Math 1530 - Chapter 4 GATHERING DATA Chapter 4 Learning Objectives: 1. 2. 3. 4.1 Experiment or Observe? Population versus Sample Types of Studies: Experimental and Observational Comparing Experimental and Observational Studies www.consumersenergy.com Population and Samples Population: Subjects of interest Sample: Subset for whom we have data Often want answers about large group but can’t measure all, so a subset is chosen Use statistical techniques to make conclusions Observational Study Merely observe values of response and explanatory variables without doing anything to the subjects creativejuice.files.wordpress.com 1 Math 1530 - Chapter 4 Sample Survey Experiment Select sample and interview Observational study Census is survey of entire population Imposes certain conditions and observes outcomes Experimental conditions = explanatory variable values = treatments www.austinchronicle.com www.law.georgetown.edu Experiments and Observational Studies Experiment reduces lurking variables and thus outside influences Experiments establish cause and effect, unlike observational studies Some experiments impractical because of ethics, time, money, etc. 4.2 What are Good and Poor Ways to Sample? www.purbeck.gov.uk Learning Objectives: 1. 2. 3. 4. 5. 6. 7. Sampling Frame & Sampling Design Simple Random Sample Random number table Margin of Error Convenience Samples Types of Bias in Sample Surveys Key Parts of Sample Survey Sampling Frame & Sampling Design Sampling frame – list of subjects in (hopefully total) population Sampling design determines how sample is selected www.nedarc.org artstuff.typepad.com 2 Math 1530 - Chapter 4 Simple Random Sampling (SRS) Random Sampling – best way to get representative sample Simple Random Sample – each possible sample of set size n has equal chance of being selected SRS: Table of Random Numbers Pg. A6 of text 1. 2. 3. Number subjects Select numbers from random number table or random number generator Include subjects with random numbers selected www.youthnoise.com Margin of Error for Population Percentages Example: Margin of Error Margin of Error – how well sample predicts population With n subjects in SRS, margin of error is about A survey result states: “The margin of error is plus or minus 3 percentage points” 1 ×100% n 2010 British PM Election i.telegraph.co.uk Convenience Samples: Poor Ways to Sample Convenience Sample: survey sample that’s easy to get Unlikely to represent population Often severe biases Results apply only to observed subjects Convenience Samples: Poor Ways to Sample Volunteer Sample: most common convenience sample where subjects volunteer – not representative www.revophth.com www.stlouiscenterforclinicalresearch.com 3 Math 1530 - Chapter 4 Types of Bias in Sample Surveys 4.bp.blogspot.com Key Parts of a Sample Survey Bias: Favoring parts of population 1. Sampling Bias: from sampling method (e.g., nonrandom samples) 2. Nonresponse bias: some subjects cannot be reached or decline 3. Response bias: subject gives incorrect response or question is misleading Identify entire population 2. Construct sampling frame 3. Random sampling design to get n subjects from frame 4. Avoid sampling bias 1. Cross-stitch Sampler www.vam.ac.uk Learning Objectives: 1. 2. 3. 4. 4.3 What are Good and Poor Ways to Experiment? 5. 6. Elements of an Experiment Experimental units: Subjects Treatment: Conditions imposed on subjects Explanatory variable: Defines groups and treatments Response variable: Outcome Identify elements of experiment Experiments 3 components of good experiments Blinding the study Statistical Significance Generalizing results of the study reactorfire.files.wordpress.com Experiments Impose treatments on subjects to observe responses Goal: compare effects of treatments on response Randomized experiments – subjects randomly assigned to treatments www.clinplus.com www.privatemdlabs.com 4 Math 1530 - Chapter 4 Placebo effect 3 Components of a Good Experiment Placebo – fake treatment; sugar pill Control or Comparison Group Randomization Replication Placebo effect – improving not from real treatment but from belief that he or she should improve pages.cs.wisc.edu www.scientificamerican.com Principle 1: Control or Comparison Group Control or Comparison Group Helps analyze effectiveness of primary treatment Placebo removes lurking variables Control group gets placebo Clinical trials may compare new treatment with existing Experiments should compare treatments rather than effect of single treatment Example: 400 volunteers asked to quit smoking with some taking and some not taking antidepressant www.redicecreations.com www.gpaulmckinneyjr.com Principle 2: Randomization 1. 2. Eliminates bias from researcher assigning subjects Balances groups on known and lurking variables Principle 3: Replication 1. 2. Reduces difference due to ordinary variation or chance Increases chance that results show true difference www.printercopierblog.com colorectal.surgery.ucsf.edu 5 Math 1530 - Chapter 4 Blinding the Experiment Statistical Significance Blind – subjects unaware of which treatment used Double-Blind Experiment - Neither subjects nor investigators know which treatment Controls bias from respondent and experimenter Statistically Significant Difference – Observed difference is larger than expected from chance gridskipper.com Generalizing Results Goal of Experimentation – Analyze association between treatment and response for entire population Generalize only to population represented by study 4.4 Other Ways to Conduct Experimental and Observational Studies static.howstuffworks.com Learning Objectives Sample Surveys: Random Sampling Designs Understand basic principles of the following: 1. Sample Surveys 2. Two Types of Observational Studies 3. Multifactor Experiment 4. Matched Pairs Design 5. Randomized Block Design Alternative to experiments 1. Simple Random Sampling 2. Cluster Random Sampling 3. Stratified Random Sampling codetechnology.files.wordpress.com www.greatstuffforkids.com 6 Math 1530 - Chapter 4 Cluster Random Sample 1. 2. 3. Divide population into large number of clusters, such as city blocks Select simple random sample of clusters Use all subjects in clusters as sample Cluster Random Sample Advantages Sampling frame unavailable Cost Disadvantage Need larger sample size for same reliability reactorfire.files.wordpress.com Stratified Random Sample Stratified Random Sample www.smh.com.au 1. 2. 3. Divide the population into groups, strata Select SRS from each strata Combine samples from each for total sample Advantage Ensures stratum representation Disadvantage Need sampling frame and to which stratum each subject belongs www.nedarc.org Comparing Random Sampling Methods Types of Observational Studies Sample Survey: current 2. Retrospective Study: past 3. Prospective Study: future Cause not proven, but studies can support beliefs 1. Back to the Future youtube.com/watch?v=c8Rngy90Q14 7 Math 1530 - Chapter 4 Retrospective Case-Control Study Prospective Case-Control Study Retrospective Prospective Studying sunlight exposure and multiple sclerosis connection… Cases – have MS Controls or don’t Explanatory variable – low sun or not www.lisisoft.com Studying effects of vegetarian diet on heart disease… Cases – have heart disease Controls or don’t Explanatory variable – vegetarian or not www.dubaiinternetmarketing.com Multifactor Experiments Single experiment analyzes two or more factors Learn more since combinations may affect response Matched Pairs Design biobreak.files.wordpress.com Subjects are somehow matched Husband/wife, two plots in same field, etc. Same individual – crossover design Randomly assign or randomize order of treatments Reduces effects of lurking variables www.marshallhammondwedding.info Randomized Block Design Block – subjects with common characteristics Randomized Block Design, RBD – within each block, randomly assign to treatments Image Sources Statistics: The Art and Science of Learning from Data, 2nd Edition, Agresti and Franklin http://www.consumersenergy.com/uploadedImages/CEWEB/YOUR_BUSI NESS/lab-tech-with-beakers.jpg http://creativejuice.files.wordpress.com/2007/10/voyeur.jpg http://www.law.georgetown.edu/ist/ISTalk/2009/pablo.jpg http://www.austinchronicle.com/binary/c506/hypo.needle.jpg http://www.privatemdlabs.com/newsimages/Medicine_2248_19541486 _0_0_7037531_300.jpg http://www.lisisoft.com/imglisi/5/Screensavers/56804sunset.jpg http://artstuff.typepad.com/.a/6a00e54ff1492b88340120a59ac8fb9 70b-800wi http://rochesterchiro.files.wordpress.com/2008/06/neworleans.jpg http://www.dubaiinternetmarketing.com/wpcontent/uploads/2009/12/Vegetarian-Restaurants-Dubai.png http://reactorfire.files.wordpress.com/2009/11/20090617155350_stan ford2.jpg http://www.nedarc.org/nedarc/collectingData/participantSelection/pro babilitySampling/images/stratifiedSample.gif http://www.youtube.com/watch?v=c8Rngy90Q14 http://www.psdgraphics.com/wp-content/uploads/2009/05/bullseye.jpg http://biobreak.files.wordpress.com/2009/04/wondertwins.jpg http://www.marshallhammondwedding.info/contentimage/bride-andgroom-portrait.jpg http://s3.amazonaws.com/picable/2009/05/25/981127_Drug-abuseby-teens_620.jpg http://www.nedarc.org/nedarc/collectingData/participantSelection/ima ges/subsetPopulation.gif http://www.purbeck.gov.uk/images/bird%20watching.jpg http://www.youthnoise.com/Assets/Image/Gill%27s%20images/coin_fli p.jpg http://hoppingintopuddles.files.wordpress.com/2008/03/randomnumbers.jpg http://i.telegraph.co.uk/telegraph/multimedia/archive/01630/electionD ay_1630464c.jpg http://www.revophth.com/publish/images/1_14244_0.gif http://www.stlouiscenterforclinicalresearch.com/slccrtest_files/volunteer_g roup.jpg http://4.bp.blogspot.com/_BSwhxCnDXs/S7roA8X0fKI/AAAAAAAAAvE/MQFz6vSJ9pg/s1600/BiasLogo White.jpg http://www.smh.com.au/ffximage/2006/09/22/sydneystudents_widewe b__470x317,0.jpg http://codetechnology.files.wordpress.com/2008/06/survey-form.jpg http://www.greatstuffforkids.com/files/2235105/uploaded/76020ABCBlocks.jpg http://static.howstuffworks.com/gif/population-six-billion-1.jpg http://fajr.files.wordpress.com/2009/01/scales.jpeg http://gridskipper.com/assets/resources/2006/11/blindfolded.jpg http://www.printercopierblog.com/wpcontent/uploads/2008/09/office-copier.jpg http://colorectal.surgery.ucsf.edu/media/1160745/randomization.gif http://www.scientificamerican.com/media/inline/placebo-effect-a-curein-the-mind_1.jpg http://www.redicecreations.com/ul_img/6999placebo_big.jpg http://www.gpaulmckinneyjr.com/Files/images/quit_smoking.gif http://www.clinplus.com/Portals/53286/images/iStock_000005903700 XSmall-resized-600.jpg 8