目標：本研究旨在建構食品素養自陳量表，並進一步探討大學生食品素養與飲食行為之 關係。方法：採取分層取樣的方式，資料蒐集時間自2017年5月至6月，正式樣本共394位大學 生。結果：經預試（n=202）分析，食品素養自陳量表可區分為「互動性」、「功能性」及 「批判性」三構面，各構面Cronbach’s α介於.76至.87，全量表信度為.89，可解釋總變異量為 66.08%。再以正式樣本（n=394）進行驗證性因素分析，量表個別項目的信度介於.40至.66；因 素負荷量介於.63至.81（p<.001）；潛在變項的組合信度為.80至.86之間；平均變異抽取量為.50 至.61之間。模式的各項指標顯示本研究的適配度佳。結論：本研究所建構之食品素養自陳量表 具有不錯的信效度。大學生自認為具有中等程度以上的食品素養能力，且有較好的消費者健康 和飲食控制行為，但均衡飲食的習慣較差，相對的，偏差飲食行為則較高。此外，大學生的批 判性食品素養比功能性和互動性食品素養更能正向預測其飲食行為。

Objectives: The purpose of this study was to construct a food literacy self-report inventory and to investigate the relationships between food literacy and dietary behaviors among college students in Taiwan. Methods: A nationally representative sample of 394 college students was surveyed using stratified sampling from May to June 2017. Results: In a pilot study (n = 202), items were loaded on the following 3 dimensions in exploratory factor analysis: functional, interactive, and critical food literacy. The Cronbach’s α was between .76 and .87, and that of the overall scale was .89 with a total variance of 66.08%. In the formal study (n = 394), we used structural equation modeling to examine the measurement model. The reliability of the individual food literacy scale items ranged from .40 to .66. Moreover, standardized factor loading ranged from .63 to .81 (p < .001), composite reliability from .80 to .86, and the average variance extracted for each dimension from .50 to .61. The indicators revealed that the measurement model had a good fit. Conclusions: The food literacy self-report inventory developed in this study is a reliable and valid measure. The status of food literacy and dietary behavior showed that the college students had at least a medium level of food literacy, and although they had good consumer health and diet control behavior, they had a poor balanced diet and deviant behavior. Additionally, compared with functional and interactive food literacy, critical food literacy positively predicted dietary behavior.