24.04.2015 Views

Chapter 6 ppt

Chapter 6 ppt

Chapter 6 ppt

SHOW MORE
SHOW LESS

You also want an ePaper? Increase the reach of your titles

YUMPU automatically turns print PDFs into web optimized ePapers that Google loves.

<strong>Chapter</strong> 6<br />

Experiments in the Real World<br />

<strong>Chapter</strong> 6 1


Thought Question 1<br />

Suppose you are interested in determining if drinking a glass<br />

of red wine each day helps prevent heartburn. You recruit<br />

40 adults age 50 and older to participate in an experiment.<br />

You want half of them to drink a glass of red wine each day<br />

and the other half to not do so. You ask them which they<br />

would prefer, and 20 say they would like to drink the red<br />

wine and the other 20 say they would not. You ask each of<br />

them to record how many cases of heartburn they have in<br />

the next six months. At the end of that time period, you<br />

compare the results reported from the two groups. Give<br />

three reasons why this is not a good experiment.<br />

<strong>Chapter</strong> 6 2


Experiments: Some Techniques<br />

Double-blinding<br />

– to control experimenter/respondent bias<br />

Pairing or blocking<br />

– to reduce a source of variability in responses<br />

– the same or similar subjects receive each<br />

treatment<br />

different from a completely randomized design,<br />

where all subjects are allocated at random<br />

among all treatments<br />

<strong>Chapter</strong> 6 3


Double-Blinded:<br />

Case Study<br />

Quitting Smoking with Nicotine Patches<br />

(JAMA, Feb. 23, 1994, pp. 595-600)<br />

Variables:<br />

– Explanatory: Treatment assignment<br />

– Response: Cessation of smoking (yes/no)<br />

Double-blinded<br />

– Participants do not know which patch they<br />

received<br />

– Nor do those measuring smoking behavior<br />

<strong>Chapter</strong> 6 4


Pairing or Blocking:<br />

Case Study<br />

Quitting Smoking with Nicotine Patches<br />

(JAMA, Feb. 23, 1994, pp. 595-600)<br />

Variables:<br />

– Explanatory: Treatment assignment<br />

– Response: Cessation of smoking (yes/no)<br />

Blocking?<br />

– consider men and women separately?<br />

– could use a matched-pairs design<br />

– self-pairing? Not if cessation is the goal<br />

<strong>Chapter</strong> 6 5


Experiments:<br />

Difficulties and Disasters<br />

Extraneous variables<br />

– Confounding variables (in chapter 5)<br />

– Interacting variables<br />

Hawthorne, placebo and experimenter<br />

effects<br />

Refusals, nonadherers, dropouts<br />

Extending the results (generalizing)<br />

<strong>Chapter</strong> 6 6


Interacting Variables<br />

The problem:<br />

– effect of explanatory variable on response variable<br />

may vary over levels of other variables.<br />

The solution:<br />

– measure and study potential interacting variables.<br />

does the relationship between explanatory and response<br />

variables change for different levels of these interacting<br />

variables?<br />

if so, report results for different groups defined by the levels of<br />

the interacting variables.<br />

<strong>Chapter</strong> 6 7


Interacting Variables:<br />

Case Study<br />

Quitting Smoking with Nicotine Patches<br />

(JAMA, Feb. 23, 1994, pp. 595-600)<br />

Researchers considered:<br />

– smoker at home<br />

found this to be an interacting variable:<br />

Percent quitting Nicotine Placebo<br />

Smoker at home 31% 20%<br />

No smoker at home 58% 20%<br />

– other variables: age, weight, depression<br />

no interactions found<br />

<strong>Chapter</strong> 6 8


Hawthorne, Placebo and<br />

The problem:<br />

Experimenter Effects<br />

– people may respond differently when they<br />

know they are part of an experiment.<br />

The solution:<br />

– use placebos, control groups, and doubleblind<br />

studies when possible.<br />

<strong>Chapter</strong> 6 9


Hawthorne, Placebo and<br />

Experimenter Effects :<br />

Case Study I<br />

1920’s Experiment by Hawthorne Works<br />

of the Western Electric Company<br />

What changes in working conditions<br />

improve productivity of workers?<br />

– More lighting?<br />

– Less lighting?<br />

– Other changes?<br />

All changes improved productivity!<br />

Workers behaved differently when they<br />

knew they were part of a study.<br />

<strong>Chapter</strong> 6 10


Hawthorne, Placebo and<br />

Experimenter Effects :<br />

Case Study II<br />

Experimenter Effects in Behavioral Research<br />

(Rosenthal, 1976, Irvington Pub., p. 410)<br />

Teachers given a list of student names<br />

– told these were students “who would show unusual<br />

academic development.”<br />

IQ was measured at end of year<br />

– first graders on list: 15 points higher<br />

– second graders on list: 9.5 points higher<br />

– older: no striking difference<br />

Great expectations = self-fulfilling prophecy<br />

– students were randomly selected (did not have high IQ)<br />

<strong>Chapter</strong> 6 11


Extending the Results<br />

( Can We Generalize? )<br />

The problem:<br />

– lack of generalizability due to:<br />

unrealistic treatments<br />

unnatural settings<br />

sample that is not representative of population<br />

The solution:<br />

– Researchers should use natural settings<br />

with a properly chosen sample.<br />

<strong>Chapter</strong> 6 12


Extending the Results :<br />

Case Study<br />

Does Aspirin Prevent Heart Attacks?<br />

(NEJM, Jan. 28, 1988, pp. 262-264)<br />

Participants were measured in their<br />

natural setting (at home)<br />

Only healthy male physicians were<br />

participants<br />

– Results may not apply to:<br />

male physical laborers<br />

women<br />

<strong>Chapter</strong> 6 13


Key Concepts<br />

Double-Blind Experiment<br />

Difficulties and Disasters<br />

Experimental Designs<br />

– Completely Randomized Design<br />

– Matched Pairs Design<br />

– Block Design<br />

<strong>Chapter</strong> 6 14

Hooray! Your file is uploaded and ready to be published.

Saved successfully!

Ooh no, something went wrong!