Read Aloud the Text Content

This audio was created by Woord's Text to Speech service by content creators from all around the world.

Text Content or SSML code:

the results from the exploratory factor analysis (EFA) indicate good sampling adequacy (KMO = 0.761). the result of Bartlett´s test of sphericity allows us to reject the null hypothesis that the correlation matrix is an identity matrix (p = 0.000), and the determinant of the correlation matrix equal to 0.001 indicates the absence of multicollinearity (Prato, Bekhor, & Pronello, 2005). the extracting method used was principal axis factoring based on Eigenvalues greater than 1 with varimax rotation (Kaiser normaliza-tion). the results indicated the existence of 5 factors (see table 2), which was supported by the scree plot that levels off after that, and by the theory, since the grouped items have a theoretical meaning (Hair, Black, Babin, & Anderson, 2014). the resulting five factors represent latent variables and composite measures: active modes preference, residential location satisfaction, public transport frequency of use, travel satisfaction and transport supply, and land-use mix at the residential location. the dominant items (marked in bold) were defined considering a cut-off value of 0.4, which is adequate to the sample size (Hair et al., 2014). the Cronbach’s alpha values of each factor are presented inside parentheses and indicate moderate to excellent internal consistency (Hinton, McMurray, & Brownlow, 2014).the confirmatory factor analysis validated the factor model construction (see table 2). As we imple-mented the BSEM approach described in Muthén and Asparouhov (2012), the estimates for the mea-surement equations include small-cross loadings between the indicators of the other factors, which are constrained to be close to zero, resulting in a structure similar to the EFA output. It is worth mentioning that the results of the confirmatory factor analysis do not include any significant cross-loading between factors but rather small non-significant cross-loadings that allow for an analysis that better reflects the substantive theories. Note that, despite the similarity, the model is regulated, and the cross-loadings are restricted to small values through the specification of priors (Muthén & Asparouhov, 2012). We have highlighted in bold the factor loadings for which the 95% credibility interval does not contain zero, making it easy to observe the structure defined for the factors in the confirmatory analysis. the Posterior Predictive P-Value (PPP) is 0.289, indicating a good fit of the results of the confirmatory factor analysis.  this study highlights the occurrence of residential self-selection and its impacts on residential and travel choices and on the derived level of satisfaction, focusing on factors influencing residential and travel satisfaction of transnational short-term residents.We found that individuals presenting lower levels of residential satisfaction (dissonant residents) are those who pay lower rents. this is associated with living away from the university/workplace, with higher transport expenditures. thus, the dissonant residential choice seems to be based on tradeoffs in-volving commuting distance, monthly rent, and transport expenses. the locational dissonance of those living further away than desired seems to occur due to budget constraints. However, as transport supply and land-use balance were found to directly influence residential satisfaction, it is unlikely that it is the distance itself that makes people less satisfied with their residential location. Rather the impact of budget constraints in the lower levels of accessibility to transport and different typologies of places seem to cause the lower residential satisfaction levels, as people could not properly self-select. Another possible cause for the residential location dissonance is that people do not have enough information on the options on transport supply and the land-use balance of the residential locations further from the university.In contrast, individuals presenting higher residential satisfaction (consonant residents) have stron-ger preferences for active modes, lower levels of public transport use, and spend less than 10 euros on transport a month. these are indicatives of a residential location close to the university/workplace, favor-ing the use of active modes for commuting. As for their residential choice, it seems that the stronger the preference for active travel one holds, the lower her/his sensitivity is to other features/tradeoffs that other residential location options offer.For both cases, the model reveals that better transport supply and land-use balance at the home location can improve both residential and travel satisfaction. this may be possible through informing better short-term residents on these attributes for different neighborhoods so that they can avoid mis-match while choosing their residential place.For city planners, it is important to identify workplaces that attract short-term residents so that people who self-select to live close to them can find a good infrastructure for walking or cycling, and those who do not manage to live close still can count on having a good public transport connection.the model also evidences that the travel behavior of an individual reflects her/his travel preferences, spatial constraints (distances), and opportunities provided by her/his residential location as the starting point for travels. these results are in line with Næss (2005), that claims that incentives and deterrents for the pursuit of specific travel behaviors are implicit in the urban structure through the creation of proximities and distances between activities.It would also be interesting to investigate the reasons why those who were living in Porto for up to 3 months, by the time of the survey, present lower levels of travel satisfaction. this may be the result of initial dificulties in traveling in an unfamiliar environment and system, which highlights the impor-tance of making it easier for newcomers to navigate the city and to use public transport.Finally, this study has some limitations due to the small sample size. the Bayesian Estimation for the Structural Equation Model assures the robustness of results, but a larger sample, including short-term resident groups other than students/researchers, will hopefully demonstrate the transferability of these results.