Many times we find statistical graphics in magazines and newspapers that instead of informing confound. Poorly designed statistical graphs are responsible for wrong conclusions. The analysis of this graphic was suggested by Carl J. Schwarz This graphic illustrates the cost of living in 1972 and 1985.

The height of the bars does not keep a proportional relation. compare for example food category with transportation and housing.

The percentages shown on the bars are not accuracy interpreted. The cost of transportation from 1972 to 1985 has doubled. So the increment is 100% instead of 202%

More examples of bad graphics are shown by Carl J. Schwarz

Now is your turn to interpret graphics !!!

Next graphic is the result of a classroom simulation. Students took samples of size 30 from the population. Each sample represents 30 surveyed families. For each of these samples, they recorded the number of families who own pets. The results are shown in the following diagram.

What sample outcomes are considered likely?

According to this sampling distribution , what percentage of the samples have more than 16 families with pets? Is this a likely outcome?

How many samples were used in the graphic?

How many samples reported to have 8 families with pets?

Suppose somebody run a survey with 30 families and found that only 4 families have pets. Is this a likely outcome?

Suppose you suspect that this outcome did not occur simple due to chance. Give several reasons that account for this unusually low sample outcome.

Make a pie chart with the data in the graphic. Which one is easier to understand?

Explore different graphics in spreadsheet using the data.

References

Carl J. Schwarz (1998) Statistics for the life sciences http://www.stat.sfu.ca/~cschwarz/Stat-301/Handouts/node9.html.

Coxford A. F. et al. (1997) Contemporary Mathematics in Context: a unified approach. . Course 1 Part B. Everyday Learning Corporation: Chicago Illinois

Return to my Homepage