下，步行時間並不是受訪者在選擇步行路線時最為重要的因素。本研究並沒有發現不同的都市設計元素間對行人路線選擇偏好有明顯的交互作用影響，但由於探討的交互作用項目有限，此部分有待後續研究進行確認。本研究共蒐集了性別、年齡等七個社會經濟變數與步行行為變數用以探討步行路線選擇偏好的異質性。我們發現不同變數影響步行路線選擇偏好的異質性之程度並不相同，其中「年齡」與「個人月收入」是最能解釋偏好差異的兩個變數。

Policies creating a friendly transportation for transit or non-motorized modes through the control on urban design or land use pattern have earned their popularity with the name of “humanity-oriented transportation” in Taiwan. However, there is still a fierce debate within the academic circle on the effeteness of these measures. Respecting of this discrepancy, the preference of pedestrian route choice has been analyzed to provide some behavioral evidence about these strategies, including the relative importance of essential urban design elements and their interaction effects. As the facts that there are various urban design elements and they often exist as a mixture in the physical world, both of which cause difficulties on the measurement and classification toward these variables, the preference data was gathered with a stated choice experiment providing simplified walking environment composed of homogenous urban design elements along each alternative route to be chosen. The stated choice experiment established was an ‘unlabelled’ one, which includes two alternatives profiled by nine urban design variables. Another nature of urban design elements inspires the modification of research design as well: computer-aided D-optimal experimental design was adopted to overcome the constraint that many of the variables can only be classified nominally or ordinally. The stated choice experiment was held in May of 2008 with subjects recruited from the intercity bus terminal in Taipei. The result of multinomial legalistic regression confirms that the improvement of urban design elements provides incentives for pedestrians to choose their preferred route. The provision of sidewalk with enough width is the most essential element among all the urban design elements, where the installation of crosswalks and pedestrian signals in crosssections takes the next place. There is no statistically significant interaction effect between different urban design elements; however this conclusion should be examined further owing to the number interactiuon factors that had been examined is constrainted by the original experiment design. In addition to the different degree of influence of every urban design elements on preference of pedestrian route choice behavior, the heterogeneity of preference across individuals with different demography and socio-economic variables is also conspicuous. The age and personal monthly income are the most two variables that can describe this heterogeneity.