Judge news, polls and political ads: bias, wording, sample size and context

Grade 12 Social Studies Worksheet

Circle the best answer.

1. Read this first

   Use this

   Cards A to C on two polls of one race in Marlow (invented, for practice). Card A, "Rule of thumb" (summary in our words): For a poll of n people chosen at random, the margin of error is about 100 divided by the square root of n, in percentage points. For a change between two separate polls of the same size, the margin is about 1.4 times that. A poll in which people choose themselves to answer is not a random sample, so the rule does not apply to it. Card B, "Poll X": a phone poll of 625 registered voters chosen at random, November 1 to 3. Ames 52 percent, Boyd 48 percent. Card C, "Poll Y": an online poll of 1,600 likely voters chosen at random from the voter list, November 2 to 4. Ames 47 percent, Boyd 53 percent.

2. 1.

   Ari compares Cards B and C. Using Card A, do the two polls contradict each other?

   • No: Y's larger sample shows X is wrong, so there is no real conflict to explain
   • Not clearly: Ames's ranges, 48 to 56 and 44.5 to 49.5, overlap
   • Yes: Ames's ranges, 48 to 56 and 44.5 to 49.5, do not overlap
   • Yes: X has Ames ahead and Y has Boyd ahead, so one of the polls must be wrong
3. 2.

   A third Marlow poll of 400 random adults finds Ames 50, Boyd 50. Bo uses Card A. How does its margin compare with X and Y, and what does it add?

   • About 5 points, which is irrelevant because the poll has a different sample than X and Y
   • About 2.5 points, narrower than Y, so it narrows the race to a tie
   • About 5 points, narrower than X, so it settles the race
   • About 5 points, wider than either; it fits a close race but narrows nothing
4. 3.

   Jun reads this claim: 'Poll Y is more accurate because its sample is bigger.' Which response do Cards A to C support?

   • Wrong: Poll X asked registered voters, a bigger group than likely voters, so X is more accurate
   • Wrong: a bigger sample gives a larger margin, so X is more accurate
   • Right: a bigger sample gives a smaller margin, 2.5 against 4, so Y is more accurate
   • Overstated: a bigger sample cuts sampling error, not bias in who was asked

5. Use this

   Cards A to C on two polls in Brook (invented, for practice). Card A, "Rule of thumb" (summary in our words): For a poll of n people chosen at random, the margin of error is about 100 divided by the square root of n, in percentage points. For a change between two separate polls of the same size, the margin is about 1.4 times that. A poll in which people choose themselves to answer is not a random sample, so the rule does not apply to it. Card B, "Two polls about bike lanes": In March, 400 randomly chosen adults were asked; 54 percent support the bike lanes. In June, 400 other randomly chosen adults were asked; 49 percent support. Card C, "A group in the June poll": 100 of the June 400 were adults under 30; 62 percent of them support the bike lanes.

6. 4.

   Luz reads Card C. The 100 adults under 30 in the June poll are 62 percent in support. What does the margin rule say about that result?

   • Too small to use, since a group of 100 cannot give a margin of error
   • About plus or minus 10 points, so 52 to 72 percent, with a wide range
   • Exactly 62 percent of adults under 30 in Brook support it
   • About plus or minus 5 points, so 57 to 67 percent, as for the whole poll
7. 5.

   Milo reads the June result in Card B: 49 percent with a margin of about 5. Can the poll say whether a majority of Brook adults support the bike lanes?

   • Yes, a majority: the range 44 to 54 is centered near 50
   • No: 44 to 54 includes 50, so the poll cannot say if a majority support them
   • No majority: 49 is below 50, so fewer than half support the bike lanes
   • No: the range is 39 to 59, so the poll is too wide to use
8. 6.

   Noah reads this claim: 'Support fell between March and June because of a news story about a bike crash.' Which response do Cards A to C support?

   • Unsupported: there is no way to find a reason for a change in a poll
   • Supported: June's lower support is the effect, and the story is the cause
   • Supported: the polls show support fell after the crash, which shows the story caused it
   • Unsupported: the cards give no reason, and the 5-point change is within its margin

9. Use this

   Cards A to D on a poll in Dunmore (invented, for practice). Card A, "Rule of thumb" (summary in our words): For a poll of n people chosen at random, the margin of error is about 100 divided by the square root of n, in percentage points. For a change between two separate polls of the same size, the margin is about 1.4 times that. A poll in which people choose themselves to answer is not a random sample, so the rule does not apply to it. Card B, "The poll": 625 randomly chosen adults who live within two miles of the main library were asked, "Do you support extending hours at the main branch?" Fifty-two percent support, 41 percent oppose, and 7 percent are not sure. Card C, "Headline 1": "Majority backs longer hours." Card D, "Headline 2": "Nearly half oppose or are unsure about longer hours." A third headline says: "Residents back longer library hours citywide."

10. 7.

    Pia reads Cards B to D. The first headline says 'Majority backs longer hours' and the second says 'Nearly half oppose or are unsure.' How should they be read?

    • Each picks a slice; 52 plus or minus 4 includes 50, so 'majority' goes too far
    • Both are false, since the poll's margin makes each percentage meaningless
    • Headline 1 is accurate and Headline 2 is false, since 48 percent is not nearly half
    • Headline 2 is false, since people who are not sure are not opposed
11. 8.

    Ren reads Card D. A third headline says 'Residents back longer library hours citywide.' What does it go beyond in Card B?

    • The question named one branch, and the sample lived near it
    • The wording of the question, which was loaded toward support
    • The sample size of 625, which is too small for any headline
    • The margin of error, which is about 4 points for 625 people
12. 9.

    Tala reads the third headline in Card D. Which fact about the sample limits the word 'residents'?

    • The poll had 7 percent not sure, so it cannot describe residents
    • The 625 adults were chosen at random, so they cannot stand for the residents
    • The margin of about 4 points, which is too wide for any resident claim
    • The 625 adults lived within two miles of the library, not citywide
13. 10.

    Zee reads this claim: 'The poll shows 48 percent of residents oppose longer hours.' Which response do Cards B and D support?

    • Right: 52 support and 48 do not, so 48 oppose
    • Wrong: 41 percent oppose; the other 7 percent are not sure
    • Wrong: the poll shows no one opposes, since 52 percent support
    • Right: 41 plus 7 is 48, and the unsure will end up opposing