I need help understanding which statistical test to use?

NP's conducted a pilot study to examine the efficacy of smoking cessation programs. A convenience sample of of 9 heavy smokers were assigned to 2 different programs for 2 months, 5 for weekly individual counseling and 4 for weekly group counseling. If the NP wants to compare the difference in the number of smoking cigarettes/day, which statistical method should be used?

Kruskal-Wallis, or Wilcoxon signed-rank, or Mann-Whitney, or Friedmans test?