Download - Recognizing Statistical Slips 2
RECOGNIZING STATISTICAL SLIPS
Presented by:
Hellaisna N. G
200912570094
Statistics is....• Plural
Statistics is information about any phenomenon or activity expressed in numerical form, such as vital statistics, college enrollment figures, and opinion poll percentages.
• Singular
Statistics is the art & science of collecting, presenting, analyzing, and interpreting numerical data.
Common Statistical Slips• The Deceptive Sample• The Misleading Average• Plausible Charts and Graphs• Presenting the Facts without a Reference
Point
The Deceptive Sample
Error in sampling.
Interpretation of statistic output doesn’t represent true fact of total of population.
How could this happen: Sample bias or too small Sample chosen to prove the writer’s point
Brand Manager of Head & Shoulders released the statement:
“Head & Shoulders is the best brand for shampoo category in Indonesia.”
*based on national consumer survey in 2010.
The Questions are:
Are samples chosen by pure random method?How many samples are surveyed?Is it enough to represent total of Indonesian population?
The Misleading Average
• Generating a wrong conclusion because the data didn’t distribute normal.
• Example:
“Indonesia can’t acces a soft loan from the World Bank because its per capita income is
above average.”
So, what’s the big deal??
It is because the World Bank put their
conclusion on per capita income.
But, what about the data?Does it show a true fact about Indonesia?
Income (I) Population (P) Sum of P x I
US$ 100 50 US$ 5.000
US$ 200 150 US$ 30.000
US$ 300 300 US$ 90.000
US$ 400 250 US$ 100.000
US$ 500 200 US$ 100.000
US$ 600 0 US$ 0
US$ 700 0 US$ 0
US$ 800 0 US$ 0
US$ 900 2000 US$ 18.000
US$ 1000 150 US$ 150.000
TOTAL 1.300 US$ 655.000
Indonesia’s Per Capita Income is US$ 503,85.-
That is why Indonesia can’t access any soft loan from the World Bank.note: only country with Per Capita less than US$ 500 can get a soft loan from the World Bank
In fact, Indonesian income is dominated by the middle-lower income
Plausible Charts and Graphs
• Manipulating charts or graphs to convey almost any desired impression.
What do feel when watching this chart?
Same like before Graph 2 is more impressive
Presenting the Facts Without a Reference Point
• Present a statistical result without suficient data in presentation.
The graph tells us students behavior on everage is good.The question:What is definition of those scale (0-10)
Exercise 1
“Students in Class Xo are good in mathematics. Data shows the average of their mathematics’ score is above 7.”• The Deceptive Sample• The Misleading Average• Plausible Charts and Graphs• Presenting the Facts Without a Reference point
Please try again
Sorry!! Wrong answer..
Hmmm... I don’t think that’s the
answer
Let’s check the exercise one more time!
back
That’s not the answer
The Answer is….
The Misleading
Average
Exercise 2
“The level of public health is going worse.”
• The Deceptive Sample• The Misleading Average• Plausible Charts and Graphs• Presenting the Facts
Without a Reference point
Please try again
Sorry!! Wrong answer..
Hmmm... I don’t think that’s the
answer
Let’s check the exercise one more time!
back
That’s not the answer
Presenting the Facts Without a Preference
Point
What is the answer?
Exercise 3
“94% of Indonesian women agree that the Garnier light cream can make their skin more shinning, 92% agree the Garnier white cream can reduce oil in their face, 93% agree it can reduce black spot of acne.”
• The Deceptive Sample• The Misleading Average• Plausible Charts and Graphs• Presenting the Facts Without a Reference point
Please try again
Sorry!! Wrong answer..
Hmmm... I don’t think that’s the
answer
Let’s check the exercise one more time!
back
That’s not the answer
Answer is…
The Deceptive
Sample
Conclusion
• The good statistics are:1. Sampling uses a random sampling method with
sufficient data to represent the total group.
2. There is no individual interests behind the survey.
3. No outlier in the data.
4. Don’t judge the fact by the graphs or charts.
5. Statistical output must attach a referenced point as a comparison standard.