Given the current crisis regarding painkiller addiction, the administrator is concerned about how patients seeking drugs might respond negatively on overall ratings. What test would be appropriate for looking for an association between whether the patient was dissatisfied with pain management (vs. satisfied/neutral, combined into one other group) and whether the patient would recommend the hospital (yes/no)?

Wilcoxon rank sum test

Kruskal-Wallis test

Spearman correlation

Fisher's exact test