measuring public opinion

Q: to understand public opinion among US adults, do we need to survery everyone

No. a well designed sample can be used to estimate the opinions of the larger population.

• scientific poling

What is the…

 The practice of securing a representative sample in a poll that accurately represents and predicts public opinion.

scientific polling

Public opinion polls: sampling

• Population (Entire subset you are studying or surveying)

  • Sampling

    • sample 

      • Subset of people from a population 

    • Inference

      • Making estimate

What is the Subset of people from a population

sample

What is the Entire subset you are studying or surveying

population

How to measure public opinion:

• Public opinion polls: scientific instrument for measuring public opinion 

  • Scientific:

    • Representative sample

    • Valid and reliable measurement

What is a public-opinion poll

scientific instrument for measuring public opinion 

What kind of sample is being descibed?

• a sample of a larger population that accuretaley mirrors the key characteristics, demographics, and traits of that entrie group

  • Sample errors

  • Representation errors 

representative sample

What kind of sample is being described?

•  survey measures that capture the concept they are designed to measure (valdity) and produce consistent results across repeated adminitstation (reliability)

  • Valididty: does the survey measure what it is intended to measure? 

  • Reliability: does the survey produce consistent results under similar conditions? 

Valid and reliable measurement:

What is this concept ..

• : does the survey measure what it is intended to measure? 

validity

What is this concept ..

does the survey produce consistent results under similar conditions? 

Reliability

Determine the reliability and valididty of these targets

• A: not valid, not reliable

• B: valid, not reliable

• C: not valid, reliable

• D: valid, reliable

Low validity or low reliability can lead to …

measurement error  

How to sample?

• Simple random sampling 

  • Selecting people from a population in such a way that each individiaual has an equal chance, or probability of selection

    • Systematic  sampling

    • Stratified sampling, by region 

    • Random digit dialing 

• In practice, random sampling (probability sampling) rarely follows a perfect simple random sampling process 

What is this called?

• Selecting people from a population in such a way that each individiaual has an equal chance, or probability of selection

  • Systematic  sampling

  • Stratified sampling, by region 

  • Random digit dialing

simple random sampling

In practice, random sampling (probability sampling) rarely follows…

a perfect simple random sampling process 

How to interpret a poll: sampling error chart

• 43% 

  • We can be 95% confident that the share of all US adults who 

  • 43-3.5 <actual share < 43 + 3.5

• You can do the same math for the rest of the percentages, to understand how it works if you want 

How to interpret a poll: sampling error

• margin of errror

• confidence interval

What is …

• An estimate of how much the sample result is likely to differ from the true population value because of sampling error

margin of error

• Margin of error = standard error x critical value= standard deviation/(square root of sample size)  x critical value

  • Dont need to know formula just rember relationship between the two

  • When sample size increase margin of error decrease**

  • Larger samples generally produce more precise estimates 

  • normally , samples of 1000 people are considered sufficient for acccuretale measuring public opinion through the use of surveys 

When sample size increases…..

margin of error decrease**

_____ samples generally produce more precise estimates 

• normally , samples of 1000 people are considered sufficient for acccuretale measuring public opinion through the use of surveys

larger

• Larger samples generally produce more precise estimates 

• normally , samples of _____ people are considered sufficient for acccuretale measuring public opinion through the use of surveys

1000

What is A range around the sample estimate, based on the margin of error, that is used to estimate the true population value 

confidence interval

Representation errors:

• Selection bias

  • Occurs when the process that determines inclusion is related to the variables being measured (non-random) 

  • The people you are surveruing is related to the topic your surveying, can create selection bias

  • Example: a survey about recycling behavior is answered mainly by people who already recycle  

    • Underrperesnt people who dont favor recycling and overrrepresent people who favor recycling

• Coverage bias

  • Occurs when the source material which a sample is drawn doesnt full match the target population 

    • Example: a survey of US adults only conatacts people with landline phones 

      • Underrepresent adults w/o landlines and overrepresents adults with landlines

Errors in sample selection can lead to underrpresentation or overrepresentation of certain groups in the sample

What is being described..

• Occurs when the process that determines inclusion is related to the variables being measured (non-random) 

• The people you are surveruing is related to the topic your surveying, can create selection bias

• Example: a survey about recycling behavior is answered mainly by people who already recycle  

  • Underrperesnt people who dont favor recycling and overrrepresent people who favor recycling

Errors in sample selection can lead to underrpresentation or overrepresentation of certain groups in the sample

selection bias

What is being descibed

• Occurs when the source material which a sample is drawn doesnt full match the target population 

  • Example: a survey of US adults only conatacts people with landline phones 

    • Underrepresent adults w/o landlines and overrepresents adults with landlines

Errors in sample selection can lead to underrpresentation or overrepresentation of certain groups in the sample

Coverage bias

Measure errors:

• Sources of measurement errors

  • Survey questions

    • Wording and order can produce different responses

  • Respondents

  • Interviewer 

  • Survey mode

    • Answers may differe whether the survey is online, by phone, or face to face 

• Social desrieability bias

  • Occurs when respondents report what they think is the socially acceptable responses, rather than what they actually believe or know to be true 

  • Example: a nonvoter reports that they voted because voting is socially valued  

Measure of errrors : Survey questions

Wording and order can produce different responses

Measures of errors survey mode

Answers may differe whether the survey is online, by phone, or face to face 

When does social desireablitity bias occur

• when respondents report what they think is the socially acceptable responses, rather than what they actually believe or know to be true 

• Example: a nonvoter reports that they voted because voting is socially valued