Using the mean $\mu = n \times p$ ($\text{sample size} \times \text{probability}$), compare it to the observed value. If it is more, than test upper, otherwise lower.