Being a survey statistician involves much more than calling people and asking who they plan to vote for in the next election. Here are some areas in which surveys are used, and some of the problems that a survey statistician might work on:

• Government Surveys

Many U.S. government agencies conduct surveys to collect data used in public policy decisions. The best known of these is the decennial Census of Population, conducted by the U.S. Census Bureau. Other large-scale surveys include the Current Population Survey, which collects data on employment, unemployment, earnings, and other labor market topics; the National Crime Victimization Survey, which measures crime against individuals; and the National Health Interview Survey, which collects information on the amount and effects of illness and disability.

A survey statistician might study:

• Efficient survey design
• Experimental methods for increasing response rate
• Accounting for nonresponse and undercoverage
• How to release data to the public but maintain confidentiality of respondents
• Social Sciences

Social research depends heavily on surveys, which can provide a snapshot or moving picture of social trends. In many instances, the only way to obtain information about behavior or opinions is to ask a representative sample of people. Political polls fall in this category.

A survey statistician might study: 

• Question wording and design
• Methods for follow-up of nonrespondents
• Deciding where to take samples in an archaeological dig 
• Analyzing survey data collected for another purpose, for example, using the Current Population Survey to study trends in marital status
• Education

Educational research and improvement of the educational system requires that performance and subsequent development be measured. A survey statistician might study: 
• Designing surveys to include underrepresented groups
• Longitudinal surveys that follow a cohort of children to adulthood
• Law

Survey research is commonly used in the legal arena. In trademark cases, consumers of a product may be surveyed and asked whether a trademark is associated with a certain brand even when the company's name is missing. Surveys may also be used by attorneys to help with jury selection in high-profile cases. 
• Forestry and Agriculture

Surveys are crucial for estimating agricultural production, and many commonly-used sampling designs were developed for agricultural surveys. 
• Biology

How do we estimate the number of caribou in a region? How many mosquitoes in New York are estimated to carry West Nile virus?

A survey statistician might study: 

• How to sample from a population when no list of the population exists (for example, the mosquitoes in New York)
• Estimating sizes of wildlife populations
• Avoiding bias in samples of wildlife (for instance, because it might be easier to catch infected animals)
• Medicine and Public Health

The National Health Interview Survey mentioned above is just one of the surveys conducted by the Centers for Disease Control to monitor public health. State governments and medical researchers conduct numerous surveys to estimate prevalence of disease and costs to society.
A survey statistician might study: 
• How to sample persons with rare disorders if no national registry exists
• How to estimate effects of environmental toxins when comparing samples from neighboring cities
• Methods for combining data from a specialized survey of AIDS patients with national survey data
• How to profile and explain the extent of cardio-vascular disease, back pain, and other chronic medical conditions
• Methods to assess the impact of specific community programs to promote health through exercise, better eating habits, and other strategies
• Business

Market research commonly relies on surveys and focus groups. Economic surveys such as those conducted by the Bureau of Labor Statistics may be used in making business decisions. 
• Internet

The internet has the potential to become a widespread medium for conducting surveys. (After all, you're reading this page, aren't you?) To date, however, most surveys have been of the "click here for yes, click there for no" variety, and the fact that web viewers decide whether or not to respond makes these surveys unreliable. How can the internet be used to conduct reliable, statistically valid surveys?