How can these polls be accurate when everyone has cellphones, and there is no directory?

Question

Seriously, I have no "land line". The only people I know of that do are folks like my parents, who have simply had the same number for 40 years. I know hardly anyone under 40 does. It is just a silly expense when you have the cell phone... Point is, unless I am mistaken the numbers these poll people call are all from public directories, which means they are only calling the people with land lines. And that would make the poll more than invalid. If that is the case it would be horribly manipulative to present these results as an accurate representation the populace. Propaganda even. Please enlighten me.

Answer 1

Pollsters are becoming concerned about the issue of land lines versus cell phones, especially among younger voters. However, contrary to your experience, many young voters still have land lines.

Polling is a matter of statistical science. The first question for the firms doing the polling (and there are other forms of survey done beyond mere random sampling) is whether there is a significant difference in the opinions of those under forty who have land lines and those under 40 who have cell phones. So far, the answer is no and there is still a sufficiently large pool of people with land lines to allow for random selection.

The second aspect of statistical science is the assignment of weights to the sample. A long time ago, pollsters did find statistically significant differences between certain groups of respondents that tended to be consistent over time (e.g. age, race, gender, income status). To assure accurate results, you need to reweight your sample to accurately reflect percentages in the population. Thus, if white males between 20-30 should be 8% of your response group and (due to the problem of cell phones) that group is only 4% of your response group, you give more weight to those answers so that it does end up counting as 8% of the total pool. This technique is considered to be scientifically accurate.

Thus, while you raise a valid issue, at the present time, there is no evidence that it alters the results of the polls.

The best way to explain polling (and why this is not an issue) is to imagine 3 groups of rubber balls (a red-colored group, a white-colored group, and a blue-colored group). We know that, in the overall population, we have 1000 of each of these groups. We now want to determine without looking at each ball, how many have a stripe, how many have an x, and how many have no marking. Before doing our sample, we dump 500 of the blue-colored balls out. We then combine the remaining blue-colored balls with the red-colored balls and the white-colored balls. Out of this group, we randomly select 81 balls and note their color and whether they have a stripe, an x, or no marking. We then adjust the weight of the results so that the results from each color represents one-third of the total pool. Assuming that there was nothing special about the blue balls we discarded (in terms of our question), there should be no difference between this result and one in which we had not discarded any blue balls.

Answer 2

Pollsters generally ask before either before or after the interview what the persons age, party affilitaion, tax bracket and the likes, that is how htey group their numbers.

And they don't just get publically listed numbers, if you ever gave your number for something on the internet, they can frequenlty sell your number to these pollsters.

Technically you can block them by using the do not call list, but it doesn't work. I am not listed, but my wife gave out our number on the internet to a mortgage company, now I am inundated every night with calls. I have a land
line for my internet DSL. Many places that is the only way to get online. It is far cheeper than trying to use my cell to surf.

Answer 3

Do you have statistics to hold up your claim that only people over 40 have landlines? how absurd!