A political pollster was interested in whether there’s any link between age and political preference.
The pollster interviewed $15$ randomly chosen subjects ages $18-24$ and found that $9$ of them said they voted Democrat the last election. This compares to $15$ randomly chosen subjects ages $50$+ where $5$ of them said they voted Democrat last election.
A two-sided test gave a p-value of $0.1432$ and test statistic of $1.46$. The pollster was confused why the sample proportions were so different ($60\%$ vs. $33.33\%$) yet failed to show any significant differences at any common alpha levels ($1\%, 5\%$, or $10\%$).
How might you describe this discrepancy to the pollster?