4.2.1 Descriptive statistics for categorical variables
For example, let us consider the variable ‘resistance to levofloxacin’. This variable was measured for 120 methicillin-resistant Staphylococcus aureus (MRSA) isolates. This variable has two levels: ‘resistant’ for 38 of the isolates and ‘susceptible’ for the others. Descriptive statistics for this variable can be reported in a single sentence of text:
- As a count: AST results showed that 38 of the 120 MRSA isolates were resistant to levofloxacin.
- As a proportion: the AST results showed that 31.7% of the 120 MRSA isolates were resistant to levofloxacin (i.e. ((38/120)*100)).
Below is an example of reporting descriptive statistics in sentence form from the literature. This sentence reports relative frequencies (as percentages) for four variables containing the resistance status (resistant/susceptible) for four antimicrobials.
“A large number of MRSA isolates showed resistance to levofloxacin (83.9%), ciprofloxacin (83%), erythromycin (77.7%) and clindamycin (72.3%).” (Kot et al., 2020)
In the examples above, the outcome variable had only two levels (resistant/susceptible), and tabular or graphical representation would not provide much more additional information. When categorical variables have more than two levels, or when multiple variables with two levels are presented in parallel, they may also be displayed in a frequency table (Table 3) or in a graphical format using a bar chart (see module Summarising and presenting AMR data). When presenting proportions, it is good practice to also provide the corresponding counts, or frequencies, especially in cases where small sample sizes are used. The example in Table 3 shows the breed of cattle from which a target organism was recovered. The categorical variable ‘breed’ has four levels: Jersey, Guernsey, Holstein Friesian and Unknown.
Breed | Frequency | Relative frequency (%) |
---|---|---|
Jersey | 234 | 20.7 |
Guernsey | 63 | 5.6 |
Holstein Friesian | 800 | 70.7 |
Unknown | 34 | 3.0 |
Total | 1131 | 100 |
4.2 Descriptive analysis