利用清代各省按月奏報的糧價資料來作科學化研究分析的學者，自從王業鍵院士在一九七○年代後期開始後，陸續有數位美國的學者（Li, 1992; 2000; Wong and Perdue, 1992; etc.）收集了直隸等省資料著文討論，大陸方面有陳春聲（1992）對於十八世紀廣東糧價寫成專著等，這些算是清代物價史的一個開始。我們則從大量糧價資料的科學化分析著手，試著將統計科學和方法應用上來，可以從中獲取多方面的訊息，並且提供一個客觀而具體的印證指標。在這期間，從陳仁義、王業鍵和胡翠華（1999）的研究過程中，我們已著手整理了時間數列的統計分解模型，來模式化清代糧價資料序列值，其中有些月份原件散失，因此資料的不完整和可靠性是難以避免的問題，陳仁義、王業鍵、溫麗平和歐昌豪（2003）已初步發展完成遺漏值填補法和資料可靠性檢測法。 在統計模式化過程中，我們將糧價的變動視為由趨勢性、季節性、週期性和不規則性等因子形成和相互影響，從蘇州府米價的實際資料分析結果看來，季節性的變動非常顯著，米價指標值以八月份為最高，隨後就往下降，到十一月時最低，隨後又開始往上升，持續到八月份，大致上與稻米生產和收成的背景相互穩合；從麥價的季節指標之相對變動中，可以看出在三、四月時指標值最大，四、五月開始下降，正逢小麥收成季節，可作為米的替代食物以緩和米價持續上升，因此對於米價有穩定的作用，此為科學化印證之一。 除了季節性循環之外，中米價格呈現緩和上升的走勢，其中主要因素是美洲白銀大量流入，市場上貨幣流通量膨脹，貨幣供給和流通速度的增加稍許超過了糧食供應的增加所致（Wang, 1990），其次是人口的成長大致上可以和糧食的供給增加量並駕齊驅（Wang, 1990; 1992; 全漢昇，1972），其中高價部份上升的幅度較低價大些，此為科學化印證之二。 我們進而引用了週期頻譜法（periodogram）來找週期性的存在和變化，運用資料轉換為頻率軸的型態之後，頻譜值檢定法可以同時且自動化地檢測出多個週期來，其中發現了顯著的二十六年左右之長週期存在，很值得做進一步的研究。當分離初系統性因子之後，隨機性的不規則因子部份，其統計分析結果亦呈現出一種穩定的隨機化序列，這些結果正可以印證當時是中國歷史上難得的物價穩定時期（王業鍵和黃國樞；Wang, 1990; 1993），此為科學化印證之三。值得一提的是，Wang（1993）所作的中國六大區域糧價變動的初步觀察，透過多組資料的並列時間序列圖中，就已指出各區域糧價同樣呈現大至二十年到三十年的週期性變動。此外，引用的頻譜法可以同時且自動化地檢測出多個週期來，也檢定初4.3與3.5年的週期顯著性，正好和Wang（1990）的直觀法所得到之可能週期為2、3、4、5年當中，出現4年的次數為最高是相近的。因此，我們引用的統計檢定法和科學化的印證是值得肯定的。 陳仁義、王業鍵和周昭宏（2002）以相關性的統計分析，來探討清代糧價市場整合性，透過科學化方法所得到的客觀性數據，來和歷史性的研究考察結論相互印證，從東南沿海四省的米價資料庫中，我們選取了較為可靠的十八世紀時期作整合性的統計分析，以相關係數矩陣和轉化的指標函數，用來模式化整體性的相關程度和區域內地方間的互通性。初步的研究分析結果，相當程度地印證了歷史性的考察結論，也進一步的支持王業鍵和黃瑩玨（1999）的考察研究結論：“十八世紀的中國二個樞紐地區 --- 長江三角洲和珠江三角洲 --- 經濟上關聯尚弱。但是，糧食不足地區與有餘地區之間的地域分工與經濟交流至為明顯，各個經濟大區市場關聯性高低各異，但並非孤立。就我們所考察的地區來看，大致可以說，以長江三角洲為中心，地理上和交通運輸上愈接近的地區，市場整合程度愈高。”此為科學化印證之四。基於此，我們認為運用了糧價紀錄的相關係數值和係數矩陣指標函數值，來印證地區間的糧價互動程度、糧食的供需情況和糧價波動的同步狀態等，進而來探討糧價市場的整合程度，是個具體可行的科學化研究法。

Wang (2004) took a lot of time to collect the invaluable real data bank of Qing's grain prices. In an argrarian society like Qing China, grains are the most important commodities in domestic trade in which food consumption makes up more than half of the average household budget. Grain prices are therefore the leading indicator in the market for which a clear knowledge of their trends and patterns will provide a key to understanding the state of economy and society (Li, 1992; 2000; Wang, 1990; 1992; 1993; Wong and Perdue, 1992). For grain price data collected, these statistical data should rank among the best produced by imperial China, but are far from perfect and should be treated with caution. To check the reliability of grain price data, Chen, Wang, Wen and Ou (2003) developed an automatic method for graphical preview or statistical test. Based on these, further data analysis can be coducted using suitable statistical modeling. Firstly, we concentrate on the statistical analysis of the prices of rice for Suzhou Prefecture which is the principal grain market in the Yangzi delta, the center of economic gravity in late imperial China. The decomposition model in a time series is chosen for studying the seasonal variations, the trend movements, the cyclical fluctuations and the irregularity in the grain prices. Specifically, the statistical methods, including moving average method, regression analysis, and periodogram are used. The pattern or systematic changes from the original data may reflect conditions of inflation, deflation, or crises of major proportion. Preliminarily, nonparametric regression is used for imputing the missing part in the original data for subsequent analysis. Once the modeling is done, iterative procedure can be applied for smoothing the missing data and the results of consequent analysis may be used for evaluating the model fitness. The conduction results are generally consistent with the conclusions of Wang (1990). The seasonal variations of prices show the effect of the multiple-crop system of Suzhou. Rive prices did not reach their peak until August since the wheat harvest took place around April and May. In addition, the price of slower rising trend confirms the possibility that its markets were relatively well developed. It also indicates that a distinct 26-year cycle shown in this data set is invaluable for further study. Details can be seen in Chen, Wang and Hu (1999). On the other hand, Chen, Wang and Chou (2002) further study the integration of rice markets on the southeastern coast of China, where thirty-four prefectures are chosen and classified into 10 groups on the basis of rice-trading route and geographic location. A statistical analysis of rice prices from 1741 to 1760 were conducted using the decomposition method of time series model. By doing these, the rice price data are averaged by year for removing the seasonal variation. Then year-based data are further de-trended for computing the correlation coefficients of pairwise prefectures in each group and forming the correlation matrix. Two index functions of the correlation matrix are produced for measuring the integration of rice markets. Our statistical results give further scientific proof of the major findings of Wang and Huang (1999), that is “centering on the Yangzi Delta, the shorter is the distance from it, the greater the market integration observable along the coast.” Expanded for this in the nationwide, Wang and Chen (2002) observed the extent of market integration for the metropolitan regions － Suzhou, Wuhan, Zhongqing, Guangzhou, Tianjin, and Xi' an. Our preliminary investigation indicates that grain markets were strongly integrated along the Yangzi River Valley, that less so was the case along the Grand Canal, that the Yangzi Delta and the Canton Delta appeared to be far from integrated, and that the Northwest centering at Xi' an was by and large autarkic. More findings are also discussed.