Home > Preview
The flashcards below were created by user
on FreezingBlue Flashcards.
Some numerical facts are called parameters.
- Average age of all eligible voters.
- the percentage of all eligible voters who are currently registered to vote.
survey organization use percentage points as the units for the difference between actual and predicted percents.
should be fair, selection people for inclusion in the sample in an impartial way so as to get a representative cross section of the public
A systematic tendency on part of the sampling procedure to exclude one kind of person or another from the sample is called Selection Bias
When a selection procedure is biased taking a large sample does not help. This just repeats the basic mistake
- Happens when people don't respond to surveys.
- If there is a high non-response rate look out for non response bias.
- lower income and higher income people tend not to respond to questionnaires
(Not a Probable Method)
- Interviews was assigned a fixed quota of subjects to interview.
- quotas was assigned to certain categories. sex, age, race, and economic status were fixed.
- interviewers where able to select anyone they liked (hand-picked)
- left room to human choice is always bias (Not Randomized)
- samples to resemble population with key characteristics. method seems reasonable, but does not work very well because of unintentional bias on the part of the interviewers.
- alternative is to us objective and impartial chance mechanisms to select the samples
- Many surveys are carried-out using Probability Methods.
- Two Important Features:
- 1. The interviewers have no discretion at all as to whom they interview
- 2. there is a definite procedure for selecting the sample, and it involves the planned use of chance
- sampling at random (taking names from a box without replacement)
- Samples are to be representative of entire population giving chance error/SE
- able to predict outcomes with great accuracy with small percentage of population.
Simple Random Sampling
- tickets drawn at random without replacement
- at each draw each ticket has an equal chance to be chosen
- interviewers have no discretion who they interview
- procedure in selecting samples are impartial.
- every has a chance to get in sample
- Procedure used for elections from 1952-1984
- 4 stages were used.
- Stage 1: divided into 4 geographical region
- stage2: simple sample divided into wards
- Stage 3: wards divided into Precincts by
- stage 4: Precincts are divided into homes
- Stage 5: homes are divided into individuals
Each Stage is carried-out by simple sampling
- In presidential election between 1/3 & 1/2 of eligible voters fail to vote
- Non voters are irrelevant and should be screened out of the sample as far as possible.
- screening out increases accuracy for election forcast.
- People who have not decided
- surverys try to minimize undecided individuals with trick questionnaire questions, to keep undecided individuals to a minimum
- information might be used to tell how voters will likely vote, but hard to tell
- answering giving in surveys are influenced to some extent by the phrasing of the questions; even tone or attitude of interviewer.
- example is changing the order of candidates names was found to change the response by 5%
- to control response bias all interviewer use same questionnaire and standardized interview procedures.
non- response bias
- Many subjects are missed or not a home
- this bias can be adjusted by givin more weight to subject who were available but hard to reach with questions that ask "whether they were home the other day)
- Making sure interviews follow instruction
- some question are redundant to check for consistency.
- a small percentage may be reinterviewed by and administrator to check for consistency
- Many surveys are now conducted by telephone.
- phone number are selected using multistage clustering based on area codes
- simple random sampling is conducted using computer (excludes businesses)
- a lot cheaper than in person interviewing (about 1/3 cheaper)
- Phone surveys are conducted on weekends when more people are likely to be home. Will call back up to 15x if no answer
- high and lower class usually left out. Middle class mostly survey.
Chance error and Bias
- 1. Chance Error is often call Sampling Error
- Example: people change their minds, not home, don't answer, don't vote, etc.
- 2. Bias is often call not Non-Sampling Error
- example: People don't respond
- Bias is more serious problem than chance error.
Home > Flashcards > Print Preview