the choice was made to work with day-level data in the analyses. Sum scores were calculated for all variables, measured from noon 12:00 to noon the following afternoon 12:00, so that these daily totals provided a complete picture on the full nightly activity per day.

We analyzed day-level data by calculating sum scores for variables measured from noon to noon the following day. This provided a complete picture of nightly activity per day.

Given that in this study we are primarily interested in nightlife activity, a day is considered from noon 12:00 to noon 12:00 in order to get a picture that captures the whole night. Whenever "Day" is mentioned further in this study, it always refers to the above interpretation. For example, when speaking of September 1, 2019, we are unambiguously referring to the time interval 01/01/2019 12:00:00 to 02/01/2019 12:00:00.

This study focuses on nightlife activity and considers a day as the time from noon to noon, capturing the entirety of the night. Whenever "Day" is mentioned, it refers to this definition. For instance, September 1, 2019, refers to the time interval from 01/01/2019 12:00:00 to 02/01/2019 12:00:00.

When additional sanitary facilities are added in the environment, we see no significant effect in contrast to the business-as-usual condition. Thus, it can be said that adding the urination column did not statistically lead to a reduction in the number of PU’s compared to a business-as-usual situation. Compared to the baseline measurement there is a non-significant increase in the estimated average number of 4.03% in PU’s. Since the added toilet facilities were present in all subsequent interventions as a constant, this condition will be referred to as 'baseline measurement' in what follows.

Adding additional sanitary facilities did not lead to a statistically significant reduction in the number of PU's compared to the business-as-usual condition. The estimated average number increased by 4.03% compared to the baseline measurement. This condition will be referred to as 'baseline measurement' in the following interventions.

However, such strong correlations do not come as a surprise. When there are more visitors, it can be expected that there will be a higher café turnover and more nuisance. Despite the strong association between different covariates, no problematic VIF scores were found. We have therefore decided to withhold all six predictors for further analyses. There are sufficient reasons to believe that these covariates each have a unique added value to the study. Consequently, it should be considered that due to the strong interrelationship between different covariates, some effects may be suppressed. When we look at the relationships between the covariates and the central dependent variables, we do see a relatively low correlation with the two covariates concerning temperature . However, for safety, we will also withhold both covariates in further analyses.

Strong correlations are not surprising. More visitors result in higher café turnover and more nuisance. No problematic VIF scores were found despite the strong association between different covariates. We will withhold all six predictors for further analyses. Each covariate has a unique added value to the study. Due to the strong interrelationship between different covariates, some effects may be suppressed. The relationships between the covariates and the central dependent variables have a relatively low correlation with the two covariates concerning temperature. For safety, both covariates will also be withheld in further analyses.

The statistically significant contrast between the 'Feet' condition and the baseline measurement (t (78) = 3.21, padj = .041, Hedges g = .769, 95%CI = [ .276,  1.26]) should be interpreted with caution. This is because the lower bound of the confidence interval is relatively low (. 276). This raises the possibility of a false positive effect. Consequently, we conclude with caution that there were more PU’s during the Zero Measurement (WC) than during the 'Feet' condition. More specifically, .769 standard deviations more PU’s were observed during the Zero Measurement than during the Feet condition.  The estimated average number of PU’s within this condition is (M =15.2, SD =4.67). Thus, it can be concluded that the 'Feet' condition is not statistically different from the business-as-usual measurement with respect to the number of PU’s, but it is significantly different in comparison to the baseline measurement, with a 51.9% decrease in the estimated average number PU’s.

The difference between the "Feet" and baseline measurements was statistically significant (t(78) = 3.21, padj = .041, Hedges g = .769, 95%CI = [.276, 1.26]). However, caution is needed because the lower bound of the confidence interval is relatively low, which increases the possibility of a false positive effect. Therefore, we cautiously conclude that there were more pressure ulcers (PUs) during the Zero Measurement (WC) than during the "Feet" condition. Specifically, 0.769 standard deviations more PUs were observed during the Zero Measurement than during the "Feet" condition. The average number of PUs within this condition was estimated to be 15.2 with a standard deviation of 4.67. Thus, we conclude that the "Feet" condition is not statistically different from the business-as-usual measurement regarding the number of PUs but is significantly different compared to the baseline measurement, with an estimated average number of PUs decreasing by 51.9%.

Given the contrast with the baseline and the 'Arrows' condition indicates a large effect (Hedges' g = 1.03), and the lower bound of the confidence interval is >.5, there is sufficient evidence to conclude that the effect really exists and it is not a false positive result. The estimated mean number of PU’s within this condition is (M =8.75, SE =4.89). This represents a 72.31% decrease in the estimated mean number of PU’s compared to the baseline condition.

The 'Signs' condition shows no significant difference from the normal measurement in terms of PU's (t (78) = .616, padj = 1.000, Hedges’ g = .179, 95%CI = [-.401, .758]). Additionally, there is no statistical effect of the 'Signs' condition when compared to the baseline measurement (t (78) = 1.154, padj = 1.000, Hedges' g = .277, 95%CI = [-.201, .754]). The average number of PU's in this condition is estimated to be (M = 25.6, SD = 4.80). Although not statistically significant, placing the signs results in a 19% decrease in the estimated average number of PU's compared to the baseline measurement.