How to lie with statistics

12. The steps in producing a “gee-whiz” graph involve truncating to the bottom and expanding the scale on the y-axis. (Pg. 62-63).

13. Making one picture twice as tall as another on a picture graph leaves an impression that one quantity is eight times as big because when one doubles the height of the picture to make it appear proportional; it becomes double the length as well. It makes the area four times as large but the mind perceives the picture as a representative of a three-dimensional object which then has eight times the volume. (pg. 69-70).

14. If I did not notice the scale, I would assign the Nissan a percentage of 95.3. Using a scale from 0 to 100, the version of the graph would be as shown below.

Model	Percentage (%)
Chevy	98.3
Ford	97.5
Toyota	96.5
Nissan	95.3

Graph showing the percentages of Chevy, Ford, Toyota, and Nissan

15. A semi-attached figure is a figure, fact, or argument that seems at first look as related to the point being made through after close inspection on it, it is not.        

16. In the excerpt from a guest commentary column in the Santa Maria Times by Ron Fink, there are several semi-attached figures that I can spot. The columnist’s thesis states that the lower air quality in Southern Santa Barbara County may be caused by the natural methane seepage. There is nothing in the article to prove the statement and almost the entire article has little evidence associated with it. The major semi-attached figures that I spot in the excerpt are:

A. how does 71 tons of methane generated on a daily basis compare to those released in the areas that meet the air quality standards. In consideration of the issue, how can the reader know that it is 71 tons? The quantity given appears and even sounds official by being precise though it is not clear whether there is a precise amount even for ten tons of methane.

B. There is also an issue with pollen which is not certain how to handle it. How does the detection equipment differentiate between pollen and methane?

C. Does the author suggest that the air quality in Santa Barbara is worse than that of the northern end of the county due to the volcanoes in the Philippines? How can ash that travels 100 miles from Mt. St. Helens in Oregon affect Santa Barbara on the central coast of California?

D. Another issue is to consider whether the ozone layer thinning over the Antarctic affect the air quality in Santa Barbara. The percentages given were not compared to any other existing figures such as the year before eruption, immediately before the eruption, the average thickness, volume, or the screening ability.

E. The author has not given any evidence at all that nature has more contribution to air pollution than the things listed in the conclusion. Even though it does, it is not certain whether it makes the things responsible for the difference between the two ends of the country.

17. Post hoc is a logical fallacy of believing that temporal or time succession implies a causal relationship. Post Hoc is a fallacy that takes the form of: A occurring before B, therefore A is the cause of B. The fallacy is committed when conducted that one event causes another simply because the proposed cause occurred before the proposed effect. The fallacy entails concluding that ‘A’ causes or caused ‘B’ since ‘A’ occurs before ‘B’ but there is inadequate evidence to warrant the claim. 

18. The various kinds of correlation that might lead to post hoc reasoning include:

Chance: One sample may have no correlation at all as such as that of Cancer versus milk (pg. 95-96)

Real relationship: It is uncertain to distinguish the cause and the effect such as the case of Spinsterhood versus college education (pg. 94)

Common cause: An example is that of smoking versus college grades (pg. 87), minister’s salary versus price of rum (pg. 90)

19. The chapter warns against the dangers of extrapolation. An example used in the discussion is Rainfall versus corn height and the years of education versus income (pg.91) whose extrapolation do not result to a true representation of the variables.

20. I think that the people that are most likely to statisticulate and the purpose include: the media to sensationalize, politicians to win the electorate votes, advertisers to sell their products, and anyone attempting to convince another of an issue especially where they have an opportunity to benefit from convincing the other person.    

21. From the maps on page 103, there is none of them drawn fairly without statisculation. It appears that there was a deliberate attempt to statisculate the maps for a defined purpose.

22. Percentages are often the source of statisculation because they mask the very small and largely meaningless results thereby hiding their counts. Percentages can also create different impressions depending on what is selected as the base which in most instances is not stated. Percentages are also not fully understood by first glance by both the writer and the reader’s ends of communication. It is apparent that percentages are not as easy to use when compared to actual figures since many people are not conversant with them.

23.A look at who is offering a statistic for consumption should be associated with bias, conscious or unconscious (pg. 123). The three aspects are critical aspects that influence the manner in which readers perceive and understand various issues.

24. If a respectable organization is cited as a source of a statistic, there are certain aspects to consider about the authority. The major to consider is whether the reputable organization stands behind the information and can substantiate it and not merely being the source. The organization ought to have credible evidence as to all the data that circulates in the media to avoid legal liability.

25. Summary of the five questions we can ask to defend ourselves as a consumer of statistics and explanation of each.

A. Who says so? Is there likelihood of bias to either the one analyzing the data or the one reporting the statistic? Is the cited authority standing behind the statistic?

B. How does he know? Is there a likelihood of bias in the sample? Is it representative?

C. What is missing? Do we have all the things required to be known to fully understand the significance of the statistic being offered?

D. Did someone change the subject? Are the definitions of all the terms fully understood, and consistent for the comparisons? Does the data has a likelihood of being accurate, or was there an opportunity and reason for the subjects to lie? Is correlation being represented as causation?

E. Does it make sense? Is it believable? Or are we being blinded by the seemingly complex analysis processes and the scientific-sounding statistic? Is it reasonable to extrapolate this far?

Reference

Huff D. (2010) How to Lie with Statistics: W.W. Norton, 2010. ISBN 0393070875,           9780393070873

Sherry Roberts is the author of this paper. A senior editor at MeldaResearch.Com in write my essay online if you need a similar paper you can place your order from write my essay for me services.