Final answer:
Nonresponse bias occurs when survey participants are not representative of the general population, leading to skewed results. This can happen in various forms, such as telephone surveys with noninclusive sample frames or self-selected samples from the internet, both of which contribute to biased survey results.
Step-by-step explanation:
Nonresponse bias exists when participants in a survey or study are not representative of the general population, often because not all individuals are equally likely to participate. The phenomenon can lead to skewed results, as it may exclude certain demographics or opinions. This was famously illustrated in the 'Bradley effect', a type of response bias evident in the 1982 California gubernatorial race, where societal pressures led to misleading poll results.
Another example of nonresponse bias can be seen in telephone surveys. With the advent of caller ID and the increase in cell phone use, fewer people are participating in random digit dial surveys, which raises concerns about the validity of such research. Factors such as not having a telephone, the exclusion of cell phone numbers from directories, and individuals not being at home during the call can contribute to a nonrepresentative sample, which in turn can lead to biased survey results.
It's important to note that large samples can still be biased, particularly in cases where participation is optional, such as internet surveys. Here, only those who choose to respond do so, which can again lead to a nonrepresentative sample. For instance, people may not respond to mail surveys, or the responses may not be from the intended demographic due to incomplete mailing lists or other factors.