Landline users are more likely to respond than cell phone users

People without landlines (often younger, lower-income) are excluded from the sample

The sample is too large to be representative

Random sampling cannot be done by phone