*2.6. Statistical Analyses*

The data of resistance to antibacterial agents were performed using the SAS statistical software package version 7.1 (SAS Inst., Cary, NC, USA). MICs were summarized and reported as susceptible (S), intermediate (I), and resistant (R; the results were classified as 1 = S, 2 = I, or 3 = R), where CLSI veterinary breakpoints were available [28]. The analyses were made with these values for the categories. There are no intermediate values between classes one, two, and three. Therefore, a generally high standard deviation has to be tolerated. In the case of completely sensitive isolates at the beginning of the tests, the values are constant at one, i.e. the standard deviation is zero and can therefore not be seen graphically. Significant differences in the means of the resistance results between the four groups of flooring designs were tested using the repeated measures ANOVA (Fisher's Least Significant Difference (LSD)). This test was also used to determine the differences between the sampling stages and the frequency of resistance between the three trials.

#### **3. Results**

In total, 864 *E. coli* were isolated and analyzed. These isolates were obtained from 648 cloacal swabs and 216 manure samples at the BT, AT, and ET stages. In the water collected at days ten to 14 and days 26 to 30 in trial 2, the enrofloxacin concentration were 50.17 and 50.62 μg/mL, respectively; in trial 3, water contained 49.87 and 50.42 μg enrofloxacin/mL, respectively.
