Midterm Study Guide

Midterm Study Guide - POLI 100 Midterm Study Guide Survey...

Info iconThis preview shows pages 1–3. Sign up to view the full content.

View Full Document Right Arrow Icon
POLI 100 Midterm Study Guide Survey Research: Method and History – Lecture #1 Two problems afflict commercial surveys o Customer ordinarily has no basis with which to evaluate the quality of the research o Doing it right is expensive, doing it wrong can be profitable Most commercial work is low quality o Question isn’t whether they know how to do it right or not, because they do know o It is much more profitable to do it wrong Question of how to make an accurate survey comes down to two things o How do you select units to study? o How many do you need to get reliable results? Randomness o Selection is random if every unit in the group we care about has an exactly equal chance of being chosen o Eliminates all bias that you aren’t aware of History of Survey Research o Early polls were an attempt to find “almost random” way of doing things Sometimes would end in disaster o Literary Digest Example: Resulted in lasting lessons learned, and gave people a distrust of polls For 1932 election, mailed out ballots to millions of people through phone book Predicted FDR would decisively win, which he did o Seemed to validate the method Not random because left out those voters without phones, cars etc and only about 20% actually returned them Repeated in 1936, and predicted an Alf Landon victory Digest died shortly thereafter, as did peoples trust in polling Social Class bias Middle class returns questionnaires, working class doesn’t Types of Surveys o Straw Polls People choose whether or not to participate Such as in the Literary Digest example Possess no validity whatsoever If one happens to be correct, it’s simply luck o Dial in Polls These are the worst type Pay a long distance fee in order to vote So not only are they voluntary, but you must pay Conductors know they are garbage, at least the Digest didn’t know o Quota Sampling
Background image of page 1

Info iconThis preview has intentionally blurred sections. Sign up to view the full version.

View Full DocumentRight Arrow Icon
Originated as a solution to Digest’s problem of unrepresentative sampling Solution was forcing it to be representative by sampling in proportion to population quotas Street Corner bias Ex. If you did it on Franklin street you’d have a disproportionate amount of students and panhandlers Natural tendency to avoid shabby residences, long staircases, and homes with dogs o Two correct but more expensive ways to do it: o Area Probability Cluster Sampling Divide population into small units of area that can be covered by one interviewer; called Primary Sampling Units Selection areas randomly in proportion to their known population Within unit, select households randomly Within households, select family members randomly Expensive part: If you don’t find the person randomly chosen, you keep going back until you do get them, you don’t just replace them with someone available Extremely expensive, therefore not used in commercial surveys Used by main academic surveys o Random Digit Dialing
Background image of page 2
Image of page 3
This is the end of the preview. Sign up to access the rest of the document.

Page1 / 10

Midterm Study Guide - POLI 100 Midterm Study Guide Survey...

This preview shows document pages 1 - 3. Sign up to view the full document.

View Full Document Right Arrow Icon
Ask a homework question - tutors are online