A survey aims to estimate the proportion of adults supporting a new policy with a margin of error of ±0.035 (±3.5 percentage points) at 95% confidence. Assuming the true proportion is unknown (worst case scenario), what minimum sample size is required?

Use: $ME = z^* \sqrt{\frac{p(1-p)}{n}}$

For 95% confidence, z* = 1.96. For unknown p, use p = 0.5 (maximum variance).