By Nicole Goutziers

1

• By Nicole Goutziers

2
Should You Believe A Statistical Study?

• Statistical research can be very complex.
• Therefore, we must look at research carefully.
• There are 8 guidelines for determining if
  research is logical and nonbiased.

3
Guideline 1

• Identify the goal, population, and type of study
• What was the goal?
• What was the population under study? Was the
  population clearly defined?
• What type of study and is it appropriate?

4
Example Peanut Butter

• Researchers gave 50 participants Jeff Peanut
  Butter and asked them if it tasted good. 80 of
  the participants stated that the peanut butter
  tasted good. Researchers concluded that most
  people like Jeff Peanut Butter.

5
Solution

• This study was observational since researchers
  only gave the participants Jeff Peanut Butter
  with no other competition. There should have
  been a controlled experiment comparing Jeff and
  another brand of peanut butter. This study was
  inappropriate and meaningless.

6
Guideline 2

• Consider the source
• Some studies tend to be biased.
• Its important to consider the source and
  evaluate the biases that might be there.

7
Example Gum Chewing

• A study done in 2004 by the Wrigley Gum Comp.
  states that gum chewing kills bad breath germs.

8
Solution

• Since this study, other researchers have debated
  that it only covers the smell temporarily. The
  Wrigley Gum Company has a bias to try and sell
  the gum.

9
Guideline 3

• Look For Bias In The Sample
• Look for bias that may prevent the sample from
  being representative of the population.
• Happens in two forms Selection bias, and
  participation bias.

10
The Difference

• Selection- when researchers select their sample
  in an unrepresentative way.

• Participation- when people choose whether or not
  they want to participate.

11
Example Teen Beat

• Writers for Teen Beat Magazine set out a survey
  for favorite vacation spots. The participants
  have to pay the 37 cents for postage when sending
  their opinions to the magazine.

12
Solution

• Only those people who would like to pay for a
  stamp and put forth the effort to send in the
  survey would participate, therefore making it a
  participation biased survey.

13
Guideline 4

• Look for Problems in Defining or Measuring the
  Variables of Interest
• Studies usually measure something, that something
  we call variables.
• Variable- any item or quantity that can vary or
  take on different values (what the problem is
  measuring.)

14
Example Out for the Day

• The GR Press produces a survey asking
  participants whether they would rather go to the
  zoo, circus, or an amusement park.

15
Solution

• There are many zoos, circuses, and amusement
  parks around the Grand Rapids area. Different
  people may have various opinions of the John Ball
  Zoo and the Binder Park Zoo. Its too hard to
  define the examples to really see peoples
  opinions of each place.

16
Guideline 5

• Watch Out for Confounding Variables
• Useless variables are sometimes included in a
  study to make it more difficult to understand.
• These variables are called confounding
  (confusing) variables.

17
Example Fast Food?

• A study finds that more overweight people live in
  America than in China because the amount of fast
  food places is greater in America.

18
Solution

• There are other variables that can cause weight
  gain such as the time a culture eats their meals,
  to the kinds of food prepared in the average
  household. This study is not fair because they
  dont mention other important variables.

19
Guideline 6

• Consider the Setting and Wording in Surveys
• Watch out for problems in the wording that might
  produce inaccurate or dishonest responses.

20
Example A Smarter Look

• Would you say that appearance is more important
  than smarts?
• Would you say that being smart is more important
  than appearance?
• 58 answered appearance/52 answered smarts in
  1st question
• 24 answered smarts/57 answered appearance in
  2nd.

21
Solution

• The order of the words appearance and smarts
  was reversed. The change in order dramatically
  changed the results. Its a proven fact that
  people pay more attention to the first half of
  the question than the latter half.

22
Guideline 7

• Check That Results Are Presented Fairly
• Studies may be misrepresented in graphs or
  concluding statements.
• Sometimes studies tend to jump to false
  conclusions.

23
Example Laundry

• 78 of housewives report that their laundry
  doesnt look clean after washing. Therefore, in
  that 78 of households laundry is always dirty
  after washing.

24
Solution

• This statistic is generalizing. They assume that
  looking clean means the same thing as dirty. The
  women could just be talking about permanent
  stains, not actual dirt.

25
Guideline 8

• Stand Back and Consider the Conclusions
• There are specific questions you must ask
  yourself when reviewing a study.

26
4 Questions

• Can you rule out alternative explanations for the
  results?

Did the study achieve its goals?

• Do the conclusions make sense?

If it made sense, does it have significance?
27
Example Hair dye

• An experiment tracks the time it takes for
  certain hair dyes to wear out. They compare two
  products, A B. They find that product A stayed
  in longer because the color looked more vibrant
  after 30 days in most cases.

28
Solution

• This experiment has no relevance. They only used
  2 groups to experiment with. The goals werent
  achieved, the conclusion is rash, and the factors
  of how long hair color usually stays in hair
  wasnt even considered. Very unpractical.

29
Overall...
Beginning