If I compare two different population samples and the p-value > 0.05 can I assume that samples came from same population (and can therefore be merged)? | |
What happen if p-value < 0.05, can I assume that populations came from different population? |
But coincidence is always a possible alternative explanation. After all, the definition of p=0.05 is that the observed difference DOES occur 5% of the time by chance when the null hypothesis (in our case, when the populations are the same) are true. Statistics never gives 100% conclusions.
So what you conclude should depend not merely on the p-value, but on what other information you have as well. If two labs both sampled people who claim to be Malay, and the p-value comparing them is not small, then we might as well combine the samples. If the p-value are small, then it might be a good idea to check the work and the methodology of collecting the data carefully. After careful checking we might decide that we did excellent science and the low p-value is therefore just coincidence, and the samples still can be combined.