Acta Geographica Sinica ›› 2018, Vol. 73 ›› Issue (2): 295-317.doi: 10.11821/dlxb201802007

• Transportation and Tourism Geography • Previous Articles     Next Articles

Sensitivity analysis of the measures of tourism seasonality

LIU Zehua1(),ZHANG Jinhe2(),PENG Hongsong2,ZHANG Yu2,TANG Guorong2   

  1. 1. School of Business Administration, Nanjing University of Finance and Economics, Nanjing 210023, China
    2. School of Geographic and Oceanographic Sciences, Nanjing University, Nanjing 210023, China
  • Received:2017-01-25 Online:2018-02-11 Published:2018-02-11
  • Supported by:
    National Natural Science Foundation of China, No.41671145;Planning Project of National Tourism Administration, No.15TAAG011;Tourism Young Expert Training Program, No.TYETP201312


Standard Deviation (Rsd), Gini coefficient (G), Theil index (T), seasonality ratio (Rsr), and ratio of highest to lowest-season demand (Rhl) are popular measures of tourism seasonality; however the relationship between these measures and their influence on changes in tourism flow varies in literatures. Two sensitivity indices were developed to reflect the effect of the changes of tourism flow on the tourism seasonality measures (si) and the relative effect of the changes of tourism flow on the tourism seasonality measures (pi). The correlation coefficient of measures of tourism seasonality in various scales (Int), the μsi, μsi*, and μpi* and changes of tourism flow (Δ) are calculated by the Monte Carlo method. The results show that: (1) All the correlation coefficients between Rsd, G and T are over 0.970. Therefore, the tourism seasonality of some destinations with the three measures is significantly consistent. The correlation coefficients of Rsr and Rsd (or G, T) are between 0.446 to 0.845, and increase with increasing Int. The differences of using Rsr and Rsd (or G, T) are increased with Int. The correlation coefficients of Rhl and Rsd (or G, T, Rsr) are in the range of 0.003-0.772, and smaller with increasing Int. The differences of using Rhl and Rsd (or G, T, Rsr) are decreased with increasing Int. (2) Besides s12 of Rhl, the only coefficient affected by m1, si of Rsr and Rhl to the other months are influenced by Rsr and Rhl, Δ, m? and mi. In addition to s12 >0, the value, even the sign, of si has no correlation to Rsr and Rhl, and cannot consider to be more sensitive to a specific month. (3) The μsi, μsi*, and μpi* of Rsd, G, T and Rsr for each month are varied by Int and Δ. Values and rank order change respectively to Int and Δ. Among them, s12 is always positive, increasing tourism flow increases the value of the measures of tourism seasonality; and to the other months, changing tourism flow may increase, decrease, or no change measures of tourism seasonality. Range of seasonal variation and changes of tourism flow could affect the sensitivity of different measures of tourism seasonality and the rank order in different months. Therefore, no such particular measure is discovered in this research that it is more sensitive than other measures related to the tourism flows, and neither are any measures more sensitive to the changes in the peak season, off-season, or shoulder season.

Key words: tourism seasonality, measure, sensitivity analysis, Monte Carlo