Analysis of Cultural Differences

Step 1: Graphs
Graph > Boxplot > with Groups, Graph variable=MTBW, Categorical Variable=Country

There is a problem with the normal assumption. We can try to fix this with a square root transform:
Calc > Calculator, Store in Sqrt(MTBW), Expression SQRT('MTBW')
Graph > Boxplot > with Groups, Graph variable=Sqrt(MTBW), Categorical variable=Country

Still not good. How about a log transform?
Calc > Calculator, Store in log(MTBW), Expression LOGT('MTBW')
Graph > Boxplot > with Groups, Graph variable=log(MTBW), Categorical variable=Country

but this does not look any better either.

Step 2: Summary Statistics
We could not even find a transformation, so we will use the median and IQR/1.35
Stat > Basic Statistics > Display Descriptive Statistics, Variable=MTBW, By variables=Country, Statistics > check IQR

Country

Groups n Median IQR/1.35

USA 20 134 32.07

Europe 15 140 19.03

Japan 12 166 14.22

Country
Groups	n	Median	IQR/1.35
USA	20	134	32.07
Europe	15	140	19.03
Japan	12	166	14.22

Step 3: Hypothesis Test
Because none of the transformations worked we will use the non-parametric Kruskall-Wallis test:

Stat > Nonparametrics > Kruskall-Wallis, Response=MTBW, Factor=Country

1) a=0.05
2) H₀: M₁ = M₂ = M₃=0 (no difference in the median MTBW for different countries)
3) H_a: some differences in the median MTBWs for different countries
4) p-value=0.001 < a
5) We reject H₀, there are some differences in the median MTBW for different countries

Step 5: Multiple Comparison
Not possible (or at least not with MINITAB) for Kruskall-Wallis

Warning If we had just done the ANOVA Country would not have been stat. significant (p-value=0.098)