A Summary of How to Lie with Statistics

ASummary of How to Lie with Statistics

Asummary of How to Lie with Statistics

Howto lie with statistics is a book written by Darrell Huff whichoutlines the errors that are always carried out in statistics as wellas the manner in which these errors may create conclusions that arenot correct. This summary will mention the first five chapters,outlining the key ideas and memorable examples used.

Chapterone talks about Sampling Biases. The main ideas in this chapter areresponse Bias, Non-response bias, representative sample,questionnaire wording and Recall bias. Response bias highlights thepropensity for people to exaggerate or belittle the truth whereasNon-response Bias shows the manner in which people who finish reviewsare methodicallydissimilar from the ones that fail to respond. RecallBias highlights propensity for a certain group to remember formerexperience in reflective surveys. Chapter two highlights well-chosenAverage. Main ideas are Arithmetic mean, that dispenses the totalamong persons equally, median, which represents a value divingdistribution into two even portions and Mode. Memorable example forarithmetic mean is ‘per capita income’ and for Median is Medianhousehold income. Mode shows frequently observed outcome, which isseldom conveyed with numeric data. The third chapter is named ‘Littlefigures Not There’. This chapter has six main ideas small samples,low incidence rates, significance levels, ranges, clearly label chartaxes and gathering among persons as opposed to population. Smallsamples idea shows how estimators with hefty regular mistakes canprovide apparently very strong effects. Much Ado About nothing iswhat is contained in chapter four. It highlights probable errors,Margin of error and Experimental importance. Chapter five hasEye-catching Graphs as its heading. There is percentage changethrough choice of ranges on charts that may have a great impact ondata analysis (i.e. there will be alterations in percentages).Capable of altering bar charts by having them commence at a positivevalue.

References

Darrell,H. (1954) Howto Lie with Statistics(illust. I. Geis), Norton, New York.