The following is a thorough list of mathematical calculations/explanations related to Essay:An analysis of another misleading poll.
Finding sample makeup by gender
According to the polling data for the 2020 Senate special election in Arizona, 34% of respondents intend to vote for Republican Martha McSally and 49% prefer Democrat Mark E. Kelly. For the two-way tables that involve the preferences by gender, 34% of men intended to vote for McSally, and 52% for Kelly; for women, 34% back McSally and 46% are for Kelly. Since it's unknown exactly what percent of the sample makeup is by gender, M can be used as a variable for male respondents and F for female respondents. Since the percent of men (out of the entire sample) that support McSally and the percent of women (again out of the entire sample) that support McSally must add up to 34%, the equation "34M + 34F = 34" can be set. With the same notion applying to Mark Kelly, the equation "52M + 46F = 49" can be set as well. And to solve:
34M + 34F = 34 → → → 34M = –34F + 34 | 52M + 46F = 49 → → → 52M = –46F + 49 ↓ ↓ ↓ ↓
− 34F = 34F → → → ÷ 34 = ÷ 34 | – 46F = 46F → → → ÷ 52 = ÷ 52 ↓ ↓ ↓ ↓
——————————————————————— → → → ————————————————— | ——————————————————————— → → → ————————————————————————— ↓ ↓ ↓ ↓
34M = –34F + 34 → → → M = –F + 1 | 52M + = –46F + 49 → → → M = –(46/52)F + 49/52 ↓ ↓ ↓ ↓
↓ ↓ ↓ ↓
↓ ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ↓ ↓ ↓
↓ ↓ ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ↓ ↓
↓ ↓ ↓ ↓ ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ←
↓ ↓ ↓ ↓
–F + 1 = –(46/52)F + 49/52 → → → (6/52)F + 49/52 = 1 → → → (6/52)F = 3/52
+ F = (52/52)F → → → – 49/52 = 49/52 → → → × (52/6) = × (52/6)
——————————————————————————— → → → ———————————————————————— → → → ——————————————————
1 = (6/52)F + 49/52 → → → (6/52)F = 3/52 → → → F = 3/6
Since exactly half of the sample makeup are women, the other half must be men. Plugging these figures back into the original equations result in identities, so the answers yielded are correct.
Finding sample makeup by age
Given the way the data had been set up for age groups, it is impossible to find the exact, correct answer for the makeup. Six age groups are given (which means six variables to solve for), and since there are only four options to choose from for candidates ("McSally", "Kelly", "Other", and "Don't Know"), that means that only four equations (each involving the six variables) can be set up. Since there more variables than equations, it is impossible to cancel out all enough variables to find out what any one of them are equal to in terms of hard numbers (rather, each variable can only be expressed in terms of other variables); thus there are infinitely many solutions possible.[1]
Finding sample makeup by region
According to the statistics for preferences by region, 31% of those in Maricopa County back Martha McSally in addition to 40% in Pima County and 37% from the rest of the counties. For Mark Kelly, 51% in Maricopa County back him, along with 53% in Pima County and 40% in the other counties. Also, 2% in Maricopa answered "Other", along with another 2% from Pima and 6% from the rest of the counties. With these given numbers, a three-variable system can be set up. Since the 31% in Maricopa County that back McSally along with the 40% in Pima County and 37% in the other counties that follow as such must add up to 34% of all respondents, the equation "31M + 40P + 37T = 34" can be set (note that "T" is used for "Other" rather than "O", as the latter can be confused with the number zero). Following the same idea for Kelly and "Other", the respective equations "51M + 53P + 40T = 49" and "2M + 2P + 6T = 3" can set up as well. Now to solve:
31M + 40P + 37T = 34 | 51M + 53P + 40T = 49 → → 10(2M + 2P + 6T) = 10(3) | 20(2M + 2P + 6T) = 20(3) ↓ ↓ ↓ ↓
51M + 53P + 40T = 49 | – 31M + 40P + 37T = 34 → → – 20M + 13P + 3T = 15 | – 31M + 40P + 37T = 34 ↓ ↓ ↓ ↓
2M + 2P + 6T = 3 | —————————————————————— → → ——————————————————————————— | ———————————————————————————— ↓ ↓ ↓ ↓
| 20M + 13P + 3T = 15 → → 7P + 57T = 15 | 9M + 83T = 26 ↓ ↓ ↓ ↓
↓ ↓ ↓ ↓
↓ ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ↓ ↓ ↓
↓ ↓ ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ↓ ↓
↓ ↓ ↓ ↓ ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ←
↓ ↓ ↓ ↓
7P + 57T = 15 → → 9M + 83T = 7P + 57T + 11 → → 9M + 26T = 7P + 11 → → 26T = 7P – 9M + 11
+ 11 = 11 → → – 57T = 57T → → – 9M = 9M → → ÷ 26 = ÷ 26
———————————————————— → → —————————————————————————— → → ————————————————————————— → → —————————————————————————————————
7P + 57T + 11 = 26 → → 9M + 26T = 7P + 11 → → 26T = 7P – 9M + 11 → → T = (7/26)P – (9/26)M + 11/26
↓ ↓ ↓ ↓
↓ ↓ ↓ ↓
↓ ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ↓ ↓ ↓
↓ ↓ ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ↓ ↓
↓ ↓ ↓ ↓ ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ←
↓ ↓ ↓ ↓
31M + 40P + 37((7/26)P – (9/26)M + 11/26) = 34 | 2M + 2P + 6((7/26)P – (9/26)M + 11/26) = 3 ↓ ↓ ↓ ↓
↓ ↓ ↓ ↓ | ↓ ↓ ↓ ↓ ↓ ↓ ↓ ↓
31M + 40P + (259/26)P – (333/26)M + 407/26 = 34 | 2M + 2P + (42/26)P – (54/26)M + (66/26) = 3 ↓ ↓ ↓ ↓
↓ ↓ ↓ ↓ | ↓ ↓ ↓ ↓ ↓ ↓ ↓ ↓
(1299/26)P + (473/26)M + 407/26 = 34 | (94/26)P – (2/26)M + (66/26) = 3 ↓ ↓ ↓ ↓
↓ ↓ ↓ ↓ | ↓ ↓ ↓ ↓ ↓ ↓ ↓ ↓
(1299/26)P + (473/26)M = (477/26) | (94/26)P – (2/26)M = 12/26 ↓ ↓ ↓ ↓
↓ ↓ ↓ ↓ | ↓ ↓ ↓ ↓ ↓ ↓ ↓ ↓
1299P + 473M = 477 | 94P – 2M = 12 ↓ ↓ ↓ ↓
↓ ↓ ↓ ↓ | ↓ ↓ ↓ ↓ ↓ ↓ ↓ ↓
473M = 477 – 1299P | –2M = 12 – 94P ↓ ↓ ↓ ↓
↓ ↓ ↓ ↓ | ↓ ↓ ↓ ↓ ↓ ↓ ↓ ↓
–(1299/473)P + (477/473) = M | M = –6 + 47P ↓ ↓ ↓ ↓
↓ ↓ ↓ ↓
↓ ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ↓ ↓ ↓
↓ ↓ ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ↓ ↓
↓ ↓ ↓ ↓ ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ← ←
↓ ↓ ↓ ↓
–(1299/473)P + 477/473 = –6 + 47P | M = –6 + 47(51/362) | 2(225/362) + 2(51/362) + 6T = 3
↓ ↓ ↓ ↓ | ↓ ↓ ↓ ↓ | ↓ ↓ ↓ ↓
477/473 = –6 + (23530/473)P | M = –6 + (2397/362) | 450/362 + 102/362 + 6T = 3
↓ ↓ ↓ ↓ | ↓ ↓ ↓ ↓ | ↓ ↓ ↓ ↓
(3315/473) = (23530/473)P | M = 225/362 | 552/362 + 6T = 3
↓ ↓ ↓ ↓ | | ↓ ↓ ↓ ↓
P = 3315/23530 | | 6T = 534/362
↓ ↓ ↓ ↓ | | ↓ ↓ ↓ ↓
P = 51/362 | | T = 89/362
Fully solving the three-variable system, the resulting numbers for sample makeup by region are ~62.15% from Maricopa County, ~14.09% from Pima County, and ~24.59% from the rest of Arizona. Note that the sum of these percentages slightly exceeds 100%.[2]
Finding sample makeup by 2016 presidential vote
According to the data for preferences between McSally and Kelly by the respondents' votes in the 2016 U.S. presidential election, 71% of those who voted for Trump in 2016 intend to back McSally, along with 4% who voted for Hillary Clinton, 14% who had back someone else, and 13% didn't vote then. For Mark E. Kelly, 15% of respondents who voted for Trump intend to support him, along with 86% who voted for Clinton, 46% who went third party, and 43% who didn't vote. For "Other", this response was given by 2% who voted for Trump then, 2% who voted for Clinton, 17% who voted for another candidate, and 6% who didn't vote. And for "Don't Know", this response was given by 11% whose vote went for Trump, 8% who went for Clinton, 23% who had voted for someone else, and 37% who didn't vote then. With such, the respective equations "71D + 4C + 14T + 13V = 34",[3] "15D + 86C + 46T + 43V = 49", "2D + 2C + 17T + 6V = 3", and "11D + 8C + 23T + 37V = 14" can be formed. Since there are four variables set up into four separate equations, a system can be set up and solved. However, due to the length it would take to solve/simplify step-by-step, WolframAlpha will be used, unlike for the previous section. According to their computation, ~42.3613% of respondents voted for Donald Trump in 2016, ~40.78134 voted for Hillary Clinton, ~2.649277% had voted for someone else, and ~14.7795% didn't vote then. Note that the sum of these percentages is slightly above 100%.[2]
Finding sample makeup by intended 2020 presidential vote
According to the data given for preferences in the 2020 U.S. Senate election in Arizona by the intended vote in the concurrent presidential election, 76% of those who plan to vote for Trump also intend on supporting Sen. McSally, along with 3% who intend on voting for Joseph Robinette Biden, Jr., 20% of those who answered "Don't Know", and 46% of those who don't intend on voting in the presidential election. For Kelly, 11% of those supporting Trump intend on voting for him, along with 90% of Biden supporters, 22% of those undecided, and 36% who won't vote. For "Other", this was the response from 1% of Trump supporters, 1% of Biden supporters, 8% of those who answered "Don't Know", and 7% who won't vote in the presidential election. And for "Don't Know", this was the response given by 12% of those intending on voting for Trump, 5% of those voting for Biden, 51% of the "Don't Know", and 11% of those who won't vote in the presidential election. Following the same notion as the section above in setting up a four-variable system and using WolframAlpha to solve, the respective equations are "76T + 3B + 20D + 46W = 34", "11T + 90B + 22D + 36W = 49", "T + B + 8D + 7W = 3", and "12T + 5B + 51D + 11W = 14". Setting up this information on WolframAlpha, the resulting numbers are that ~28.6522% of respondents intend to vote for Trump, ~40.4671% intend to vote for Biden, ~12.7776% don't know, and ~18.38% won't vote in the presidential election; note that the sum of these numbers slightly exceeds 100%.[2]
Notes
- ↑ Further explanation: Normally in a 2-D graph, two lines would intersect at one point. In a 3-D graph, two planes would intersect at a line and three planes would intersect at one point, etc. Suppose a point can be considered to the 0-D analogue of a square, a line an infinitely extending 1-D analogue of a square, a plane a square extending infinitely, etc. Now suppose that two infinitely extending 5-D analogues of squares are graphed on a 6-D plane. The intersection would be an infinitely extending 4-D analogue of a square. Now graph another equation, which now yields three infinitely extending 5-D analogues of squares that intersect at an infinitely extending 3-D space. Graph yet another equation, and the four graphed infinite extensions of 5-D analogues of squares will intersect at a plane. Because that plane has infinitely many points lying on it, with each point representing a valid solution, there are an infinite amount of valid ordered pairs, thus an infinite amount of solutions.
- ↑ 2.0 2.1 2.2 Since the original percentages given are more likely than not rounded from long decimal figures, the sum of the final answers may be slightly off from exactly 100%.
- ↑ Since "T" is used for "Other" rather than "O" (to avoid confusion with the number zero), "D" is used for Donald Trump rather than T. And since "D" is used for Trump, "V" is then used for "Didn't Vote".