Determining the confidence interval for non-probabilistic surveys: Method proposal and validation

This paper proposes a novel method for determining confidence intervals in non-probabilistic survey samples, addressing the challenge of statistically quantifying uncertainty in such samples. The method utilizes combinatorial calculations to assess the myriad combinations possible in the unsampled population, applying the symmetric property of binomial coefficients to optimize computational efficiency. The proposed approach offers a practical solution, making it feasible to calculate confidence intervals even with large sample sizes. This methodology represents a significant advancement in survey research, providing a tool to evaluate and generalize non-probabilistic data, particularly in contexts where probabilistic sampling is not viable. The paper also validates the method using extensive testing, demonstrating its applicability and effectiveness in real-world scenarios.