If population is normally distributed and n < 30, first find the value of ±t0.025 so that 2.5% ((100-95)%÷2) of the area is within each tail for n - 1 (or 13 - 1 = 12) degrees of freedom. Enter a t Distribution Table with df = 12 under 0.025 to obtain 2.179.
Then µ = X-bar ± 2.179(s/√n) or 14.3 ± 2.179(2/√13) or 14.3 ± 1.209. This gives 13.09 < µ < 15.51 with 95% confidence.
If population is not normally distributed, then use Chebyshev's Theorem: set 1 - (1/K2) = 0.95 or 0.05 = 1/K2 or K2 = √(1/0.05) and K ≈ 4.47. Then µ = X-bar ± K(s/√n) or 14.3 ± 4.47(2/√13). This gives 11.82 < µ < 16.78 with a 95% level of confidence.