This study examines the determinants of a household’s threshold price for a restaurant meal cost increase; the level of cost increase that would cause households to either eat in restaurants less frequently or change what they would typically purchase. The design of the study is formulated using a Tobit model to examine the threshold price by differing social, economic and demographic characteristics of households in Kentucky as well as their preferences for restaurant-specific characteristics. The empirical estimates suggest that households that frequently have dinner at restaurants, households with higher incomes and households that strongly prefer full-service restaurants have a positive threshold price-range; which suggests that such households are more willing to pay an additional cost increase in restaurant meals. Conversely, households that always notice taxes before paying their checks, households close to retirement-age, and households that do not strongly prefer local-food restaurants have negative threshold price-range and are consequently less willing to pay an additional cost increase in restaurant meals.

在台灣，未來很有可能帶領台灣高科技產業向上突破的新英雄---薄晶電晶體液晶顯示器（TFT-LCD），已逐漸為人所重視，近來，TFT-LCD產業的更是利多頻傳，不但在股票成交量創下新高，股價更是一再突破新高。本文則是希望以資料包絡分析法(DEA)來比較分析台灣各家廠商（友達、華映、奇美電、廣輝、彩晶）的相對經營績效，研究範圍從2001年第四季一直到2004年的第一季，所採用的投入變數即為資產總額、營業成本、營業費用和員工人數，產出變數我們則採用營業收入和營業淨利。 總結來看，廠商在CCR和BCC模式下，雖然在平均排名上，兩個模式並不盡相同，但由時間趨勢來看，五家廠商在趨勢變動上，大致是一樣的；效率值表現較好的時期為2002Q1、和2002Q2和2004Q1，較差的則為2002Q4和2003Q1，推測其原因，一部分當然是和景氣的波動有關，但也和面板的價格波動有一定程度的相關。 藉由Tobit 迴歸分析，意味著廠商可藉由提高每人配備率、總資產週轉率、流動比率、研發費用率及經營年限等變數，可提昇廠商的整體技術效率，提供了一些政策方向供廠商參考。 以Malmquist生產力指數來看，總要素生產力的變動（Tfpch）大部分的因素，是來自於生產技術的變動(Tech）；彩晶在總要素生產力的變動上，平均有16.8%的成長，是第1名，其他的廠商則呈現不大的差別；但若以時間趨勢來看，2001Q4到2002Q1和2003Q1到2003Q2兩各階段都有滿大的成長，但在2002Q2到2002Q3和2002Q3到2002Q4兩個階段卻呈現衰退的表現。 / In Taiwan, TFT-LCD, which is likely to lead high-tech to strive upward in the future, has been receiving more and more attention. Recently, TFT-LCD industry has come into great blossom. Not only does the stock transaction amount achieve a record-breaking peak, but the stock price also makes breakthroughs at all times. This research is to use Data Envelopment Analysis(DEA) to compare and analyze the relative operating efficiency of chief leaders in the high-tech industry, including AUO, CPT, CMO, QDI to HannStar. The research ranges from the 4th quarter in 2001 to the 1st quarter in 2004. The researcher adopts input variables, which include total assets, business costs, business expenses and employees, and output variables, business revenues and business net profits, as well. On the whole, in spite of CCR and BCC model, the average order is different, but approximately similar. The quarters, which perform well, are the 1st quarter in 2002 and the 1st quarter in 2004; the bad quarters are the 4th quarter in 2002 and the 1st quarter in 2003. To speculate reasons, it has to do with the business cycle and panel board price fluctuation. Under the frame, by use of Tobit analysis, factories can strengthen total technology efficiency by raising equipment per employee, total asserts turnover, current ratio, R & D ratio, and the period of operating. It provides these factories a referential direction. In conclusion, according to Malmquist index analysis, the Tfpch is the better part from the tech. HannStar, with 16.8% average growth in Tfpch, is the top one. The other factories are nearly close. In the time period, from the 4th quarter in 2001 to the 1st quarter in 2002, and from the 1st quarter in 2003 to the 2nd quarter in 2003, the other factories has a high growth, but from the 2nd quarter in 2002 to the 3rd quarter in 2002 and from the 3rd quarter in 2002 to the 4th quarter in 2002, there is a decline in growth.