. 2009 Dec 23;48(3):952–956. doi: 10.1128/JCM.01860-09

TABLE 2.

Nonparametric comparison of MICs for six antimicrobial agents tested against 39 Brucella isolates at three test sites for each incubation atmosphere

Antimicrobial agent	Mean rank^a		P value^b
Antimicrobial agent	Ambient air (n = 101)	CO₂ (n = 117)	P value^b
Doxycycline	121.0	98.7	0.004
Gentamicin	85.6	130.2	<0.001
Rifampin	108.4	109.6	0.876
Streptomycin	81.6	133.6	<0.001
Tetracycline	119.4	101.0	0.016
Trimethoprim-sulfamethoxazole	98.3	119.2	0.008^c

n = the number of MIC results for each antimicrobial agent. There were 35 ambient air (4/39 isolates required CO₂) MICs for 2 sites and only 31 at one site, for a total of 101 results. n = 100 available ambient air MICs for doxycycline and rifampin. The mean rank values were calculated by first converting the MICs to whole numbers in a linear fashion (e.g., MICs of 0.25, 0.5, 1, and 2 were converted to 1, 2, 3, and 4, respectively). These relative values were then used for statistical analysis.

Computed using the Wilcoxon rank sum test. A difference is significant if P is <0.05. If the mean ranks are not significantly different, the implication is that no significant shift in MIC has been observed based on these sample data. These results are supported by the Kuiper test for all agents except trimethoprim-sulfamethoxazole.

P = 0.110 by Kuiper test.