INTR STATISTCL REASONING
INTR STATISTCL REASONING STAT 110
Popular in Course
Popular in Statistics
verified elite notetaker
This 10 page Class Notes was uploaded by Shane Marks on Monday October 26, 2015. The Class Notes belongs to STAT 110 at University of South Carolina - Columbia taught by G. Baker in Fall. Since its upload, it has received 30 views. For similar materials see /class/229647/stat-110-university-of-south-carolina-columbia in Statistics at University of South Carolina - Columbia.
Reviews for INTR STATISTCL REASONING
Report this Material
What is Karma?
Karma is the currency of StudySoup.
Date Created: 10/26/15
Experiments in the Real World Chapter 6 I Ideally we want to randomly choose groups I Randomize out any confounding variables I Comparing groups that are equal except for the treatment they are receiving I Demonstrate cause and effect between explanatory and response variables a mm o u nnnn a Experiments Some Real World Problems In a randomized comparative experiment all subjects experience the same conditions except for the treatments that the experiment is trying to compare Easier said than done 52m Depzvlmerl y scum Mmmmmcmm Slide 2 Experimental Bias I Just as poor sampling creates bias in survey results unequal conditions among groups creates bias in an experiment a my Depzvlmerl y scum Univevxilyofsmlhczvdim Slide 4 DoubleBlind Experiments I In a double blind experiment neither the subjects nor the people who are working with them know which treatment each subject is receiving More Problems Nonadherers and Dropouts IA nonadherer is a subject who participates but does not follow the experimental protocol IA dropout IS a subject who begins the experiment but does not complete it Nonadherers amp Bias Results 6 52m Depzvlmerl y scum Univevxilyofsmlhczvdim Slide 5 Dropouts Generalizing from Experiment to Population I First our findings must be statistically significant I Treatments subjects and environment must be realistic for the population Often the direction ofthe experimental study is co amount ofthe change is disappointing U35 w Completely Randomized Experimental Design I In a completely randomized experimental design all the subjects are randomly placed into the groups and the treatments are randomly assigned to each group I If there is more than one explanatory variable the number of treatments Groups number of values for explanatory variable 1 number of values for explanatory variable 2 number of values for explanatory variable 3 a an DEDzvlmerl a Slzhxliz Uninintyarsinmoinim Slide x Completely Randomized Experimental Design I By varying more than one explanatory variable at once experimenters can study the combined effects of several explanatory variables simultaneously I The combined effect of more than one explanatory variable is called an interaction 6 an DEDzvlmerl a Slzhxliz Uninintyarsinmoinim Slide 9 Interactions Can Change Results Time to Relief for 2 Treatments and a Control Taste Scores for Cake mix 10 Yummy And 1 Yuk 300 320 340 Avg 1 Hour 8 8 8 8 125 9 7 5 hours Matched pairs Design 9 w gt t ash W m mmsz ncmm39mms Matched Pairs Design lReduces persun m persun simiier subiets lCDmparEs as m ream em IRandumizaliun and cum parisun is Matched Pairs Design ITHE warmed pair ten be me subject 1mm 1mm Matched Pairs Design with Both Treatments being Applied to One Subject IWhere s the randomizing IWhere s the comparing IStatisticaI Significance a my Depzvlmerl y shim Unvevxilyofsoulh claim sia is Randomized Block Design I Instead of pairing similar subjects we group many similar subjects into a block I Each block contains subjects that are similar in some way that is expected to affect the response to treatments I Random assignment of treatments is within blocks I Comparison can be between blocks as well as between treatments Wigwam claim sia n Randomized Block Design mnewn l Ei g AB c MM gt gt aim BLOCK gt RANDOMIZE gtc0MPARE TREAT S 13335axiiii i Principals of Good Design I Ra ndomize g I Com pare I Enough subjects Data Ethics Chapter 7 Experimental Ethics I Are sham placebo surgeries ethical 6 am a 95mm Univevxilyufsmlhczvdim Slide 2 Sometimes There is No Question It s Unethical I Researcher fakes data I Surveyor starts by saying You have won a prize I Surveyor coaches responder I Researcher biases study when choosing subjects I Researcher does several studies only publishes the one that supports his conclusions Univ When Experiments Impose Treatments on People clinical trials IThe law requires that studies funded by the federal government follow these basic principles 0rganizations carrying out studies must have institutional review boards that review all planned studies in order to protect the subjects from possible harm a m Univ When Experiments Impose Treatments on People clinical trials I The law requires that studies funded by the federal government follow these basic principles All individuals who are subjects in a study must give their informed consent All individual data must be kept confidential Only statistical summaries can be made public a 52m Depzvlmerl y scum Univevxilyufsmlhczvdimi Slide 5 Clinical Trials Spotlight Ethical Problems IWe need randomized comparative experiments to see the true effects of treatments but that means someone is getting a placebo I The benefits of clinical trials go to future patients while the risks go to current subjects I 1964 Helsinki Declaration of the World Medical Association the interests of the subject must always prevail over the interests of science and sggcjeetx 1 pm scum mmcmim Slide 5 Behavioral amp Social Science Experiments Pose Difficult Ethical Issues I We want to observe people s behavior I But people will change their behavior if they know they are being observed I How do we get informed consent I If we don t get informed consent before observing are we invading personal privacy 6 my Depzvlmerl y shim Univuxilvofsuulhczvdim Slide 7 American Psychological Ass n Ethical Principles I Require consent unless study observes behavior in public places I Allow deception when It is necessary to the study and It does not hide information that would influence a subject s willingness to participate and It is explained to subjects as soon as possible afterwards a mm o u quotquotquotquot a