In survey sampling, what does the design effect describe?

• A. The factor by which the sample size must be increased to account for clustering
• B. The influence of design on randomization
• C. The bias introduced by nonresponse
• D. The cost savings from sampling

Answer: (A)

The design effect is about how much the sampling design changes the precision of your estimates compared with simple random sampling. It is the factor by which the variance increases (and thus the sample size would need to be increased) to achieve the same level of precision when clustering or other complex design is used. In cluster sampling, observations inside the same cluster are more similar, so you gain less independent information per unit, making the design effect greater than one. So you multiply the planned sample size by this factor to maintain the same accuracy as a simple random sample. The other ideas describe different issues (randomization, nonresponse bias, or cost savings) and don’t capture what the design effect measures.