WEBVTT 00:00:00.000 --> 00:00:02.455 MALE NEWS PRESENTER 1: A major winter storm [OVERLAPPING] 00:00:02.455 --> 00:00:05.340 MALE NEWS PRESENTER 2: The bruising fight [NOISE] to win Tuesday's Republican [OVERLAPPING] 00:00:05.340 --> 00:00:07.960 MALE NEWS PRESENTER 3: Most dangerous virus. [NOISE] 00:00:09.050 --> 00:00:13.050 MR GINGRICH: With your help, we're going to win a historic victory. 00:00:13.050 --> 00:00:17.060 FEMALE NEWS PRESENTER: The 2012 presidential campaign is in full swing, 00:00:17.060 --> 00:00:20.370 with Republican candidates battling state by state to 00:00:20.370 --> 00:00:23.850 determine who will face President Barack Obama on election day. 00:00:23.850 --> 00:00:26.910 MALE NEWS PRESENTER 4: In the poll numbers we are debuting tonight, 00:00:26.910 --> 00:00:29.770 there is a new GOP front-runner in this race. 00:00:29.770 --> 00:00:33.930 FEMALE NEWS PRESENTER: A crucial part of this grueling electoral process is polls, 00:00:33.930 --> 00:00:36.930 the almost daily snapshots of public opinion, 00:00:36.930 --> 00:00:40.290 that help measure who's up and who's down among the candidates. [MUSIC] 00:00:40.290 --> 00:00:42.470 PRESIDENTIAL CANDIDATE: Thank you. 00:00:42.470 --> 00:00:47.440 DR VINCENT HUTCHINGS: The aim of an opinion poll is to get a sense about a population and where they [NOISE] 00:00:47.440 --> 00:00:52.670 stand on some particular set of issues or policies to measure their attitudes and effect. 00:00:52.670 --> 00:00:53.920 FEMALE NEWS PRESENTER: [MUSIC] Vincent Hutchings, 00:00:53.920 --> 00:00:57.940 an NSF-funded professor of political science at the University of Michigan, 00:00:57.940 --> 00:01:03.420 says that public opinion polls rely on a concept in statistics known as random sampling. 00:01:03.420 --> 00:01:06.800 The idea that it is possible to draw a clear picture about 00:01:06.800 --> 00:01:10.860 the feelings of a large group of people by examining how a small, 00:01:10.860 --> 00:01:14.190 randomly assembled slice of that group feels. 00:01:14.190 --> 00:01:19.470 DR VINCENT HUTCHINGS: The process involves interviewing a very small subset of that population 00:01:19.470 --> 00:01:25.035 and doing so in a scientific manner so that you'll have a better accurate sense 00:01:25.035 --> 00:01:26.670 but more efficient sense, 00:01:26.670 --> 00:01:29.170 of what the entire population understands. 00:01:29.170 --> 00:01:32.220 FEMALE NEWS PRESENTER: An opinion poll might focus on a particular state, 00:01:32.220 --> 00:01:37.735 such as Florida, or the entire US population of more than 300 million people. 00:01:37.735 --> 00:01:41.500 Pollsters use computer programs to generate a random list of 00:01:41.500 --> 00:01:45.500 a few hundred or a few thousand telephone numbers from the larger group. 00:01:45.500 --> 00:01:50.040 Then they call each number to survey people for their opinions. 00:01:50.040 --> 00:01:53.920 Hutchings says making a random sample is a lot like cooking soup. 00:01:53.920 --> 00:01:56.960 You don't need to eat the whole pot to know if it tastes good, 00:01:56.960 --> 00:01:59.345 you just need a spoonful. 00:01:59.345 --> 00:02:03.580 DR VINCENT HUTCHINGS: The cook merely needs to get a spoon out, taste it. 00:02:03.580 --> 00:02:05.860 That spoon represents a sample, 00:02:05.860 --> 00:02:08.920 as it were, of the actual contents of the pot. 00:02:08.920 --> 00:02:12.450 FEMALE NEWS PRESENTER: [MUSIC] Samples, in general, give better poll results when they include 00:02:12.450 --> 00:02:14.770 more characteristics of the whole population 00:02:14.770 --> 00:02:17.950 and in the same proportions as the population. 00:02:17.950 --> 00:02:19.790 Back to the soup analogy, 00:02:19.790 --> 00:02:21.430 if you're cooking chicken soup, 00:02:21.430 --> 00:02:24.530 you want your sample to include the ingredients such as chicken, 00:02:24.530 --> 00:02:28.865 noodles, and broth in the same proportions as the overall soup. 00:02:28.865 --> 00:02:32.230 DR VINCENT HUTCHINGS: We want to make sure that the sample has 00:02:32.230 --> 00:02:35.390 the population characteristics that are 00:02:35.390 --> 00:02:39.815 demonstrated in the larger population to which we want to make an inference. 00:02:39.815 --> 00:02:43.860 FEMALE NEWS PRESENTER: When reading a poll, it's important to also study the fine print, 00:02:43.860 --> 00:02:45.460 usually at the bottom. 00:02:45.460 --> 00:02:48.120 You'll find information there about how the poll was 00:02:48.120 --> 00:02:51.375 conducted and what the size of the random sample was. 00:02:51.375 --> 00:02:54.000 You'll also find the margin of error, 00:02:54.000 --> 00:02:56.880 a number with a plus or minus sign in front of it. 00:02:56.880 --> 00:03:00.025 This number tells you the range of accuracy of the poll. 00:03:00.025 --> 00:03:02.940 In this case, 3-5% points. 00:03:02.940 --> 00:03:05.580 Typically, the smaller the margin of error, 00:03:05.580 --> 00:03:07.730 the more accurate the poll is. 00:03:07.730 --> 00:03:12.020 DR VINCENT HUTCHINGS: What that means is that the number that's reported in that survey, 00:03:12.020 --> 00:03:15.780 we have a sense that given the size of the sample, 00:03:15.780 --> 00:03:19.450 and given the level of uncertainty associated with that size, 00:03:19.450 --> 00:03:23.755 the number could be 3-5% points higher or lower. 00:03:23.755 --> 00:03:25.770 But we know it's in that range. 00:03:25.770 --> 00:03:27.440 FEMALE NEWS PRESENTER: [MUSIC] Even though the polls rely on 00:03:27.440 --> 00:03:30.840 just a slice of the population to gauge public opinion, 00:03:30.840 --> 00:03:34.200 they are far more accurate than you might think. 00:03:34.920 --> 00:03:39.875 [NOISE] Which is one reason why they play such a special role in politics. 00:03:39.875 --> 00:03:43.900 DR VINCENT HUTCHINGS: Public opinion polls provide us with a way 00:03:43.900 --> 00:03:49.855 absent an actual election to discern where the public stands on various issues. 00:03:49.855 --> 00:03:52.620 FEMALE NEWS PRESENTER: As election day 2012 draws closer, 00:03:52.620 --> 00:03:55.100 the science of public opinion polls will help give 00:03:55.100 --> 00:03:59.920 a clear snapshot of who might be our next president.