Polls

William Christensen 4/10/08 (13)
The headlines in our newspaper often say something like this: New Poll Shows That Utahans Approve of (insert something, usually bad). Then down below there will be a table that looks something like this:

Strongly Approve
30%
Somewhat Approve
25%
Somewhat Disapprove
25%
Strongly Disapprove
10%
Don’t Know
15%
This is a poll of 300 taken by Dave Jones &
Associates. It has a margin error of 7.5%

Most people read the poll, believe that 55% of Utahans somewhat or strongly approve of (insert something, usually bad). They may not even notice, but after reading stories and seeing polls all about how (something bad) is good, it affects the reader sub consciously.

Polls can be very misleading
Here are a few tips on how to decide if a poll is trustworthy or not.

How many people
Read the small print at the bottom of the poll! It tells you a lot of things about the poll. Look at the small print under the poll above. It starts out by saying: This is a poll of 300 Utahans. The news paper takes the liberty to say, because there is 55% that approve it in some way, that a New Poll Shows That Utahans Approve of (insert something, usually bad). The only thing this proves is that 165 people in Utah approve of (insert something, usually bad). And I would also like to know where these people were found. They could have been found in a rally for (insert something, usually bad). This common tactic used to skew polls. You decide if the amount of people polled is enough to say that on average all Utahans belive that way. For me, 300 is not nearly enough.

Biased polls
The second thing it says in the fine print: taken by Dave Jones & Associates. Lets say we know nothing about Dave Jones. So we go online and type in Dave Jones & Associates. Up pops a whole ton of information about him and his group. We find that his brother Jim Jones is a congressman, and his sister is married to the governor, and his wife is a lobbyist for (insert something, usually bad). We see that this guy probably has political bias. If a pollster has credentials like these, this is when I decide if the people were probably chosen over a wide area, and are a wide range of race, color, ect.

Margin error
What Wikipedia says: "All polls based on samples are subject to sampling error which reflects the effects of chance in the sampling process. The uncertainty is often expressed as a margin of error. The margin of error does not reflect other sources of error, such as measurement error. A poll with a random sample of 1,000 people has margin of sampling error of 3% for the estimated percentage of the whole population. A 3% margin of error means that 95% of the time the procedure used would give an estimate within 3% of the percentage to be estimated. The margin of error can be reduced by using a larger sample, however if a pollster wishes to reduce the margin of error to 1% they would need a sample of around 10,000 people. In practice pollsters need to balance the cost of a large sample against the reduction in sampling error and a sample size of around 500-1,000 is a typical compromise for political polls. (Note that to get 500 complete responses it may be necessary to make thousands of phone calls.)" Lets see what we can do to the poll with an error of 7.5 %:

Strongly Approve
22.5%
Somewhat Approve
17.5%
Somewhat Disapprove
32.5%
Strongly Disapprove
17.5%
Don’t Know
15%

Totally changes the meaning of the poll, doesn’t it? Plus, the margin error does not include any others errors they may have made.

Polls can be very useful
I am not saying that all polls are bad, just that we should be very careful before we believe in what one has to say.