7/29/2019 sem12ersh8750_lecture12
1/70
StructuralEquationModels:
PathAnalysiswithLatentVariables
IntroductiontoStructuralEquationModeling
Lecture#12(andlast)
April25,2012
7/29/2019 sem12ersh8750_lecture12
2/70
TodaysClass
Puttingitalltogether:
PathAnalysis
Observedvariables
ConfirmatoryFactorAnalysis/MeasurementModels
Latentvariables
Concernsinbuildingstructuralequationmodels Modelpredictedcovariancematricesforpathanalysiswith
observedandlatentvariables
ExamplesofSEMuses
7/29/2019 sem12ersh8750_lecture12
3/70
UNDERLYINGTHEORYOFSTRUCTURALEQUATIONMODELS
7/29/2019 sem12ersh8750_lecture12
4/70
StructuralEquationModels
AlthoughthetermSEMcanbeappliedtomanysettings,I
viewthelabelasbeingusedtodescribeanalyseswith
observedandlatentvariables
Astructuralequationmodelconsistsoftwoparts:
Measurementmodel(s)foreachlatentvariable Pathanalysisbetweenthelatentandobservedvariables
Uptothispoint,wehavecoveredbothinisolation todayweputthemtogethertoshowhowtheprocessworks
Youwillseethisextrastepisprettystraightforward
7/29/2019 sem12ersh8750_lecture12
5/70
REVIEWOFPATHANALYSIS
7/29/2019 sem12ersh8750_lecture12
6/70
TypesofVariablesintheAnalysis
AnimportantdistinctioninpathanalysisandSEMis
betweenendogenousandexogenousvariables
Endogenousvariable(s): variableswhosevariabilityisexplainedbyoneormorevariablesinamodel
Inlinearregression,thedependentvariableistheonlyendogenousvariableinananalysis
Exogenousvariable(s):variableswhosevariabilityisnotexplainedbyanyvariablesinamodel
Inlinearregression,theindependentvariable(s)arethe
exogenousvariablesintheanalysis
ERSH8750:Lecture#4 6
7/29/2019 sem12ersh8750_lecture12
7/70
DirectandIndirectEffectsofHSLonMSE
ERSH8750:Lecture#4 7
Mathematics
SelfEfficacy
CollegeMath
Experience
HighSchool
Math
Experience
DirectEffect
Residual(Endogenous)
Variance
ExogenousVariances
ExogenousCovariances
4.1590.463
0.696
0.258
0.363
0.105
33.7972.712
103.1288.508
1.7430.140
NotShownOnPathDiagram:
6.9041.304
49.3132.412
4.9120.074
IndirectEffect
7/29/2019 sem12ersh8750_lecture12
8/70
PathAnalysisinMatrixForm
Ourpathmodelsimultaneousequationswere:
2 endogenousvariables 1 exogenousvariable
Alternatively,wecouldrephrasethisinmatrixform:
Where:
(matrixofsize1 containingobservedexogenousvariables)
(matrixofsize1 containingobservedendogenousvariables)
Then:
(matrixofsize1 containinginterceptsforendogenousvariables)
0 0
0
(amatrixofcoefficientsrelatingtheendogenousvariablestothemselves)
(matrixofsize relatingexogenousvariablestoendogenousvariable(s))
, (where isthe residualcovariancematrix)
Here, willbediagonal(nocovariance)aswedonothaveanymoredegreesoffreedom
ERSH8750:Lecture#4 8
7/29/2019 sem12ersh8750_lecture12
9/70
PathAnalysisinMatrixForm
Theequationsfromthepreviousslidearecalledthestructuralform ofthepathmodel
Anotherformthatexistsinliteratureisthereducedform,whereallendogenousvariablesareonthelefthandside
Where
Thereducedformisnotasfrequentlyusedinpractice,butdoesariseinsomeresearchareasandinidentification
ERSH8750:Lecture#4 9
7/29/2019 sem12ersh8750_lecture12
10/70
PathAnalysiswithMatrices
Althoughnotexplainedbyourmodel,wecouldstatethat
themeanvectorofexogenousvariableswas:
Likewise,wecanstatethatthecovariancematrixofthe
exogenousvariablesis
Wewillusethesetermsinourmatrixversionofthemodelpredictedmeanandcovariancematrix
ERSH8750:Lecture#4 10
7/29/2019 sem12ersh8750_lecture12
11/70
ModelPredictedMeanVectorand
CovarianceMatrix
Theunconditionalmeanoftheendogenousvariablesis:
Thecovariancematrixoftheexogenousandendogenousvariablesisthen:
,
Thepoint:thatmodelspecificationshavedirectimplicationsfortheparametersofthemultivariatenormaldistribution
ERSH8750:Lecture#4 11
7/29/2019 sem12ersh8750_lecture12
12/70
MatchingMatriceswithResults
Tomorespecificallylinkourresultstothematricesfrom
thepreviouspage:
ERSH8750:Lecture#4 12
Name Matrix ModelEstimates
ResidualCovariance Matrix 33.797 0
0 103.128
Regression WeightsofExogenousonto
Endogenous
0.696
4.159
CovarianceMatrixofExogenous Variables 1.743
MeanVectorof ExogenousVariables 4.912
VectorofEndogenous VariableIntercepts 6.904
49.313
MatrixofEndogenousRegression Weights 0 0
0.363 0
Inverse matrixusedincalculations 1 0
0.363 1
7/29/2019 sem12ersh8750_lecture12
13/70
ModelPredictedMeanVectorand
CovarianceMatrix
Theestimatedconditionalmeanoftheendogenousvariablesis:
Thesevaluescorrespondexactly(saturatedmodel)
Theestimatedcovariancematrixoftheexogenousand
endogenousvariablesis:
Thesearemostlyexact smalldifferences
ERSH8750:Lecture#4 13
7/29/2019 sem12ersh8750_lecture12
14/70
REVIEWOFCONFIRMATORY
FACTORANALYSIS
7/29/2019 sem12ersh8750_lecture12
15/70
OneFactorModelofFiveGRIItems
TheCFAmodelforthefiveGRIitems:
Here: responseofsubject onitem interceptofitem (listedasameanasthisistypicallywhatitbecomes) factorloadingofitem onfactor1(onlyonefactortoday) latentfactorscoreforsubject (sameforallitems)tofactor1(onlyonetoday) regressionlikeresidualforsubject onitem
Weassume 0, ;
iscalledtheuniquevariance ofitem Wealsoassume and areindependent
Also,wewillassume ,
Typically 0 (butnotalways) Factorvariancecanbeestimatedorfixed(moreonbothinidentification)
ERSH8750:Lecture#5 15
7/29/2019 sem12ersh8750_lecture12
16/70
OurCFAModelPathDiagram
MeasurementModel:
s=factorloadings
es=errorvariances
s=itemintercepts
StructuralModel:
=factorvariance
F1 =factormean
(Someofthesevalueswillhavetobe
restrictedforthemodeltobeidentified)
ERSH8750:Lecture#5
e1 e2 e3 e4 e5
Y4 Y5Y3Y2Y1
F1
1121 31 41
51
1
2
3
4 5
F1
1
16
7/29/2019 sem12ersh8750_lecture12
17/70
ModelPredictedMeanVector
Combiningacrossallitems,themeanvectorfortheitemsisgivenby:
ERSH8750:Lecture#5 17
7/29/2019 sem12ersh8750_lecture12
18/70
ModelImpliedCovarianceMatrix
Combiningacrossallitems,thecovariancematrixfortheitemsisgivenby:
Getusedtoseeingthis althoughyoualreadyhave(seetheregression
slides)
, , , ,
, , , ,
, , , ,
, , , ,
, , , ,
0 0 0 0
0 0 0 0
0 0 0 0
0 0 0 0
0 0 0 0
ERSH8750:Lecture#5 18
7/29/2019 sem12ersh8750_lecture12
19/70
PUTTINGITTOGETHER:PATH
ANALYSISWITHLATENTVARIABLES
7/29/2019 sem12ersh8750_lecture12
20/70
ASmallSEMExample
TodemonstratehowSEMworks,wewilluseaverysmall
example:
Measurementmodel:threeGRIitemsformingonelatent
construct(gambling)
Note:withthreeitems,themeasurementmodelisjustidentified
(meaningperfectfit)
Pathmodel:ThepredictionofgamblingbytheSOGscore
Note:herewetreatSOGSscoreasbeingobservedwithouterror
GRI1
GRI3
GRI5
Gambling SOGS
7/29/2019 sem12ersh8750_lecture12
21/70
Step#1:BuildingtheMeasurementModel
Thefirststepinastructuralequationmodelistobuildthe
measurementmodel
Here,themeasurementmodelissimplifiedsoastoshowhow
SEMworks
7/29/2019 sem12ersh8750_lecture12
22/70
MeasurementModelFitAssessment
Ourthreeitemmeasurementmodelfitsperfectly
7/29/2019 sem12ersh8750_lecture12
23/70
MeasurementModelParameterEstimates
7/29/2019 sem12ersh8750_lecture12
24/70
MeasurementModel:
ImpliedCovarianceMatrix
Themeasurementmodelimpliedcovariancematrixis:
1.000
0.7260.996
0.407 1.000 0.726 0.996
0.648 0 0
0 0.535 0
0 0 0.546
1.055 0.295 0.405
0.295 0.749 0.294
0.405 0.297 0.950
7/29/2019 sem12ersh8750_lecture12
25/70
Step2:Estimatingthe
StructuralEquationModel
Oncethemeasurementmodelisfoundtofit,thenext
stepistoestimatethefullstructuralequationmodel
SOGSsum istreatedasanexogenousvariable
Alsocalledanindependentvariable
GAMBLING(andtheitemsmeasuringit)aretreatedasendogenousvariables
Alsocalleddependentvariables
7/29/2019 sem12ersh8750_lecture12
26/70
SEM:ModelIdentification
AsSEMintegratesbothmeasurementandpathmodels,theidentificationrulesforSEMborrowfromboth
Themeasurementmodel(foralllatentvariables)mustbelocallyidentified Includingrulesforsettingscaleoflatentfactor(s)
Thepathmodelmustbeidentified
Anecessarybutnotsufficientwayofensuringidentificationisthetrule(countingrule)
Thenumberofparametersmustbelessthanthetotalnumberofmeans+
variances/covariancesofall observedvariablesintheanalysis
Numberofobservedvariablesinouranalysis:4 Numberofvariances/covariances:4*(4+1)/2=10 Numberofmeans:4 Total:14
Numberofparametersinouranalysis 2factorloadings+1factorvariance+3uniquevariances+1directeffect+
3itemintercepts+1exogenousvariance=12
7/29/2019 sem12ersh8750_lecture12
27/70
SEM:MplusSyntax
TheMplussyntaxisacombinationofpathand
measurementmodels
7/29/2019 sem12ersh8750_lecture12
28/70
SEM:ModelFitAssessment
Becausewehavefewerparametersthanthetotal
possible,wemustnowassessourmodelfit
7/29/2019 sem12ersh8750_lecture12
29/70
SEM:ModelParameterOutput
Note:ourmeasurementmodelparametershavechanged
slightly(moreonwhyinamoment)
7/29/2019 sem12ersh8750_lecture12
30/70
SEM:StandardizedModelParameters
Here,weseethattheSOGShasacorrelationof.577with
theGAMBLINGlatentvariable
7/29/2019 sem12ersh8750_lecture12
31/70
SEM:ModelImpliedCovarianceMatrices
Noticeourmodelimpliedcovariancematrix:
Andtheresidualsfromthesaturatedmodel:
NEWWRINKLES: MEASUREMENTMODELDOESNOTFITSATURATEDMODEL
PERFECTLY OFFDIAGONALOFBLOCKCANCAUSEMODELMISFIT
E ti F f
7/29/2019 sem12ersh8750_lecture12
32/70
EquationFormof
OverallStructuralEquationModel
Ourstructuralequationmodelsimultaneousequations
were,forthemeasurementportion:
And,forthestructuralportion:
3 endogenousvariables
1 exogenousvariable
7/29/2019 sem12ersh8750_lecture12
33/70
MatrixFormofStructuralEquationModel
Inmatrices,themeasurementportionofthemodelisgivenby:
with
Further,thestructuralportionofthemodelisgivenby:
with
Intermsofourmodel: Therearenodirectexogenouspredictorsofourendogenous
measurementmodelparameters(so )
Therearenodirectpredictorsofourendogenouslatentvariablesbyotherlatentvariables(so )
Westandardizedourfactormeantozero(so
Putting Values into Matrices:
7/29/2019 sem12ersh8750_lecture12
34/70
PuttingValuesintoMatrices:
MeasurementModel
itemintercepts
factorloadingsforendogenousvariables
uniquevariancesof
endogenousvariables
Putting Values into Matrices:
7/29/2019 sem12ersh8750_lecture12
35/70
PuttingValuesintoMatrices:
StructuralModel
directregressioncoefficientofexogenous
variableontoendogenousfactor
residualvarianceofendogenousfactor
variance/covariancematrixforthe
exogenousvariables
Model Predicted Mean Vector and
7/29/2019 sem12ersh8750_lecture12
36/70
ModelPredictedMeanVectorand
CovarianceMatrix
Thecovariancematrixoftheexogenousandendogenous
variablesisthen:
,
Thepoint:thestructuralequationmodelcanhave
significantmodelmisfitduetoboththemeasurementmodelandthestructuralmodel
ERSH8750:Lecture#4 36
7/29/2019 sem12ersh8750_lecture12
37/70
IssuesinBuildingStructuralEquationModels
BecauseofthemultiplewaysSEMscanexhibitmodelmisfit,theprocessofbuildingSEMscanbedifficult
Ingeneral,currentpracticestatesthatmeasurementmodelsshouldbebuiltfirst thenthefullSEM
Someresearchersofferquestionableadvice: Useonlyjustidentifiedmeasurementmodels
Why:fewerdegreesoffreedomwheremisfitcanhappen Badidea:poorreliabilityforlatentconstructs
BuildmeasurementmodelswithSEMssimultaneously Why:fullcalibrationcanleadtobetteroverallmodelfit Badidea:measurementshouldhappeninabsenceofexogenousvariables
UsetwostageanalysesforSEMs Why:measurementmodelthencannotchange Badidea:propagationofmeasurementerrorforsomefactorscoremethods
7/29/2019 sem12ersh8750_lecture12
38/70
SEMINPRACTICE:EXAMPLES
FROMREALWORLDANALYSES
7/29/2019 sem12ersh8750_lecture12
39/70
SEMinPractice
Todemonstratethepracticalsideofbuildingstructural
equationmodels,Iwillgooveracoupleexamplesfrom
realdataanalyses
Intheseexamples,themodelbuildingprocesswillbe
discussed,alongwithvaryingmethodsforanalysis
Thedatafortheseexamplesisnotavailable butthe
practiceshouldshowhowdecisionsaremadeabouthowSEMsareconstructedandinterpreted
7/29/2019 sem12ersh8750_lecture12
40/70
Example#1:EvaluationofAcademicProgress
Thisexamplecomesfromdatafromalargesoutheasternuniversity
Datainclude: PRE:scoresonapretestofmathematicsability,administered
tostudentswhentheyarriveattheuniversity Scoresarefromtotalnumbercorrect alphareliabilityof.81
POST:scoresonaposttestofmathematicsability(usingthesameitems),administeredtostudentsaftertwoyearsattheuniversity
Scoresarefromtotalnumbercorrect alphareliabilityof.81
CourseEnrollments: Ifastudenthadenrolledinoneof29coursesrelatedtomathand
scienceeducationattheuniversity Dataarebinary 0=didnotenroll;1=enrolled
l h i
7/29/2019 sem12ersh8750_lecture12
41/70
Example#1:ResearchQuestions
Theevaluationsoughttoanswerthefollowingquestions:
Didscoresimproveontheposttestwhencomparedwiththe
pretest?
Didcourseworksignificantlyaffecttheposttestscores?
Didthescoreonthepretestpredictthecourseworkstudents
took?
Didcourseworkmediatetherelationshipbetweenpretestand
posttest?
B ildi h SEM M d li I
7/29/2019 sem12ersh8750_lecture12
42/70
BuildingtheSEM:ModelingIssues
Becauseofthenatureofthedata,severalmodelingissuesmustbeconsideredwhenusingSEMtoanswertheresearchquestions
Becausepretestandposttestaresumscores(withaknown
reliability),eachcanbeusedasasingleindicator Inthiscase,theposttestsingleindicatorwillbeproblematicbecauseofthe
residualvariance(afterprediction)islessthantheoverallvariance Somustputsingleindicatormodelinlast
Eachofthecoursesisbinary(dichotomous),soincludingtheminthemodeldirectlyisnotanoption
Modelwouldtreatthemasnormallydistributedifnotcategorical Softwarewontallowcategoricalmediators
Couldusethemas: Countsforspecificcategories(thentreatcountasapproximatelynormal)
Whatwedid
Indicatorsofacourseworkfactor Hardtoenvision
M d li St t
7/29/2019 sem12ersh8750_lecture12
43/70
ModelingStrategy
Courses:
Createcountsofeachcoursecategory(3categoriestotal)
Treatcountsasapproximatelynormal(anduseMLR)
Useallvariablesinapathmodelwhere:
Pretestpredictscoursecountsandposttestscore Coursecountspredictposttestscore
Treatpretestandposttestassingleindicatorswherevariance
ofeachisweightedbythe.81reliabilityofeach Finalstepintheanalysis
I iti l S t F D i ti St ti ti
7/29/2019 sem12ersh8750_lecture12
44/70
InitialSyntax:ForDescriptiveStatistics
Initial Output Descriptive Statistics
7/29/2019 sem12ersh8750_lecture12
45/70
InitialOutput:DescriptiveStatistics
Model#1:PathModelw/o
7/29/2019 sem12ersh8750_lecture12
46/70
PosttestSingleIndicator
TheMplussyntax:
Modelfit:
Model #1: Relevant Output
7/29/2019 sem12ersh8750_lecture12
47/70
Model#1:RelevantOutput
Forbuildingasingleindicatoroutofposttest:
Model #2: Pre/Post Single Indicators
7/29/2019 sem12ersh8750_lecture12
48/70
Model#2:Pre/PostSingleIndicators
MplusSyntax:
Model #2: Model Fit Assessment
7/29/2019 sem12ersh8750_lecture12
49/70
Model#2:ModelFitAssessment
MplusOutput:
Normalizedresiduals:
Needforresidualcovariancesbetweencourseworksums
Model#3:SingleIndicatorswith
7/29/2019 sem12ersh8750_lecture12
50/70
ResidualCovariances
Mplussyntax:
Note:thismodelhasnodegreesoffreedomleft itis
justidentified Thereforemodelfitisperfect
Model #3: Results
7/29/2019 sem12ersh8750_lecture12
51/70
Model#3:Results
Model #3 Results
7/29/2019 sem12ersh8750_lecture12
52/70
Model#3Results
Model #3 Results
7/29/2019 sem12ersh8750_lecture12
53/70
Model#3Results
Model #3 Results
7/29/2019 sem12ersh8750_lecture12
54/70
Model#3Results
Model #3 Results
7/29/2019 sem12ersh8750_lecture12
55/70
Model#3Results
Example#1:ResearchQuestionsAnswered
7/29/2019 sem12ersh8750_lecture12
56/70
p Q
Theevaluationsoughttoanswerthefollowingquestions:
Didscoresimproveontheposttestwhencomparedwiththepretest?
Yes,posttestscoresimprovedby.654SDforeveryoneSDincreaseinthepretestscore(p
7/29/2019 sem12ersh8750_lecture12
57/70
p j
Thesecondexamplecomesfromarecentdissertation
titled:
RISKSANDBENEFITSOFSEEKINGANDRECEIVING
EMOTIONALSUPPORTDURINGTHEDIVORCEPROCESS:AN
EXAMINATIONOFDIVORCEEINDIVIDUALADJUSTMENT,
CLOSENESS,ANDRELATIONALSATISFACTIONWITH
MULTIPLEPARTNERSFROMASOCIALNETWORK
Thedissertationusedastructuralequationmodelto
investigatethepsychometricpropertiesofascaleand
theninvestigatedtheroleofthefactorsinpredicting
outcomesrelevanttotheresearch
Example#2:ResearchQuestions
7/29/2019 sem12ersh8750_lecture12
58/70
p Q
Thefollowinghypotheseswereexamined:H1:Initiatorstatusoffilingforthelegaldivorceisrelatedtohigheradjustmenttodivorce.
H2:Femaleshavehigheradjustmenttodivorce.
H3:Havingchildrenwiththeexspouseisrelatedtolowerlevels
ofadjustment.
H4:Thereisapositiverelationshipbetweenlengthoftimesincethedivorceandadjustmenttodivorce.
H5:Thereisanegativerelationshipbetweenageandadjustmenttodivorce.
7/29/2019 sem12ersh8750_lecture12
59/70
AnalysisofRiskinSeekingSupportScale
7/29/2019 sem12ersh8750_lecture12
60/70
y g pp
RiskinSeekingSupportScale:
6itemsinitially
Target exspouse:RMSEA=.12;CFI=.85
Target familymember:RMSEA=.06;CFI=.96 Target friend:RMSEA=.09;CFI=.91
Oneitemremoved: Target exspouse:RMSEA=.13;CFI=.87;Alpha=.69
Target familymember:RMSEA=.00;CFI=1.00;Alpha=.81
Target friend:RMSEA=.08;CFI=.95;Alpha=.75
AnalysisofEmotionalSupportProvision
Effectiveness Scale
7/29/2019 sem12ersh8750_lecture12
61/70
EffectivenessScale
EmotionalSupportProvisionEffectivenessScale
13items
Target exspouse:RMSEA=.12;CFI=.86
Target familymember:RMSEA=.11;CFI=.87 Target friend:RMSEA=.07;CFI=.92
Removed3reversecodeditems: 10items
Target exspouse:RMSEA=.09;CFI=.92;Alpha=.95
Target familymember:RMSEA=.11;CFI=.91;Alpha=.95
Target friend:RMSEA=.06;CFI=.96;Alpha=.92
AnalysisofClosenessScale
7/29/2019 sem12ersh8750_lecture12
62/70
ClosenessScale:
18items
Target exspouse:RMSEA=.11;CFI=.78
Target familymember:RMSEA=.09;CFI=.81 Target friend:RMSEA=.08;CFI=.86
Removed7reversecodeditems 11items
Target exspouse:RMSEA=.09;CFI=.88;Alpha=.89
Target familymember:RMSEA=.06;CFI=.95;Alpha=.90
Target friend:RMSEA=.06;CFI=.95;Alpha=.90
AnalysisofRelationalSatisfactionScale
7/29/2019 sem12ersh8750_lecture12
63/70
Relationalsatisfactionscale:
7items
Target exspouse:RMSEA=.07;CFI=.96;Alpha=.83
Target familymember:RMSEA=.05;CFI=.99;Alpha=.92 Target friend:RMSEA=.05;CFI=.98;Alpha=.88
InvarianceTesting
7/29/2019 sem12ersh8750_lecture12
64/70
Becausemanyofthescalesaskedaboutspecifictargets,
thesewereusedasmultiplegroupstoinspectfactorial
invariance
Modelspreviouslyreportedrepresentedtheconfigural models
RiskinSeekingSupportScale:
Metricinvarianceheld Partialscalarinvariance(3itemswithdifferentintercepts)
Partialresidualinvariance(2itemswithdifferentunique
variances)
InvarianceTesting
7/29/2019 sem12ersh8750_lecture12
65/70
Emotionalsupportprovisioneffectivenessscale Partialmetricinvariance(5itemsnotinvariant)
Partialscalarinvariance(5itemsnotinvariant)
Closenessscale
Partialmetricinvariance(1itemnotinvariant)
Partialscalarinvariance(5itemsnotinvariant)
RelationalSatisfactionScale
Partialmetricinvariance(3itemsnotinvariant) Partialscalarinvariance(3itemsnotinvariant)
SubsequentPathAnalysis
7/29/2019 sem12ersh8750_lecture12
66/70
FinalPathAnalysis
7/29/2019 sem12ersh8750_lecture12
67/70
7/29/2019 sem12ersh8750_lecture12
68/70
CONCLUDINGREMARKS
WrappingUp
7/29/2019 sem12ersh8750_lecture12
69/70
Todaywasaboutputtingitalltogether:pathanalysisand
measurementmodels
TheSEMframeworkallowsforpowerfulinferential
analysestobeconductedinastatisticallyrigorousmanner
Butwiththepowercomesalotoffrustration datadonot
alwayscooperate
Youwillfindthatpeopletakegreatlibertieswithhowthey
conductSEManalyses Ihopethisclasshelpsyouunderstandhowpeopledo
SEMandhowtodoSEMyourself
UpNext
7/29/2019 sem12ersh8750_lecture12
70/70
Todayisourlastclassmeeting Nofinal;Nofinalhomework
Extracreditopportunity: Filloutcoursesurveyonline
https://portal.coe.uga.edu/apps/authorize/
EnsureyougetanAinthecourseItakethecomments
veryseriously!