在旅游社区,旅游从业者是一个特殊的群体,研究他们对旅游影响的认知水平与其背景特征之间的内在关系,有助于旅游业的可持续发展。本文利用除趋势典范对应分析(Detrended Canonical Correspondence Analysis,DCCA)对芦芽山旅游从业者对旅游影响程度的认知水平进行了分析。结果表明:DCCA第一排序轴与性别、旅游收入和受教育程度有较高的相关性,而第二排序轴与违法情况、性别的关系较为密切。这些变化趋势与DCA分析结果基本一致。DCCA作为一种限定排序,在研究样本类型与其背景特征的关系上优于DCA,但是在描述样本类型之间的关系上,则DCA优于DCCA。

The residents engaging in tourism industry are a kind of special crowd in the tourism community. It is of help to tourism industry sustainable development that the inherent relationship between their expression and background toward its impact is analyzed. The relationship between the residents engaging in tourism industry and their expression toward its impact was analyzed by Detrended Canonical Correspondence Analysis(DCCA)in Luya Mountains Nature Reserve in the paper.The result indicates that there is obvious correlation between the first axis of DCCA and sex, tourism income and the education level, and there is evident correlation between the second axis and the case of violations and sex. From the left lower part to the right upper part in DCCA,four groups transit from the mature active group,the non-mature active group and the receded group to the bland group gradually. The length of arrowhead denotes the relationship between sample groups and different factors,and the angle between the arrowhead line and the ordination axis shows the magnitude of their correlation. Therefore,the DCCA figure can show the relationship between the groups and their characteristics visually. These trends are consistent with the Detrended Correspondence Analysis(DCA)results. Although DCA can also study the relationship between samples types and their background, we need to judge it based on actual condition after completing the ordination. The calculation of constrained ordinations combines the data of sample groups with that of their background characteristic,so their results are superior to the result of DCA on studying their relationship. Moreover,DCCA is the best in all the constrained ordinations because of removing"arch effect".However, DCCA can make the distribution of samples denser, and make the borderline among different types much blurrier. Therefore, if DCA is combined with some classfication ways on studying the relationship among the sample groups, its results are much better than that of DCCA.