Final answer:
The type of bias in the survey is nonresponse bias. The polling organization should try contacting households that do not respond by phone or face-to-face and consider mailing the questionnaire to a greater number of households.
Step-by-step explanation:
The type of bias in this survey is nonresponse bias. This bias occurs when a significant number of selected households do not respond to the survey. In this case, only 21 households out of 1,476 responded.
To remedy this bias, the polling organization should try contacting households that do not respond by phone or face-to-face. This would increase the response rate and ensure a more representative sample.
Additionally, the organization could also consider mailing the questionnaire to a greater number of households to increase the chances of getting a higher response rate. This would help reduce the impact of nonresponse bias on the survey results.