I am conducting a study using questionnaire survey for tow purposes: first: ranking risk factors effecting time delay in construction by taking the average score of respondents based on five point likert-scal, second: making prediction using Bayesian network. For the first objective, One paper aimed only to rank risk factors suggested to consider Z= 2.575 for 99% confidence level, p percentage picking choice = 0.5, C margin of error= 9%, for unknown population the sample size is =2.575^2*0.5*(1-0.5)/0.09^2=205 respondents. Random sampling was adopted.
Given that my population is unknown, can I take the same confidence interval and margin of error to calculate sample size for prediction purposes? as many colleagues take Z and C equal to 1.96 and 0.05 instead of 2.575 and 0.09?
Any suggestion to reduce sample size?
- Badges
- Science topic
- Similar topics
- Mathematics
- Statistics
- Sampling Design
- Sample Size