**3. Material and Methods**

The research carried out was twofold. Initially, the desk research method was applied, the analysis of secondary and primary data (subject literature, strategies and plans for tourism development in Wrocław [48,49], reports from the tourist traffic survey, Statistics Poland data, website data). The authors reviewed scientific publications on conditions of tourism development in urban entities and agglomerations, processes of sustainable development of tourism economy in large cities, and the effects of tourism concentration (including overtourism) in urban space.

Direct research was of key methodological importance. It was conducted in Wrocław in November 2019 as a questionnaire survey and aimed at a preliminary recognition of the investigated overtourism phenomenon. The research was performed in three stages. The first stage was a pilot study (questionnaire test) involving 10 key experts: representatives of Wrocław city authorities and local government (directors of tourism development offices), business communities (including representatives of tourism industry associations), academic communities (of recognized position in the research of tourist phenomena), and planning institutions (actively occupied with creating the sphere of tourism in Wrocław). The second stage was to collect opinions (on the basis of a survey questionnaire) of people directly involved in tourist services (90 persons: tour guides and tour leaders, also affiliated with the industry associations, as well as organizers of the incoming tourism to Wrocław), whom we treat as experienced observers and tourist experts with established views on tourist phenomena taking place in urban space. The third stage of the survey (survey questionnaire) involved students of tourism (60 persons: Wrocław residents), as a group of additional experts, having in-depth knowledge of tourist traffic, as well as a more critical view of the current phenomena in tourism. It is important that the respondents (experts) were residents of Wrocław.

Purposive sampling was applied with the consideration of the respondents' suitability for assessing the potential overtourism. It should be emphasized that the essence and purpose of the research was not to measure the entire population, as it was a qualitative study aimed at identifying and describing the phenomenon of overtourism on the basis of the opinions of people actively involved in tourism services or constantly observing the market for these services. The sample size in the survey was not substantial since the key variables in the selection of respondents were their competences and experience which would guarantee deepened opinions on the manifestations of overtourism.

Direct research with key experts was carried out by using the computer-assisted web interview method; in surveying people directly involved in providing services to tourists (tourist experts), we employed official meetings organized by the Wrocław authorities to present directions of tourism development. The additional experts, a group of students, were selected from among the best students of tourism (master's degree).

The research was based on a questionnaire with 13 groups of variables. The variables included:


The purpose of the empirical research conducted in this manner was to identify the negative and positive effects of tourist traffic in Wrocław, particularly in terms of sustainable development of tourism as well as the potential threats to this process resulting from overtourism.

The statistical analysis of the survey results was conducted by using the R 3.5.3 software [50]. The response rate diagrams were obtained in the likert [51] package in R. In order to present a comprehensive picture of the results, the following calculations were performed: the mean rate of positive and negative indications for the main research issues for the groups of tourist experts and additional experts (hereinafter referred to as other experts) (Mro—mean rate of other experts' responses) and key experts (Mre—mean rate of key experts' responses), the coefficient of variation for other experts (Vo) and key experts (Ve), as well as the mean point values (Mpv1–5 or Mpv1–7) of opinions for particular detailed questions, expressed in the ranges of 1–5 and 1–7.

The analysis of the significance of differences in the rates of particular responses was also carried out and verified with a Fisher's exact test, with the statistical significance level of α = 0.05. In turn, the significance of differences in the mean point values ascribed to particular responses was verified with the Wilcoxon test, with the statistical significance level of α = 0.05. An analysis of the correlation of total scores was performed for 5 content-related questions from the survey questionnaire. The total scores were determined by summing up all responses of a respondent separately for each of the indicated questions and then scaling them to the [0;100] range by using the following formula:

> *sum o f points* − *minimum sum o f points maximum sum o f points* – *minimum sum o f points* <sup>×</sup> <sup>100</sup>

In drawing conclusions, the methods of deduction and comparative analysis with the technique of describing diversity and similarities were applied. In order to simplify the problem presentation, the obtained data were aggregated, limiting the number of presented observations to key results only. The authors also relied on their own observations and experiences as professional and active participants in the system of planning, managing, organizing, and servicing the tourist traffic in Wrocław (observation participant method).
