math 124: lecture for week 4 of 17online.sfsu.edu/meredith/statistics/s2008/lecture_4_slides.pdfweek...

87
Math 124: Week 3 Questions and Answers Quiz Review Standard Deviation Review Case I Review Case II Review Case III Sampling Designing Studies Observational Studies Experimental Studies Surveys Math 124: Lecture for Week 4 of 17 David Meredith Department of Mathematics San Francisco State University February 19, 2008

Upload: others

Post on 21-Sep-2020

0 views

Category:

Documents


0 download

TRANSCRIPT

Page 1: Math 124: Lecture for Week 4 of 17online.sfsu.edu/meredith/statistics/S2008/Lecture_4_slides.pdfWeek 3 Questions and Answers Quiz Review Standard Deviation Review Case I Review Case

Math 124:Week 3

Questions andAnswers

Quiz

ReviewStandard Deviation

Review Case I

Review Case II

Review Case III

Sampling

DesigningStudiesObservationalStudies

Experimental Studies

Surveys

Math 124: Lecture for Week 4 of 17

David MeredithDepartment of Mathematics

San Francisco State University

February 19, 2008

Page 2: Math 124: Lecture for Week 4 of 17online.sfsu.edu/meredith/statistics/S2008/Lecture_4_slides.pdfWeek 3 Questions and Answers Quiz Review Standard Deviation Review Case I Review Case

Math 124:Week 3

Questions andAnswers

Quiz

ReviewStandard Deviation

Review Case I

Review Case II

Review Case III

Sampling

DesigningStudiesObservationalStudies

Experimental Studies

Surveys

Class progress

• As of 5:30PM• only 10 people had completed the Lab Exercise• only 4 people had read the complete assignment

• Passing this class is up to you.• If you do the work, you will get a good grade• An hour a day, every day, is what you need to do

• Do two 30-minute sessions if that works better

• First on-line quiz due Thursday, second on-line quizdue one week from Thursday.

• First midterm in two weeks

Page 3: Math 124: Lecture for Week 4 of 17online.sfsu.edu/meredith/statistics/S2008/Lecture_4_slides.pdfWeek 3 Questions and Answers Quiz Review Standard Deviation Review Case I Review Case

Math 124:Week 3

Questions andAnswers

Quiz

ReviewStandard Deviation

Review Case I

Review Case II

Review Case III

Sampling

DesigningStudiesObservationalStudies

Experimental Studies

Surveys

What we will do tonight

1 Questions and Answers

2 Quiz

3 ReviewStandard DeviationReview Case IReview Case IIReview Case III

4 Sampling

5 Designing StudiesObservational StudiesExperimental StudiesSurveys

Page 4: Math 124: Lecture for Week 4 of 17online.sfsu.edu/meredith/statistics/S2008/Lecture_4_slides.pdfWeek 3 Questions and Answers Quiz Review Standard Deviation Review Case I Review Case

Math 124:Week 3

Questions andAnswers

Quiz

ReviewStandard Deviation

Review Case I

Review Case II

Review Case III

Sampling

DesigningStudiesObservationalStudies

Experimental Studies

Surveys

What we will do tonight

1 Questions and Answers

2 Quiz

3 ReviewStandard DeviationReview Case IReview Case IIReview Case III

4 Sampling

5 Designing StudiesObservational StudiesExperimental StudiesSurveys

Page 5: Math 124: Lecture for Week 4 of 17online.sfsu.edu/meredith/statistics/S2008/Lecture_4_slides.pdfWeek 3 Questions and Answers Quiz Review Standard Deviation Review Case I Review Case

Math 124:Week 3

Questions andAnswers

Quiz

ReviewStandard Deviation

Review Case I

Review Case II

Review Case III

Sampling

DesigningStudiesObservationalStudies

Experimental Studies

Surveys

What we will do tonight

1 Questions and Answers

2 Quiz

3 ReviewStandard DeviationReview Case IReview Case IIReview Case III

4 Sampling

5 Designing StudiesObservational StudiesExperimental StudiesSurveys

Page 6: Math 124: Lecture for Week 4 of 17online.sfsu.edu/meredith/statistics/S2008/Lecture_4_slides.pdfWeek 3 Questions and Answers Quiz Review Standard Deviation Review Case I Review Case

Math 124:Week 3

Questions andAnswers

Quiz

ReviewStandard Deviation

Review Case I

Review Case II

Review Case III

Sampling

DesigningStudiesObservationalStudies

Experimental Studies

Surveys

What we will do tonight

1 Questions and Answers

2 Quiz

3 ReviewStandard DeviationReview Case IReview Case IIReview Case III

4 Sampling

5 Designing StudiesObservational StudiesExperimental StudiesSurveys

Page 7: Math 124: Lecture for Week 4 of 17online.sfsu.edu/meredith/statistics/S2008/Lecture_4_slides.pdfWeek 3 Questions and Answers Quiz Review Standard Deviation Review Case I Review Case

Math 124:Week 3

Questions andAnswers

Quiz

ReviewStandard Deviation

Review Case I

Review Case II

Review Case III

Sampling

DesigningStudiesObservationalStudies

Experimental Studies

Surveys

What we will do tonight

1 Questions and Answers

2 Quiz

3 ReviewStandard DeviationReview Case IReview Case IIReview Case III

4 Sampling

5 Designing StudiesObservational StudiesExperimental StudiesSurveys

Page 8: Math 124: Lecture for Week 4 of 17online.sfsu.edu/meredith/statistics/S2008/Lecture_4_slides.pdfWeek 3 Questions and Answers Quiz Review Standard Deviation Review Case I Review Case

Math 124:Week 3

Questions andAnswers

Quiz

ReviewStandard Deviation

Review Case I

Review Case II

Review Case III

Sampling

DesigningStudiesObservationalStudies

Experimental Studies

Surveys

Questions and Answers

Page 9: Math 124: Lecture for Week 4 of 17online.sfsu.edu/meredith/statistics/S2008/Lecture_4_slides.pdfWeek 3 Questions and Answers Quiz Review Standard Deviation Review Case I Review Case

Math 124:Week 3

Questions andAnswers

Quiz

ReviewStandard Deviation

Review Case I

Review Case II

Review Case III

Sampling

DesigningStudiesObservationalStudies

Experimental Studies

Surveys

Go over quiz

www.cmu.edu/oli

Page 10: Math 124: Lecture for Week 4 of 17online.sfsu.edu/meredith/statistics/S2008/Lecture_4_slides.pdfWeek 3 Questions and Answers Quiz Review Standard Deviation Review Case I Review Case

Math 124:Week 3

Questions andAnswers

Quiz

ReviewStandard Deviation

Review Case I

Review Case II

Review Case III

Sampling

DesigningStudiesObservationalStudies

Experimental Studies

Surveys

Apply the standard deviation rule

Question 1:

• Calculate the mean and standard deviation of thedataset:

5.1 6.4 6.5 8.8 8.8 9.4 9.9 10.110.2 13.1 13.5 13.5 13.6 14.4 14.6

• Is this data approximately normally distributed?• Implicit understanding for this class• Whenever you get a question like the one above, the

answer isn’t just “yes” or “no”; the answer is “yes,because . . . ” or “no, because . . . ”.

Page 11: Math 124: Lecture for Week 4 of 17online.sfsu.edu/meredith/statistics/S2008/Lecture_4_slides.pdfWeek 3 Questions and Answers Quiz Review Standard Deviation Review Case I Review Case

Math 124:Week 3

Questions andAnswers

Quiz

ReviewStandard Deviation

Review Case I

Review Case II

Review Case III

Sampling

DesigningStudiesObservationalStudies

Experimental Studies

Surveys

Apply the standard deviation rule

Question 1:

• Calculate the mean and standard deviation of thedataset:

5.1 6.4 6.5 8.8 8.8 9.4 9.9 10.110.2 13.1 13.5 13.5 13.6 14.4 14.6

• Is this data approximately normally distributed?

• Implicit understanding for this class• Whenever you get a question like the one above, the

answer isn’t just “yes” or “no”; the answer is “yes,because . . . ” or “no, because . . . ”.

Page 12: Math 124: Lecture for Week 4 of 17online.sfsu.edu/meredith/statistics/S2008/Lecture_4_slides.pdfWeek 3 Questions and Answers Quiz Review Standard Deviation Review Case I Review Case

Math 124:Week 3

Questions andAnswers

Quiz

ReviewStandard Deviation

Review Case I

Review Case II

Review Case III

Sampling

DesigningStudiesObservationalStudies

Experimental Studies

Surveys

Apply the standard deviation rule

Question 1:

• Calculate the mean and standard deviation of thedataset:

5.1 6.4 6.5 8.8 8.8 9.4 9.9 10.110.2 13.1 13.5 13.5 13.6 14.4 14.6

• Is this data approximately normally distributed?• Implicit understanding for this class

• Whenever you get a question like the one above, theanswer isn’t just “yes” or “no”; the answer is “yes,because . . . ” or “no, because . . . ”.

Page 13: Math 124: Lecture for Week 4 of 17online.sfsu.edu/meredith/statistics/S2008/Lecture_4_slides.pdfWeek 3 Questions and Answers Quiz Review Standard Deviation Review Case I Review Case

Math 124:Week 3

Questions andAnswers

Quiz

ReviewStandard Deviation

Review Case I

Review Case II

Review Case III

Sampling

DesigningStudiesObservationalStudies

Experimental Studies

Surveys

Apply the standard deviation rule

Question 1:

• Calculate the mean and standard deviation of thedataset:

5.1 6.4 6.5 8.8 8.8 9.4 9.9 10.110.2 13.1 13.5 13.5 13.6 14.4 14.6

• Is this data approximately normally distributed?• Implicit understanding for this class• Whenever you get a question like the one above, the

answer isn’t just “yes” or “no”; the answer is “yes,because . . . ” or “no, because . . . ”.

Page 14: Math 124: Lecture for Week 4 of 17online.sfsu.edu/meredith/statistics/S2008/Lecture_4_slides.pdfWeek 3 Questions and Answers Quiz Review Standard Deviation Review Case I Review Case

Math 124:Week 3

Questions andAnswers

Quiz

ReviewStandard Deviation

Review Case I

Review Case II

Review Case III

Sampling

DesigningStudiesObservationalStudies

Experimental Studies

Surveys

Compare and contrast distributions of quantitative data from

two or more groups

Question 2: here are scores from a test I gave to Math 124last semester.

100 100 92 92 83 83 83 83 75Women 75 67 67 67 58 58 58 50 50

50 50 50 42 33 33 17100 100 100 92 92 92 92 83 83

Men 75 67 67 58 58 58 58 50 4242 33

1 How would you analyze and compare the results formen and women??

Page 15: Math 124: Lecture for Week 4 of 17online.sfsu.edu/meredith/statistics/S2008/Lecture_4_slides.pdfWeek 3 Questions and Answers Quiz Review Standard Deviation Review Case I Review Case

Math 124:Week 3

Questions andAnswers

Quiz

ReviewStandard Deviation

Review Case I

Review Case II

Review Case III

Sampling

DesigningStudiesObservationalStudies

Experimental Studies

Surveys

Compare and contrast distributions of quantitative data from

two or more groups

Question 2: here are scores from a test I gave to Math 124last semester.

100 100 92 92 83 83 83 83 75Women 75 67 67 67 58 58 58 50 50

50 50 50 42 33 33 17100 100 100 92 92 92 92 83 83

Men 75 67 67 58 58 58 58 50 4242 33

1 How would you analyze and compare the results formen and women??

Page 16: Math 124: Lecture for Week 4 of 17online.sfsu.edu/meredith/statistics/S2008/Lecture_4_slides.pdfWeek 3 Questions and Answers Quiz Review Standard Deviation Review Case I Review Case

Math 124:Week 3

Questions andAnswers

Quiz

ReviewStandard Deviation

Review Case I

Review Case II

Review Case III

Sampling

DesigningStudiesObservationalStudies

Experimental Studies

Surveys

Compare and contrast distributions of quantitative data from

two or more groups

Question 2 continued: here are five-number summaries forthe scores from a test I gave to Math 124 last semester.

Min. 1st Qu. Median 3rd Qu. Max.Women 17.00 50.00 67.00 83.00 100.00Men 33.0 58.0 71.0 92.0 100.0

1 Draw the boxplots.2 Would you say that the difference between men and

women was significant?

Page 17: Math 124: Lecture for Week 4 of 17online.sfsu.edu/meredith/statistics/S2008/Lecture_4_slides.pdfWeek 3 Questions and Answers Quiz Review Standard Deviation Review Case I Review Case

Math 124:Week 3

Questions andAnswers

Quiz

ReviewStandard Deviation

Review Case I

Review Case II

Review Case III

Sampling

DesigningStudiesObservationalStudies

Experimental Studies

Surveys

Compare and contrast distributions of quantitative data from

two or more groups

Question 2 continued: here are five-number summaries forthe scores from a test I gave to Math 124 last semester.

Min. 1st Qu. Median 3rd Qu. Max.Women 17.00 50.00 67.00 83.00 100.00Men 33.0 58.0 71.0 92.0 100.0

1 Draw the boxplots.

2 Would you say that the difference between men andwomen was significant?

Page 18: Math 124: Lecture for Week 4 of 17online.sfsu.edu/meredith/statistics/S2008/Lecture_4_slides.pdfWeek 3 Questions and Answers Quiz Review Standard Deviation Review Case I Review Case

Math 124:Week 3

Questions andAnswers

Quiz

ReviewStandard Deviation

Review Case I

Review Case II

Review Case III

Sampling

DesigningStudiesObservationalStudies

Experimental Studies

Surveys

Compare and contrast distributions of quantitative data from

two or more groups

Question 2 continued: here are five-number summaries forthe scores from a test I gave to Math 124 last semester.

Min. 1st Qu. Median 3rd Qu. Max.Women 17.00 50.00 67.00 83.00 100.00Men 33.0 58.0 71.0 92.0 100.0

1 Draw the boxplots.2 Would you say that the difference between men and

women was significant?

Page 19: Math 124: Lecture for Week 4 of 17online.sfsu.edu/meredith/statistics/S2008/Lecture_4_slides.pdfWeek 3 Questions and Answers Quiz Review Standard Deviation Review Case I Review Case

Math 124:Week 3

Questions andAnswers

Quiz

ReviewStandard Deviation

Review Case I

Review Case II

Review Case III

Sampling

DesigningStudiesObservationalStudies

Experimental Studies

Surveys

Produce a two-way table and compare conditional

percentages

Question 3: Here’s some imaginary data.

Explanatory A B A A A B B A B AResponse X Z Z Z X Y X X X X

Explanatory A B B A B A B B A AResponse X X Z Y Z X Y Z X X

• How would you analyze and summarize these datanumerically and graphically?

• Do it.

Page 20: Math 124: Lecture for Week 4 of 17online.sfsu.edu/meredith/statistics/S2008/Lecture_4_slides.pdfWeek 3 Questions and Answers Quiz Review Standard Deviation Review Case I Review Case

Math 124:Week 3

Questions andAnswers

Quiz

ReviewStandard Deviation

Review Case I

Review Case II

Review Case III

Sampling

DesigningStudiesObservationalStudies

Experimental Studies

Surveys

Produce a two-way table and compare conditional

percentages

Question 3: Here’s some imaginary data.Explanatory A B A A A B B A B A

Response X Z Z Z X Y X X X X

Explanatory A B B A B A B B A AResponse X X Z Y Z X Y Z X X

• How would you analyze and summarize these datanumerically and graphically?

• Do it.

Page 21: Math 124: Lecture for Week 4 of 17online.sfsu.edu/meredith/statistics/S2008/Lecture_4_slides.pdfWeek 3 Questions and Answers Quiz Review Standard Deviation Review Case I Review Case

Math 124:Week 3

Questions andAnswers

Quiz

ReviewStandard Deviation

Review Case I

Review Case II

Review Case III

Sampling

DesigningStudiesObservationalStudies

Experimental Studies

Surveys

Produce a two-way table and compare conditional

percentages

Question 3: Here’s some imaginary data.Explanatory A B A A A B B A B A

Response X Z Z Z X Y X X X X

Explanatory A B B A B A B B A AResponse X X Z Y Z X Y Z X X

• How would you analyze and summarize these datanumerically and graphically?

• Do it.

Page 22: Math 124: Lecture for Week 4 of 17online.sfsu.edu/meredith/statistics/S2008/Lecture_4_slides.pdfWeek 3 Questions and Answers Quiz Review Standard Deviation Review Case I Review Case

Math 124:Week 3

Questions andAnswers

Quiz

ReviewStandard Deviation

Review Case I

Review Case II

Review Case III

Sampling

DesigningStudiesObservationalStudies

Experimental Studies

Surveys

Produce a two-way table and compare conditional

percentages

Question 3: Here’s some imaginary data.Explanatory A B A A A B B A B A

Response X Z Z Z X Y X X X X

Explanatory A B B A B A B B A AResponse X X Z Y Z X Y Z X X

• How would you analyze and summarize these datanumerically and graphically?

• Do it.

Page 23: Math 124: Lecture for Week 4 of 17online.sfsu.edu/meredith/statistics/S2008/Lecture_4_slides.pdfWeek 3 Questions and Answers Quiz Review Standard Deviation Review Case I Review Case

Math 124:Week 3

Questions andAnswers

Quiz

ReviewStandard Deviation

Review Case I

Review Case II

Review Case III

Sampling

DesigningStudiesObservationalStudies

Experimental Studies

Surveys

Interpret the value of the correlation coefficient

• This question is taken from a sample CSET test. CSETis a test that qualifies potential high school mathteachers as competent in mathematics.

• Question 4: Suppose the equation for the regressionline through some data is Y = 4.5 + 2.3X . Then thecorrelation coefficient for the data could be:

1 2.42 0.63 -0.74 -1.3

• If the regression equation was Y = 4.5− 2.3X , whatcould the correlation coefficient be?

• Explain your reasons.

Page 24: Math 124: Lecture for Week 4 of 17online.sfsu.edu/meredith/statistics/S2008/Lecture_4_slides.pdfWeek 3 Questions and Answers Quiz Review Standard Deviation Review Case I Review Case

Math 124:Week 3

Questions andAnswers

Quiz

ReviewStandard Deviation

Review Case I

Review Case II

Review Case III

Sampling

DesigningStudiesObservationalStudies

Experimental Studies

Surveys

Interpret the value of the correlation coefficient

• This question is taken from a sample CSET test. CSETis a test that qualifies potential high school mathteachers as competent in mathematics.

• Question 4: Suppose the equation for the regressionline through some data is Y = 4.5 + 2.3X . Then thecorrelation coefficient for the data could be:

1 2.42 0.63 -0.74 -1.3

• If the regression equation was Y = 4.5− 2.3X , whatcould the correlation coefficient be?

• Explain your reasons.

Page 25: Math 124: Lecture for Week 4 of 17online.sfsu.edu/meredith/statistics/S2008/Lecture_4_slides.pdfWeek 3 Questions and Answers Quiz Review Standard Deviation Review Case I Review Case

Math 124:Week 3

Questions andAnswers

Quiz

ReviewStandard Deviation

Review Case I

Review Case II

Review Case III

Sampling

DesigningStudiesObservationalStudies

Experimental Studies

Surveys

Interpret the value of the correlation coefficient

Question 5: Here’s the data from the text relating thelongevity of animals to their gestation period.

Page 26: Math 124: Lecture for Week 4 of 17online.sfsu.edu/meredith/statistics/S2008/Lecture_4_slides.pdfWeek 3 Questions and Answers Quiz Review Standard Deviation Review Case I Review Case

Math 124:Week 3

Questions andAnswers

Quiz

ReviewStandard Deviation

Review Case I

Review Case II

Review Case III

Sampling

DesigningStudiesObservationalStudies

Experimental Studies

Surveys

Interpret the value of the correlation coefficient

• Is the correlation coefficient for these data positive ornegative?

• Does including the outlier increase or decrease thecorrelation coefficient?

• The correlation with the outlier is 0.66; without theoutlier the correlation is 0.52.

Page 27: Math 124: Lecture for Week 4 of 17online.sfsu.edu/meredith/statistics/S2008/Lecture_4_slides.pdfWeek 3 Questions and Answers Quiz Review Standard Deviation Review Case I Review Case

Math 124:Week 3

Questions andAnswers

Quiz

ReviewStandard Deviation

Review Case I

Review Case II

Review Case III

Sampling

DesigningStudiesObservationalStudies

Experimental Studies

Surveys

Interpret the value of the correlation coefficient

• Is the correlation coefficient for these data positive ornegative?

• Does including the outlier increase or decrease thecorrelation coefficient?

• The correlation with the outlier is 0.66; without theoutlier the correlation is 0.52.

Page 28: Math 124: Lecture for Week 4 of 17online.sfsu.edu/meredith/statistics/S2008/Lecture_4_slides.pdfWeek 3 Questions and Answers Quiz Review Standard Deviation Review Case I Review Case

Math 124:Week 3

Questions andAnswers

Quiz

ReviewStandard Deviation

Review Case I

Review Case II

Review Case III

Sampling

DesigningStudiesObservationalStudies

Experimental Studies

Surveys

In the special case of linear relationship, use the least

squares regression line

Question 6: Here’s a subset of the data from the textrelating the longevity of animals to their gestation period.

Page 29: Math 124: Lecture for Week 4 of 17online.sfsu.edu/meredith/statistics/S2008/Lecture_4_slides.pdfWeek 3 Questions and Answers Quiz Review Standard Deviation Review Case I Review Case

Math 124:Week 3

Questions andAnswers

Quiz

ReviewStandard Deviation

Review Case I

Review Case II

Review Case III

Sampling

DesigningStudiesObservationalStudies

Experimental Studies

Surveys

In the special case of linear relationship, use the least

squares regression line

• Is the correlation coefficient for these data positive ornegative?

• Is the correlation coefficient:

1 -0.632 -0.283 0.414 0.755 0.96

• The correlation coefficient is 0.75

Page 30: Math 124: Lecture for Week 4 of 17online.sfsu.edu/meredith/statistics/S2008/Lecture_4_slides.pdfWeek 3 Questions and Answers Quiz Review Standard Deviation Review Case I Review Case

Math 124:Week 3

Questions andAnswers

Quiz

ReviewStandard Deviation

Review Case I

Review Case II

Review Case III

Sampling

DesigningStudiesObservationalStudies

Experimental Studies

Surveys

In the special case of linear relationship, use the least

squares regression line

• Is the correlation coefficient for these data positive ornegative?

• Is the correlation coefficient:

1 -0.632 -0.283 0.414 0.755 0.96

• The correlation coefficient is 0.75

Page 31: Math 124: Lecture for Week 4 of 17online.sfsu.edu/meredith/statistics/S2008/Lecture_4_slides.pdfWeek 3 Questions and Answers Quiz Review Standard Deviation Review Case I Review Case

Math 124:Week 3

Questions andAnswers

Quiz

ReviewStandard Deviation

Review Case I

Review Case II

Review Case III

Sampling

DesigningStudiesObservationalStudies

Experimental Studies

Surveys

In the special case of linear relationship, use the least

squares regression line

The regression equation for these data is Y = 12.9 + 10.3X

Page 32: Math 124: Lecture for Week 4 of 17online.sfsu.edu/meredith/statistics/S2008/Lecture_4_slides.pdfWeek 3 Questions and Answers Quiz Review Standard Deviation Review Case I Review Case

Math 124:Week 3

Questions andAnswers

Quiz

ReviewStandard Deviation

Review Case I

Review Case II

Review Case III

Sampling

DesigningStudiesObservationalStudies

Experimental Studies

Surveys

In the special case of linear relationship, use the least

squares regression line

Question 7:

1 Can you estimate the gestation period for an animalthat lives for 15 years?

2 Can you estimate the gestation period for an animalthat lives for 40 years?

3 Can you estimate the life span of an animal with agestation period of 150 days?

167 days; no, that would require extrapolation; 13.3 years

Page 33: Math 124: Lecture for Week 4 of 17online.sfsu.edu/meredith/statistics/S2008/Lecture_4_slides.pdfWeek 3 Questions and Answers Quiz Review Standard Deviation Review Case I Review Case

Math 124:Week 3

Questions andAnswers

Quiz

ReviewStandard Deviation

Review Case I

Review Case II

Review Case III

Sampling

DesigningStudiesObservationalStudies

Experimental Studies

Surveys

In the special case of linear relationship, use the least

squares regression line

Question 7:

1 Can you estimate the gestation period for an animalthat lives for 15 years?

2 Can you estimate the gestation period for an animalthat lives for 40 years?

3 Can you estimate the life span of an animal with agestation period of 150 days?

167 days; no, that would require extrapolation; 13.3 years

Page 34: Math 124: Lecture for Week 4 of 17online.sfsu.edu/meredith/statistics/S2008/Lecture_4_slides.pdfWeek 3 Questions and Answers Quiz Review Standard Deviation Review Case I Review Case

Math 124:Week 3

Questions andAnswers

Quiz

ReviewStandard Deviation

Review Case I

Review Case II

Review Case III

Sampling

DesigningStudiesObservationalStudies

Experimental Studies

Surveys

Recognize and explain the phenomenon of Simpson’s

Paradox

• Simpson’s paradox occurs when data indicates oneconclusion, but when divided into two parts each partindicates the opposite conclusion, or when divided intomany parts most parts indicate the opposite conclusion

• Obviously this cannot happen, but it does• example from UC Bekeley (Wikipedia–Simpson’s

Paradox)http://en.wikipedia.org/wiki/Simpson’s_paradox

Page 35: Math 124: Lecture for Week 4 of 17online.sfsu.edu/meredith/statistics/S2008/Lecture_4_slides.pdfWeek 3 Questions and Answers Quiz Review Standard Deviation Review Case I Review Case

Math 124:Week 3

Questions andAnswers

Quiz

ReviewStandard Deviation

Review Case I

Review Case II

Review Case III

Sampling

DesigningStudiesObservationalStudies

Experimental Studies

Surveys

Break

Let’s take a fifteen minute break.

Page 36: Math 124: Lecture for Week 4 of 17online.sfsu.edu/meredith/statistics/S2008/Lecture_4_slides.pdfWeek 3 Questions and Answers Quiz Review Standard Deviation Review Case I Review Case

Math 124:Week 3

Questions andAnswers

Quiz

ReviewStandard Deviation

Review Case I

Review Case II

Review Case III

Sampling

DesigningStudiesObservationalStudies

Experimental Studies

Surveys

Identify the sampling method used in a study

• A statistical study usually begins with a sample of somesort

• Sample of people from the population you want to study• Sample of rocks from the area you want to study• Sample of blood from the person you want to study

Page 37: Math 124: Lecture for Week 4 of 17online.sfsu.edu/meredith/statistics/S2008/Lecture_4_slides.pdfWeek 3 Questions and Answers Quiz Review Standard Deviation Review Case I Review Case

Math 124:Week 3

Questions andAnswers

Quiz

ReviewStandard Deviation

Review Case I

Review Case II

Review Case III

Sampling

DesigningStudiesObservationalStudies

Experimental Studies

Surveys

Identify the sampling method used in a study

• Good types of samples• a simple random sample (selected at random and

without replacement) is, as the name suggests, thesimplest probability sampling plan.

• in cluster sampling we randomly choose whole orentire groups (clusters) of individuals belongingtogether in one way or another, such as taking a simplerandom sample of households and studying allindividuals in each sampled household.

Page 38: Math 124: Lecture for Week 4 of 17online.sfsu.edu/meredith/statistics/S2008/Lecture_4_slides.pdfWeek 3 Questions and Answers Quiz Review Standard Deviation Review Case I Review Case

Math 124:Week 3

Questions andAnswers

Quiz

ReviewStandard Deviation

Review Case I

Review Case II

Review Case III

Sampling

DesigningStudiesObservationalStudies

Experimental Studies

Surveys

Identify the sampling method used in a study

In stratified sampling we first divide the population intogroups of individuals that are similar to each other in animportant way with respect to the variable of interest (forexample, by gender, or year in college), and then choose asimple random sample from within each group. The groupsare called strata and hence the name "stratified sampling".Using this (a bit more cumbersome) method may help toobtain a representative sample more efficiently.

Page 39: Math 124: Lecture for Week 4 of 17online.sfsu.edu/meredith/statistics/S2008/Lecture_4_slides.pdfWeek 3 Questions and Answers Quiz Review Standard Deviation Review Case I Review Case

Math 124:Week 3

Questions andAnswers

Quiz

ReviewStandard Deviation

Review Case I

Review Case II

Review Case III

Sampling

DesigningStudiesObservationalStudies

Experimental Studies

Surveys

Identify the sampling method used in a study

Question 8: What kind of study might cluster sampling

Page 40: Math 124: Lecture for Week 4 of 17online.sfsu.edu/meredith/statistics/S2008/Lecture_4_slides.pdfWeek 3 Questions and Answers Quiz Review Standard Deviation Review Case I Review Case

Math 124:Week 3

Questions andAnswers

Quiz

ReviewStandard Deviation

Review Case I

Review Case II

Review Case III

Sampling

DesigningStudiesObservationalStudies

Experimental Studies

Surveys

Identify the sampling method used in a study

Whatever kind of sampling you use, a larger sample isusually more accurate than a smaller sample.• Later in this course we study how much improvement a

larger sample can give you• There’s always a trade-off between cost and accuracy.

Page 41: Math 124: Lecture for Week 4 of 17online.sfsu.edu/meredith/statistics/S2008/Lecture_4_slides.pdfWeek 3 Questions and Answers Quiz Review Standard Deviation Review Case I Review Case

Math 124:Week 3

Questions andAnswers

Quiz

ReviewStandard Deviation

Review Case I

Review Case II

Review Case III

Sampling

DesigningStudiesObservationalStudies

Experimental Studies

Surveys

Identify the sampling method used in a study

• Bad sorts of samples• volunteer sample• convenience sample• wrong sampling frame• systematic sampling

Page 42: Math 124: Lecture for Week 4 of 17online.sfsu.edu/meredith/statistics/S2008/Lecture_4_slides.pdfWeek 3 Questions and Answers Quiz Review Standard Deviation Review Case I Review Case

Math 124:Week 3

Questions andAnswers

Quiz

ReviewStandard Deviation

Review Case I

Review Case II

Review Case III

Sampling

DesigningStudiesObservationalStudies

Experimental Studies

Surveys

Identify the sampling method used in a study

• [X ]press’ own unscientific, on-campus exit pollsurveyed 409 people who voted on campus. Sixty-ninepercent of the people polled said they voted for BarackObama while only 20 percent voted for Hillary Clinton.

• Question 10• What kind of sample was this?• How many people voted Republican?

Page 43: Math 124: Lecture for Week 4 of 17online.sfsu.edu/meredith/statistics/S2008/Lecture_4_slides.pdfWeek 3 Questions and Answers Quiz Review Standard Deviation Review Case I Review Case

Math 124:Week 3

Questions andAnswers

Quiz

ReviewStandard Deviation

Review Case I

Review Case II

Review Case III

Sampling

DesigningStudiesObservationalStudies

Experimental Studies

Surveys

Identify the design of a study (controlled experiment vs.

observational study)

• There are three kinds of statistical studies

• Observational• Experimental• Survey

Page 44: Math 124: Lecture for Week 4 of 17online.sfsu.edu/meredith/statistics/S2008/Lecture_4_slides.pdfWeek 3 Questions and Answers Quiz Review Standard Deviation Review Case I Review Case

Math 124:Week 3

Questions andAnswers

Quiz

ReviewStandard Deviation

Review Case I

Review Case II

Review Case III

Sampling

DesigningStudiesObservationalStudies

Experimental Studies

Surveys

Identify the design of a study (controlled experiment vs.

observational study)

• There are three kinds of statistical studies• Observational

• Experimental• Survey

Page 45: Math 124: Lecture for Week 4 of 17online.sfsu.edu/meredith/statistics/S2008/Lecture_4_slides.pdfWeek 3 Questions and Answers Quiz Review Standard Deviation Review Case I Review Case

Math 124:Week 3

Questions andAnswers

Quiz

ReviewStandard Deviation

Review Case I

Review Case II

Review Case III

Sampling

DesigningStudiesObservationalStudies

Experimental Studies

Surveys

Identify the design of a study (controlled experiment vs.

observational study)

• There are three kinds of statistical studies• Observational• Experimental

• Survey

Page 46: Math 124: Lecture for Week 4 of 17online.sfsu.edu/meredith/statistics/S2008/Lecture_4_slides.pdfWeek 3 Questions and Answers Quiz Review Standard Deviation Review Case I Review Case

Math 124:Week 3

Questions andAnswers

Quiz

ReviewStandard Deviation

Review Case I

Review Case II

Review Case III

Sampling

DesigningStudiesObservationalStudies

Experimental Studies

Surveys

Identify the design of a study (controlled experiment vs.

observational study)

• There are three kinds of statistical studies• Observational• Experimental• Survey

Page 47: Math 124: Lecture for Week 4 of 17online.sfsu.edu/meredith/statistics/S2008/Lecture_4_slides.pdfWeek 3 Questions and Answers Quiz Review Standard Deviation Review Case I Review Case

Math 124:Week 3

Questions andAnswers

Quiz

ReviewStandard Deviation

Review Case I

Review Case II

Review Case III

Sampling

DesigningStudiesObservationalStudies

Experimental Studies

Surveys

Identify the design and other features of a study

• Observational studies observe variables as theynaturally occur

• Advantage: ecological validity—subjects behavenaturally

• May be negated by subjects’ knowledge that they arebeing observed (Hawthorne effect)

• Hawthorne works of the Western Electric Company inChicago 1924-32

• Disadvantage: difficult to control for lurking variables• For example, if quit-smoking programs are studied by

allowing people to choose one or another quit-smokingmethod spontaneously, then those more determined toquit may tend to choose a different method than thosewho aren’t really trying.

• This will make the method selected by the quitters lookgood

• You overcome this difficulty by conducting an experimentin which people who want to quit smoking are randomlyassigned to different treatments.

Page 48: Math 124: Lecture for Week 4 of 17online.sfsu.edu/meredith/statistics/S2008/Lecture_4_slides.pdfWeek 3 Questions and Answers Quiz Review Standard Deviation Review Case I Review Case

Math 124:Week 3

Questions andAnswers

Quiz

ReviewStandard Deviation

Review Case I

Review Case II

Review Case III

Sampling

DesigningStudiesObservationalStudies

Experimental Studies

Surveys

Identify the design and other features of a study

• Observational studies observe variables as theynaturally occur

• Advantage: ecological validity—subjects behavenaturally

• May be negated by subjects’ knowledge that they arebeing observed (Hawthorne effect)

• Hawthorne works of the Western Electric Company inChicago 1924-32

• Disadvantage: difficult to control for lurking variables• For example, if quit-smoking programs are studied by

allowing people to choose one or another quit-smokingmethod spontaneously, then those more determined toquit may tend to choose a different method than thosewho aren’t really trying.

• This will make the method selected by the quitters lookgood

• You overcome this difficulty by conducting an experimentin which people who want to quit smoking are randomlyassigned to different treatments.

Page 49: Math 124: Lecture for Week 4 of 17online.sfsu.edu/meredith/statistics/S2008/Lecture_4_slides.pdfWeek 3 Questions and Answers Quiz Review Standard Deviation Review Case I Review Case

Math 124:Week 3

Questions andAnswers

Quiz

ReviewStandard Deviation

Review Case I

Review Case II

Review Case III

Sampling

DesigningStudiesObservationalStudies

Experimental Studies

Surveys

Identify the design and other features of a study

• Observational studies observe variables as theynaturally occur

• Advantage: ecological validity—subjects behavenaturally

• May be negated by subjects’ knowledge that they arebeing observed (Hawthorne effect)

• Hawthorne works of the Western Electric Company inChicago 1924-32

• Disadvantage: difficult to control for lurking variables• For example, if quit-smoking programs are studied by

allowing people to choose one or another quit-smokingmethod spontaneously, then those more determined toquit may tend to choose a different method than thosewho aren’t really trying.

• This will make the method selected by the quitters lookgood

• You overcome this difficulty by conducting an experimentin which people who want to quit smoking are randomlyassigned to different treatments.

Page 50: Math 124: Lecture for Week 4 of 17online.sfsu.edu/meredith/statistics/S2008/Lecture_4_slides.pdfWeek 3 Questions and Answers Quiz Review Standard Deviation Review Case I Review Case

Math 124:Week 3

Questions andAnswers

Quiz

ReviewStandard Deviation

Review Case I

Review Case II

Review Case III

Sampling

DesigningStudiesObservationalStudies

Experimental Studies

Surveys

Identify the design and other features of a study

• Observational studies observe variables as theynaturally occur

• Advantage: ecological validity—subjects behavenaturally

• May be negated by subjects’ knowledge that they arebeing observed (Hawthorne effect)

• Hawthorne works of the Western Electric Company inChicago 1924-32

• Disadvantage: difficult to control for lurking variables

• For example, if quit-smoking programs are studied byallowing people to choose one or another quit-smokingmethod spontaneously, then those more determined toquit may tend to choose a different method than thosewho aren’t really trying.

• This will make the method selected by the quitters lookgood

• You overcome this difficulty by conducting an experimentin which people who want to quit smoking are randomlyassigned to different treatments.

Page 51: Math 124: Lecture for Week 4 of 17online.sfsu.edu/meredith/statistics/S2008/Lecture_4_slides.pdfWeek 3 Questions and Answers Quiz Review Standard Deviation Review Case I Review Case

Math 124:Week 3

Questions andAnswers

Quiz

ReviewStandard Deviation

Review Case I

Review Case II

Review Case III

Sampling

DesigningStudiesObservationalStudies

Experimental Studies

Surveys

Identify the design and other features of a study

• Observational studies observe variables as theynaturally occur

• Advantage: ecological validity—subjects behavenaturally

• May be negated by subjects’ knowledge that they arebeing observed (Hawthorne effect)

• Hawthorne works of the Western Electric Company inChicago 1924-32

• Disadvantage: difficult to control for lurking variables• For example, if quit-smoking programs are studied by

allowing people to choose one or another quit-smokingmethod spontaneously, then those more determined toquit may tend to choose a different method than thosewho aren’t really trying.

• This will make the method selected by the quitters lookgood

• You overcome this difficulty by conducting an experimentin which people who want to quit smoking are randomlyassigned to different treatments.

Page 52: Math 124: Lecture for Week 4 of 17online.sfsu.edu/meredith/statistics/S2008/Lecture_4_slides.pdfWeek 3 Questions and Answers Quiz Review Standard Deviation Review Case I Review Case

Math 124:Week 3

Questions andAnswers

Quiz

ReviewStandard Deviation

Review Case I

Review Case II

Review Case III

Sampling

DesigningStudiesObservationalStudies

Experimental Studies

Surveys

Identify the design and other features of a study

• On a test, please don’t limit yourself to saying: “Adisadvantage of observational studies is that they mayfail to uncover lurking variables”.

• Always add: “For example, ...”• Your answer should demonstrate an understanding of

the terms “observational study” and “lurking variable”.

Page 53: Math 124: Lecture for Week 4 of 17online.sfsu.edu/meredith/statistics/S2008/Lecture_4_slides.pdfWeek 3 Questions and Answers Quiz Review Standard Deviation Review Case I Review Case

Math 124:Week 3

Questions andAnswers

Quiz

ReviewStandard Deviation

Review Case I

Review Case II

Review Case III

Sampling

DesigningStudiesObservationalStudies

Experimental Studies

Surveys

Identify the design and other features of a study

• On a test, please don’t limit yourself to saying: “Adisadvantage of observational studies is that they mayfail to uncover lurking variables”.

• Always add: “For example, ...”

• Your answer should demonstrate an understanding ofthe terms “observational study” and “lurking variable”.

Page 54: Math 124: Lecture for Week 4 of 17online.sfsu.edu/meredith/statistics/S2008/Lecture_4_slides.pdfWeek 3 Questions and Answers Quiz Review Standard Deviation Review Case I Review Case

Math 124:Week 3

Questions andAnswers

Quiz

ReviewStandard Deviation

Review Case I

Review Case II

Review Case III

Sampling

DesigningStudiesObservationalStudies

Experimental Studies

Surveys

Identify the design and other features of a study

• On a test, please don’t limit yourself to saying: “Adisadvantage of observational studies is that they mayfail to uncover lurking variables”.

• Always add: “For example, ...”• Your answer should demonstrate an understanding of

the terms “observational study” and “lurking variable”.

Page 55: Math 124: Lecture for Week 4 of 17online.sfsu.edu/meredith/statistics/S2008/Lecture_4_slides.pdfWeek 3 Questions and Answers Quiz Review Standard Deviation Review Case I Review Case

Math 124:Week 3

Questions andAnswers

Quiz

ReviewStandard Deviation

Review Case I

Review Case II

Review Case III

Sampling

DesigningStudiesObservationalStudies

Experimental Studies

Surveys

Explain how the study design impacts the type of

conclusions that can be drawn

• Experimental studies put subjects in artificial situationsto observe behavior

• Advantage: can control for lurking variables bycontrolling situation

• Example: studying smoking

Page 56: Math 124: Lecture for Week 4 of 17online.sfsu.edu/meredith/statistics/S2008/Lecture_4_slides.pdfWeek 3 Questions and Answers Quiz Review Standard Deviation Review Case I Review Case

Math 124:Week 3

Questions andAnswers

Quiz

ReviewStandard Deviation

Review Case I

Review Case II

Review Case III

Sampling

DesigningStudiesObservationalStudies

Experimental Studies

Surveys

Explain how the study design impacts the type of

conclusions that can be drawn

• Experimental studies put subjects in artificial situationsto observe behavior

• Advantage: can control for lurking variables bycontrolling situation

• Example: studying smoking

Page 57: Math 124: Lecture for Week 4 of 17online.sfsu.edu/meredith/statistics/S2008/Lecture_4_slides.pdfWeek 3 Questions and Answers Quiz Review Standard Deviation Review Case I Review Case

Math 124:Week 3

Questions andAnswers

Quiz

ReviewStandard Deviation

Review Case I

Review Case II

Review Case III

Sampling

DesigningStudiesObservationalStudies

Experimental Studies

Surveys

Explain how the study design impacts the type of

conclusions that can be drawn

Disadvantages of experimental studies

• subjects may not behave naturally (lack of ecologicalvalidity)

• Yale University psychologist Stanley Milgram 1963conducted experiment where subjects supposedlyadministered shocks to help learners (who were reallyactors). Learners simulated pain, but subject, on ordersof experimenter kept inflicting greater and greatershocks.

• What did this experiment prove?• Question 11: can you see any ethical problems with this

experiment

Page 58: Math 124: Lecture for Week 4 of 17online.sfsu.edu/meredith/statistics/S2008/Lecture_4_slides.pdfWeek 3 Questions and Answers Quiz Review Standard Deviation Review Case I Review Case

Math 124:Week 3

Questions andAnswers

Quiz

ReviewStandard Deviation

Review Case I

Review Case II

Review Case III

Sampling

DesigningStudiesObservationalStudies

Experimental Studies

Surveys

Explain how the study design impacts the type of

conclusions that can be drawn

Disadvantages of experimental studies• subjects may not behave naturally (lack of ecological

validity)

• Yale University psychologist Stanley Milgram 1963conducted experiment where subjects supposedlyadministered shocks to help learners (who were reallyactors). Learners simulated pain, but subject, on ordersof experimenter kept inflicting greater and greatershocks.

• What did this experiment prove?• Question 11: can you see any ethical problems with this

experiment

Page 59: Math 124: Lecture for Week 4 of 17online.sfsu.edu/meredith/statistics/S2008/Lecture_4_slides.pdfWeek 3 Questions and Answers Quiz Review Standard Deviation Review Case I Review Case

Math 124:Week 3

Questions andAnswers

Quiz

ReviewStandard Deviation

Review Case I

Review Case II

Review Case III

Sampling

DesigningStudiesObservationalStudies

Experimental Studies

Surveys

Explain how the study design impacts the type of

conclusions that can be drawn

Disadvantages of experimental studies• subjects may not behave naturally (lack of ecological

validity)• Yale University psychologist Stanley Milgram 1963

conducted experiment where subjects supposedlyadministered shocks to help learners (who were reallyactors). Learners simulated pain, but subject, on ordersof experimenter kept inflicting greater and greatershocks.

• What did this experiment prove?• Question 11: can you see any ethical problems with this

experiment

Page 60: Math 124: Lecture for Week 4 of 17online.sfsu.edu/meredith/statistics/S2008/Lecture_4_slides.pdfWeek 3 Questions and Answers Quiz Review Standard Deviation Review Case I Review Case

Math 124:Week 3

Questions andAnswers

Quiz

ReviewStandard Deviation

Review Case I

Review Case II

Review Case III

Sampling

DesigningStudiesObservationalStudies

Experimental Studies

Surveys

Explain how the study design impacts the type of

conclusions that can be drawn

Disadvantages of experimental studies• subjects may not behave naturally (lack of ecological

validity)• Yale University psychologist Stanley Milgram 1963

conducted experiment where subjects supposedlyadministered shocks to help learners (who were reallyactors). Learners simulated pain, but subject, on ordersof experimenter kept inflicting greater and greatershocks.

• What did this experiment prove?

• Question 11: can you see any ethical problems with thisexperiment

Page 61: Math 124: Lecture for Week 4 of 17online.sfsu.edu/meredith/statistics/S2008/Lecture_4_slides.pdfWeek 3 Questions and Answers Quiz Review Standard Deviation Review Case I Review Case

Math 124:Week 3

Questions andAnswers

Quiz

ReviewStandard Deviation

Review Case I

Review Case II

Review Case III

Sampling

DesigningStudiesObservationalStudies

Experimental Studies

Surveys

Explain how the study design impacts the type of

conclusions that can be drawn

Disadvantages of experimental studies• subjects may not behave naturally (lack of ecological

validity)• Yale University psychologist Stanley Milgram 1963

conducted experiment where subjects supposedlyadministered shocks to help learners (who were reallyactors). Learners simulated pain, but subject, on ordersof experimenter kept inflicting greater and greatershocks.

• What did this experiment prove?• Question 11: can you see any ethical problems with this

experiment

Page 62: Math 124: Lecture for Week 4 of 17online.sfsu.edu/meredith/statistics/S2008/Lecture_4_slides.pdfWeek 3 Questions and Answers Quiz Review Standard Deviation Review Case I Review Case

Math 124:Week 3

Questions andAnswers

Quiz

ReviewStandard Deviation

Review Case I

Review Case II

Review Case III

Sampling

DesigningStudiesObservationalStudies

Experimental Studies

Surveys

Explain how the study design impacts the type of

conclusions that can be drawn

Disadvantages of experimental studies• subjects may not behave naturally (lack of ecological

validity)• Yale University psychologist Stanley Milgram 1963

conducted experiment where subjects supposedlyadministered shocks to help learners (who were reallyactors). Learners simulated pain, but subject, on ordersof experimenter kept inflicting greater and greatershocks.

• What did this experiment prove?• Question 11: can you see any ethical problems with this

experiment

Page 63: Math 124: Lecture for Week 4 of 17online.sfsu.edu/meredith/statistics/S2008/Lecture_4_slides.pdfWeek 3 Questions and Answers Quiz Review Standard Deviation Review Case I Review Case

Math 124:Week 3

Questions andAnswers

Quiz

ReviewStandard Deviation

Review Case I

Review Case II

Review Case III

Sampling

DesigningStudiesObservationalStudies

Experimental Studies

Surveys

Explain how the study design impacts the type of

conclusions that can be drawn

Blind and double-blind studies, use of placebos areattempts to improve ecological validity of studies

• An experiment is blind if the subject doesn’t know whichtreatment they are receiving

• An experiment is double blind if the subject and thoseadministering the treatments don’t know whichtreatment they are using

• A placebo is a treatment that has no effect. It is used sothat subjects and those administering treatments don’tknow who is in the “no treatment” group.

• Many but not all experiments benefit from the use ofblind and double-blind protocols and the use ofplacebos.

Page 64: Math 124: Lecture for Week 4 of 17online.sfsu.edu/meredith/statistics/S2008/Lecture_4_slides.pdfWeek 3 Questions and Answers Quiz Review Standard Deviation Review Case I Review Case

Math 124:Week 3

Questions andAnswers

Quiz

ReviewStandard Deviation

Review Case I

Review Case II

Review Case III

Sampling

DesigningStudiesObservationalStudies

Experimental Studies

Surveys

Explain how the study design impacts the type of

conclusions that can be drawn

Blind and double-blind studies, use of placebos areattempts to improve ecological validity of studies• An experiment is blind if the subject doesn’t know which

treatment they are receiving

• An experiment is double blind if the subject and thoseadministering the treatments don’t know whichtreatment they are using

• A placebo is a treatment that has no effect. It is used sothat subjects and those administering treatments don’tknow who is in the “no treatment” group.

• Many but not all experiments benefit from the use ofblind and double-blind protocols and the use ofplacebos.

Page 65: Math 124: Lecture for Week 4 of 17online.sfsu.edu/meredith/statistics/S2008/Lecture_4_slides.pdfWeek 3 Questions and Answers Quiz Review Standard Deviation Review Case I Review Case

Math 124:Week 3

Questions andAnswers

Quiz

ReviewStandard Deviation

Review Case I

Review Case II

Review Case III

Sampling

DesigningStudiesObservationalStudies

Experimental Studies

Surveys

Explain how the study design impacts the type of

conclusions that can be drawn

Blind and double-blind studies, use of placebos areattempts to improve ecological validity of studies• An experiment is blind if the subject doesn’t know which

treatment they are receiving• An experiment is double blind if the subject and those

administering the treatments don’t know whichtreatment they are using

• A placebo is a treatment that has no effect. It is used sothat subjects and those administering treatments don’tknow who is in the “no treatment” group.

• Many but not all experiments benefit from the use ofblind and double-blind protocols and the use ofplacebos.

Page 66: Math 124: Lecture for Week 4 of 17online.sfsu.edu/meredith/statistics/S2008/Lecture_4_slides.pdfWeek 3 Questions and Answers Quiz Review Standard Deviation Review Case I Review Case

Math 124:Week 3

Questions andAnswers

Quiz

ReviewStandard Deviation

Review Case I

Review Case II

Review Case III

Sampling

DesigningStudiesObservationalStudies

Experimental Studies

Surveys

Explain how the study design impacts the type of

conclusions that can be drawn

Blind and double-blind studies, use of placebos areattempts to improve ecological validity of studies• An experiment is blind if the subject doesn’t know which

treatment they are receiving• An experiment is double blind if the subject and those

administering the treatments don’t know whichtreatment they are using

• A placebo is a treatment that has no effect. It is used sothat subjects and those administering treatments don’tknow who is in the “no treatment” group.

• Many but not all experiments benefit from the use ofblind and double-blind protocols and the use ofplacebos.

Page 67: Math 124: Lecture for Week 4 of 17online.sfsu.edu/meredith/statistics/S2008/Lecture_4_slides.pdfWeek 3 Questions and Answers Quiz Review Standard Deviation Review Case I Review Case

Math 124:Week 3

Questions andAnswers

Quiz

ReviewStandard Deviation

Review Case I

Review Case II

Review Case III

Sampling

DesigningStudiesObservationalStudies

Experimental Studies

Surveys

Explain how the study design impacts the type of

conclusions that can be drawn

More disadvantages of experimental studies

• Subjects may not follow protocols or report inaccurately• Some experiments may be immoral.

• Could you ask a random group of drivers not to useseatbelts?

• Medical experiments are sometimes stoppedmidstream when dangers become clear

• On March 2, 2004, NIH announced that it had stoppedthe estrogen-alone study in the interest of safety aftercareful consideration of preliminary data and an averagefollow-up of nearly 7 years http://www.nhlbi.nih.gov/whi/estro_alone.htm

Page 68: Math 124: Lecture for Week 4 of 17online.sfsu.edu/meredith/statistics/S2008/Lecture_4_slides.pdfWeek 3 Questions and Answers Quiz Review Standard Deviation Review Case I Review Case

Math 124:Week 3

Questions andAnswers

Quiz

ReviewStandard Deviation

Review Case I

Review Case II

Review Case III

Sampling

DesigningStudiesObservationalStudies

Experimental Studies

Surveys

Explain how the study design impacts the type of

conclusions that can be drawn

More disadvantages of experimental studies• Subjects may not follow protocols or report inaccurately

• Some experiments may be immoral.• Could you ask a random group of drivers not to use

seatbelts?• Medical experiments are sometimes stopped

midstream when dangers become clear• On March 2, 2004, NIH announced that it had stopped

the estrogen-alone study in the interest of safety aftercareful consideration of preliminary data and an averagefollow-up of nearly 7 years http://www.nhlbi.nih.gov/whi/estro_alone.htm

Page 69: Math 124: Lecture for Week 4 of 17online.sfsu.edu/meredith/statistics/S2008/Lecture_4_slides.pdfWeek 3 Questions and Answers Quiz Review Standard Deviation Review Case I Review Case

Math 124:Week 3

Questions andAnswers

Quiz

ReviewStandard Deviation

Review Case I

Review Case II

Review Case III

Sampling

DesigningStudiesObservationalStudies

Experimental Studies

Surveys

Explain how the study design impacts the type of

conclusions that can be drawn

More disadvantages of experimental studies• Subjects may not follow protocols or report inaccurately• Some experiments may be immoral.

• Could you ask a random group of drivers not to useseatbelts?

• Medical experiments are sometimes stoppedmidstream when dangers become clear

• On March 2, 2004, NIH announced that it had stoppedthe estrogen-alone study in the interest of safety aftercareful consideration of preliminary data and an averagefollow-up of nearly 7 years http://www.nhlbi.nih.gov/whi/estro_alone.htm

Page 70: Math 124: Lecture for Week 4 of 17online.sfsu.edu/meredith/statistics/S2008/Lecture_4_slides.pdfWeek 3 Questions and Answers Quiz Review Standard Deviation Review Case I Review Case

Math 124:Week 3

Questions andAnswers

Quiz

ReviewStandard Deviation

Review Case I

Review Case II

Review Case III

Sampling

DesigningStudiesObservationalStudies

Experimental Studies

Surveys

Explain how the study design impacts the type of

conclusions that can be drawn

More disadvantages of experimental studies• Subjects may not follow protocols or report inaccurately• Some experiments may be immoral.

• Could you ask a random group of drivers not to useseatbelts?

• Medical experiments are sometimes stoppedmidstream when dangers become clear

• On March 2, 2004, NIH announced that it had stoppedthe estrogen-alone study in the interest of safety aftercareful consideration of preliminary data and an averagefollow-up of nearly 7 years http://www.nhlbi.nih.gov/whi/estro_alone.htm

Page 71: Math 124: Lecture for Week 4 of 17online.sfsu.edu/meredith/statistics/S2008/Lecture_4_slides.pdfWeek 3 Questions and Answers Quiz Review Standard Deviation Review Case I Review Case

Math 124:Week 3

Questions andAnswers

Quiz

ReviewStandard Deviation

Review Case I

Review Case II

Review Case III

Sampling

DesigningStudiesObservationalStudies

Experimental Studies

Surveys

Explain how the study design impacts the type of

conclusions that can be drawn

More disadvantages of experimental studies• Subjects may not follow protocols or report inaccurately• Some experiments may be immoral.

• Could you ask a random group of drivers not to useseatbelts?

• Medical experiments are sometimes stoppedmidstream when dangers become clear

• On March 2, 2004, NIH announced that it had stoppedthe estrogen-alone study in the interest of safety aftercareful consideration of preliminary data and an averagefollow-up of nearly 7 years http://www.nhlbi.nih.gov/whi/estro_alone.htm

Page 72: Math 124: Lecture for Week 4 of 17online.sfsu.edu/meredith/statistics/S2008/Lecture_4_slides.pdfWeek 3 Questions and Answers Quiz Review Standard Deviation Review Case I Review Case

Math 124:Week 3

Questions andAnswers

Quiz

ReviewStandard Deviation

Review Case I

Review Case II

Review Case III

Sampling

DesigningStudiesObservationalStudies

Experimental Studies

Surveys

Determine how features of a survey impact the collected

data and its accuracy

Surveys

• Advantage: Cheap, easy to perform• Disadvantage: must be carefully constructed

Page 73: Math 124: Lecture for Week 4 of 17online.sfsu.edu/meredith/statistics/S2008/Lecture_4_slides.pdfWeek 3 Questions and Answers Quiz Review Standard Deviation Review Case I Review Case

Math 124:Week 3

Questions andAnswers

Quiz

ReviewStandard Deviation

Review Case I

Review Case II

Review Case III

Sampling

DesigningStudiesObservationalStudies

Experimental Studies

Surveys

Determine how features of a survey impact the collected

data and its accuracy

Surveys

• Advantage: Cheap, easy to perform

• Disadvantage: must be carefully constructed

Page 74: Math 124: Lecture for Week 4 of 17online.sfsu.edu/meredith/statistics/S2008/Lecture_4_slides.pdfWeek 3 Questions and Answers Quiz Review Standard Deviation Review Case I Review Case

Math 124:Week 3

Questions andAnswers

Quiz

ReviewStandard Deviation

Review Case I

Review Case II

Review Case III

Sampling

DesigningStudiesObservationalStudies

Experimental Studies

Surveys

Determine how features of a survey impact the collected

data and its accuracy

Surveys

• Advantage: Cheap, easy to perform• Disadvantage: must be carefully constructed

Page 75: Math 124: Lecture for Week 4 of 17online.sfsu.edu/meredith/statistics/S2008/Lecture_4_slides.pdfWeek 3 Questions and Answers Quiz Review Standard Deviation Review Case I Review Case

Math 124:Week 3

Questions andAnswers

Quiz

ReviewStandard Deviation

Review Case I

Review Case II

Review Case III

Sampling

DesigningStudiesObservationalStudies

Experimental Studies

Surveys

Determine how features of a survey impact the collected

data and its accuracy

Problematic survey questions

• limited response options• unbalanced response options• leading questions

• Here’s a real question taken from a political survey:"Some people say (your Democratic incumbent) is astrong supporter of Hillary Clinton and will support herliberal agenda of big government and higher taxes ifshe becomes President. Do you plan to vote for thisperson in November?" (Washington Post, Letter to theEditor, 9/29/07)

• Political surveys like this are called push polls

• planting ideas with questions• complicated questions• sensitive questions

Page 76: Math 124: Lecture for Week 4 of 17online.sfsu.edu/meredith/statistics/S2008/Lecture_4_slides.pdfWeek 3 Questions and Answers Quiz Review Standard Deviation Review Case I Review Case

Math 124:Week 3

Questions andAnswers

Quiz

ReviewStandard Deviation

Review Case I

Review Case II

Review Case III

Sampling

DesigningStudiesObservationalStudies

Experimental Studies

Surveys

Determine how features of a survey impact the collected

data and its accuracy

Problematic survey questions

• limited response options

• unbalanced response options• leading questions

• Here’s a real question taken from a political survey:"Some people say (your Democratic incumbent) is astrong supporter of Hillary Clinton and will support herliberal agenda of big government and higher taxes ifshe becomes President. Do you plan to vote for thisperson in November?" (Washington Post, Letter to theEditor, 9/29/07)

• Political surveys like this are called push polls

• planting ideas with questions• complicated questions• sensitive questions

Page 77: Math 124: Lecture for Week 4 of 17online.sfsu.edu/meredith/statistics/S2008/Lecture_4_slides.pdfWeek 3 Questions and Answers Quiz Review Standard Deviation Review Case I Review Case

Math 124:Week 3

Questions andAnswers

Quiz

ReviewStandard Deviation

Review Case I

Review Case II

Review Case III

Sampling

DesigningStudiesObservationalStudies

Experimental Studies

Surveys

Determine how features of a survey impact the collected

data and its accuracy

Problematic survey questions

• limited response options• unbalanced response options

• leading questions• Here’s a real question taken from a political survey:

"Some people say (your Democratic incumbent) is astrong supporter of Hillary Clinton and will support herliberal agenda of big government and higher taxes ifshe becomes President. Do you plan to vote for thisperson in November?" (Washington Post, Letter to theEditor, 9/29/07)

• Political surveys like this are called push polls

• planting ideas with questions• complicated questions• sensitive questions

Page 78: Math 124: Lecture for Week 4 of 17online.sfsu.edu/meredith/statistics/S2008/Lecture_4_slides.pdfWeek 3 Questions and Answers Quiz Review Standard Deviation Review Case I Review Case

Math 124:Week 3

Questions andAnswers

Quiz

ReviewStandard Deviation

Review Case I

Review Case II

Review Case III

Sampling

DesigningStudiesObservationalStudies

Experimental Studies

Surveys

Determine how features of a survey impact the collected

data and its accuracy

Problematic survey questions

• limited response options• unbalanced response options• leading questions

• Here’s a real question taken from a political survey:"Some people say (your Democratic incumbent) is astrong supporter of Hillary Clinton and will support herliberal agenda of big government and higher taxes ifshe becomes President. Do you plan to vote for thisperson in November?" (Washington Post, Letter to theEditor, 9/29/07)

• Political surveys like this are called push polls

• planting ideas with questions• complicated questions• sensitive questions

Page 79: Math 124: Lecture for Week 4 of 17online.sfsu.edu/meredith/statistics/S2008/Lecture_4_slides.pdfWeek 3 Questions and Answers Quiz Review Standard Deviation Review Case I Review Case

Math 124:Week 3

Questions andAnswers

Quiz

ReviewStandard Deviation

Review Case I

Review Case II

Review Case III

Sampling

DesigningStudiesObservationalStudies

Experimental Studies

Surveys

Determine how features of a survey impact the collected

data and its accuracy

Problematic survey questions

• limited response options• unbalanced response options• leading questions

• Here’s a real question taken from a political survey:"Some people say (your Democratic incumbent) is astrong supporter of Hillary Clinton and will support herliberal agenda of big government and higher taxes ifshe becomes President. Do you plan to vote for thisperson in November?" (Washington Post, Letter to theEditor, 9/29/07)

• Political surveys like this are called push polls

• planting ideas with questions

• complicated questions• sensitive questions

Page 80: Math 124: Lecture for Week 4 of 17online.sfsu.edu/meredith/statistics/S2008/Lecture_4_slides.pdfWeek 3 Questions and Answers Quiz Review Standard Deviation Review Case I Review Case

Math 124:Week 3

Questions andAnswers

Quiz

ReviewStandard Deviation

Review Case I

Review Case II

Review Case III

Sampling

DesigningStudiesObservationalStudies

Experimental Studies

Surveys

Determine how features of a survey impact the collected

data and its accuracy

Problematic survey questions

• limited response options• unbalanced response options• leading questions

• Here’s a real question taken from a political survey:"Some people say (your Democratic incumbent) is astrong supporter of Hillary Clinton and will support herliberal agenda of big government and higher taxes ifshe becomes President. Do you plan to vote for thisperson in November?" (Washington Post, Letter to theEditor, 9/29/07)

• Political surveys like this are called push polls

• planting ideas with questions• complicated questions

• sensitive questions

Page 81: Math 124: Lecture for Week 4 of 17online.sfsu.edu/meredith/statistics/S2008/Lecture_4_slides.pdfWeek 3 Questions and Answers Quiz Review Standard Deviation Review Case I Review Case

Math 124:Week 3

Questions andAnswers

Quiz

ReviewStandard Deviation

Review Case I

Review Case II

Review Case III

Sampling

DesigningStudiesObservationalStudies

Experimental Studies

Surveys

Determine how features of a survey impact the collected

data and its accuracy

Problematic survey questions

• limited response options• unbalanced response options• leading questions

• Here’s a real question taken from a political survey:"Some people say (your Democratic incumbent) is astrong supporter of Hillary Clinton and will support herliberal agenda of big government and higher taxes ifshe becomes President. Do you plan to vote for thisperson in November?" (Washington Post, Letter to theEditor, 9/29/07)

• Political surveys like this are called push polls

• planting ideas with questions• complicated questions• sensitive questions

Page 82: Math 124: Lecture for Week 4 of 17online.sfsu.edu/meredith/statistics/S2008/Lecture_4_slides.pdfWeek 3 Questions and Answers Quiz Review Standard Deviation Review Case I Review Case

Math 124:Week 3

Questions andAnswers

Quiz

ReviewStandard Deviation

Review Case I

Review Case II

Review Case III

Sampling

DesigningStudiesObservationalStudies

Experimental Studies

Surveys

Determine how features of a survey impact the collected

data and its accuracy

“Randomized responses” to deal with sensitive questions

• Want to ask people if they have done something bad(fill in your own example)

• Tell them to flip a coin and not show it to you• If heads, answer truthfully• If tails, answer “yes”

• At end of study, remove a number of yes answers equalto half the subjects used

• Question 12: Would this method work if the instructionsread: If tails, answer “no”

Page 83: Math 124: Lecture for Week 4 of 17online.sfsu.edu/meredith/statistics/S2008/Lecture_4_slides.pdfWeek 3 Questions and Answers Quiz Review Standard Deviation Review Case I Review Case

Math 124:Week 3

Questions andAnswers

Quiz

ReviewStandard Deviation

Review Case I

Review Case II

Review Case III

Sampling

DesigningStudiesObservationalStudies

Experimental Studies

Surveys

Determine how features of a survey impact the collected

data and its accuracy

“Randomized responses” to deal with sensitive questions• Want to ask people if they have done something bad

(fill in your own example)

• Tell them to flip a coin and not show it to you• If heads, answer truthfully• If tails, answer “yes”

• At end of study, remove a number of yes answers equalto half the subjects used

• Question 12: Would this method work if the instructionsread: If tails, answer “no”

Page 84: Math 124: Lecture for Week 4 of 17online.sfsu.edu/meredith/statistics/S2008/Lecture_4_slides.pdfWeek 3 Questions and Answers Quiz Review Standard Deviation Review Case I Review Case

Math 124:Week 3

Questions andAnswers

Quiz

ReviewStandard Deviation

Review Case I

Review Case II

Review Case III

Sampling

DesigningStudiesObservationalStudies

Experimental Studies

Surveys

Determine how features of a survey impact the collected

data and its accuracy

“Randomized responses” to deal with sensitive questions• Want to ask people if they have done something bad

(fill in your own example)• Tell them to flip a coin and not show it to you

• If heads, answer truthfully• If tails, answer “yes”

• At end of study, remove a number of yes answers equalto half the subjects used

• Question 12: Would this method work if the instructionsread: If tails, answer “no”

Page 85: Math 124: Lecture for Week 4 of 17online.sfsu.edu/meredith/statistics/S2008/Lecture_4_slides.pdfWeek 3 Questions and Answers Quiz Review Standard Deviation Review Case I Review Case

Math 124:Week 3

Questions andAnswers

Quiz

ReviewStandard Deviation

Review Case I

Review Case II

Review Case III

Sampling

DesigningStudiesObservationalStudies

Experimental Studies

Surveys

Determine how features of a survey impact the collected

data and its accuracy

“Randomized responses” to deal with sensitive questions• Want to ask people if they have done something bad

(fill in your own example)• Tell them to flip a coin and not show it to you

• If heads, answer truthfully• If tails, answer “yes”

• At end of study, remove a number of yes answers equalto half the subjects used

• Question 12: Would this method work if the instructionsread: If tails, answer “no”

Page 86: Math 124: Lecture for Week 4 of 17online.sfsu.edu/meredith/statistics/S2008/Lecture_4_slides.pdfWeek 3 Questions and Answers Quiz Review Standard Deviation Review Case I Review Case

Math 124:Week 3

Questions andAnswers

Quiz

ReviewStandard Deviation

Review Case I

Review Case II

Review Case III

Sampling

DesigningStudiesObservationalStudies

Experimental Studies

Surveys

Determine how features of a survey impact the collected

data and its accuracy

“Randomized responses” to deal with sensitive questions• Want to ask people if they have done something bad

(fill in your own example)• Tell them to flip a coin and not show it to you

• If heads, answer truthfully• If tails, answer “yes”

• At end of study, remove a number of yes answers equalto half the subjects used

• Question 12: Would this method work if the instructionsread: If tails, answer “no”

Page 87: Math 124: Lecture for Week 4 of 17online.sfsu.edu/meredith/statistics/S2008/Lecture_4_slides.pdfWeek 3 Questions and Answers Quiz Review Standard Deviation Review Case I Review Case

Math 124:Week 3

Questions andAnswers

Quiz

ReviewStandard Deviation

Review Case I

Review Case II

Review Case III

Sampling

DesigningStudiesObservationalStudies

Experimental Studies

Surveys

Determine how features of a survey impact the collected

data and its accuracy

“Randomized responses” to deal with sensitive questions• Want to ask people if they have done something bad

(fill in your own example)• Tell them to flip a coin and not show it to you

• If heads, answer truthfully• If tails, answer “yes”

• At end of study, remove a number of yes answers equalto half the subjects used

• Question 12: Would this method work if the instructionsread: If tails, answer “no”