Min: 26 Min: 26 Q1: 1! Q1: "6 Med: 166 Med: 1#$ Q3: 236 Q3: 22 Ma%: ##6 Ma%: ##6 xx : 1$6&'#2 xx : 1"6&2!" s: 112&$' s: 126&3'" Row ID Weight 36 150 4 348 26 46 42 148 43 446 39 202 21 125 37 270 33 316 5 166 31 65 45 236 24 40 29 182 19 120 16 166 6 220 30 150 44 62 9 204 2 344 7 262 25 220 14 180 46 212 17 204 20 436 22 132 15 105 1 80 3 416 51 48 18 26 35 86 13 140 ROW ID WEIGHT 6 220 10 144 14 180 18 26 22 132 26 46 30 150 34 94 38 202 42 148 46 212 50 76 54 140 4 348 8 360 12 34 16 166 20 436 24 40 28 116 32 356 36 150 40 365 44 62 48 64 52 29 3 416 11 332 19 120 27 154 35 86 43 446 51 48 5 166 15 105 Random Sample Set: Systematic Sample Set: #& T(e t)o (isto*rams di++er a lot more t(an I e%pected t(ey )ould& T(e random sample set (as ,ot( a (i*(er mean and median- and t(e (isto*ram clearly re+lects t(is )it( t(e values pea.in* at t(e ! t( class /202#'1& T(e sample set t(at )as systematically *at(ered- (o)ever- (as a lo)er mean and median )(ic( pulls t(e pea. o+ t(e *rap( several classes le+t /around 101#'1& T(e systematic samplin* also +ailed to *at(er any values ,et)een 2! and 2''- so t(ere is an o,vious *ap in t(e (isto*ram& T(e ,o%plots appear muc( more similar /alt(ou*( it is more di++icult to visually notice di++erence in spread )it( ,o%plots1& 2ne noticea,le di++erence in t(e s(ape is t(e +irst 3uartile o+ t(e random sample set- )(ic( is 1! versus t(e ,o%plot +or t(e systematic sample set )(ic( is "6& T(e ,o% itsel+ is s(i+ted over to t(e ri*(t +or t(e sample set- re+lectin* t(e lar*er values /and *reater mean4media1 o+ t(at sample set& Term Project Part # Individual 1& a& 5e can ,e '!6 certain t(at t(e con+idence interval /&16!- $11 actually does contain t(e true value o+ t(e population proportion p& ,& 5e can ,e '!6 certain t(at t(e con+idence interval /1#$&#'- 22!&'11 actually does contain t(e true value o+ 7& c& 5e can ,e '!6 certain t(at t(e con+idence interval /'!&$!- 16&331 actually does contain t(e true value o+ 8& 2& a& 9or t(e +irst random sample set- )e calculated a con+idence interval o+ /&16!- $11& Si%teen ,ears )ere actually measured in t(e mont( o+ Septem,er- so t(e proportion o+ ,ears t(at )ere measured in Septem,er is &2'63 /164!#1& T(us- t(e con+idence interval does contain t(e population parameter& T(e same +ollo)s +or t(e ne%t t)o sample sets: t(e mean )ei*(t o+ t(e population is 1$2&$'- )(ic( is contained in t(e con+idence interval t(at )e calculated o+ /1#$&#'- 22!&'11& Similarly- t(e true standard deviation o+ t(e population is 121&$- )(ic( is included in t(e con+idence interval t(at )e calculated o+ /'!&$!- 16&331& Simple Random Sample Graphs: Systematic Sample Graphs: 2& 2ur *roup collected t(e second data set t(rou*( a systematic samplin* o+ t(e ;data collection mont( o+ )ild ,ears< data set& 2ur met(od entailed ta.in* every +ourt( value +rom t(e ori*inal data set- startin* at a randomly c(osen point /t(e t(ird value in t(e ori*inal data set1& 5(en )e came to t(e end o+ t(e data set- cycled ,ac. to t(e start and continued ta.in* ne) values until )e (ad reac(ed 3!& 3& T(e pie c(arts appear very similar& T(e +re3uencies o+ eac( mont( +or ,ot( data sets never vary more t(an one di*it- and 3 o+ t(e " mont(s (ave t(e same +re3uencies& I t(in. t(at minor di++erences are muc( (arder to notice in pie c(arts- since it is more di++icult to compare )ed*es versus ,ars& T(e Pareto *rap(s ma.e t(e di++erences more o,vious since t(e *rap( (as inde%ed values on t(e y0a%is& =lt(ou*( t(e values in t(ese data sets are very similar- t(e di++erences are more o,vious in t(e Pareto *rap(s& #& 2ur results are very similar to t(e population& T(e main di++erence is t(at t(e si%t( mont(- >une- didn?t *et pulled in eit(er o+ our sample sets /)it( a +re3uency o+ only 1&$!6 in t(e population- it ma.es sense t(at it can easily ,e passed over in a samplin*1- resultin* in a Pareto *rap( and pie c(art t(at (ad one less ,ar /or )ed*e1 t(an t(e population& T(e overall s(apes o+ ,ot( t(e pie c(art and t(e Pareto *rap( are- (o)ever- very similar- and it seems as t(ou*( t(e sample sets are *ood representations o+ t(e population& Term Project Part ! 1& @ A &!: !6 si*ni+icance level 2& Bate*ory: mont( o+ data collection +or )ild ,ears& @C: p = &2'63: 2'&636 o+ ,ears )ere c(osen in Septem,er @C: p D &2'63: 2'&636 o+ ,ears )ere not c(osen in Septem,er A $&!## s A 1&$'## n A 3! xx @ A &31#3 E &2'63 &2'63 &"3"1 3! z A &1#6$ Fsin* t(e G ta,le- )e +ind t(at t(e P0value is &!!!"%2A1&111#& 5e multiply t(e ori*inal value ,y 2 since it is a t)o0tailed test& T(e P0value is *reater t(an t(e si*ni+icance level o+ &!- t(us )e +ail to reject t(e null (ypot(esis& 5e conclude t(at t(ere is not su++icient sample evidence to )arrant rejection o+ t(e ori*inal claim t(at 2'&636 o+ ,ears )ere c(osen in Septem,er& 3& @C: @ A 1$2&$$' @C: @ D 1$2&$$' A 1$6&'#2 s A 112&$' n A 3! xx @ A 1$6&'#2 E 1$2&$$' 112&$' 3! Fsin* tec(nolo*y- )e +ind t(at t(e P0value is 1&66#& T(is means t(at t(e P0value is *reater t(an t(e si*ni+icance level o+ &!& T(us- )e +ail to reject t(e null (ypot(esis& T(ere is not su++icient evidence to support t(e claim t(at t(e population mean is e3ual to 1$2&$$'& #& Hot( samples meet t(e conditions o+ t(ese tests& 9or I2- t(e sample is a simple random sample: t(e conditions +or a ,inomial distri,ution are met ,ecause eac( ,ear )as eit(er measured in Septem,er or it )asnJt: and ,ot( np and nq are *reater t(an !& 9or I3- t(e sample is a simple random sample: and t(e population is *reater t(an 3& Term Project Part 6 Re+lection =t t(e very least- t(is term project (as tau*(t me )(at in+ormation can ,e *at(ered +rom ,ot( a population and a sample- and (o) to compare4contrast t(at collected in+ormation& I (ave never ,een a mat(0oriented t(in.er and (ave c(osen a career pat( in social sciences ,ecause it ,etter suits my stren*t(sKnone o+ )(ic( are even closely related to mat( or num,ers& T(e social sciences do- (o)ever- tend to rely on statistical in+ormation in countless applications- and t(is class (as (elped me develop an understandin* o+ t(e met(ods used to produce and discuss suc( in+ormation& =s I pursue +urt(er education- t(e s.ills and concepts t(at I learned t(rou*( t(is term project /and t(e class in *eneral1 )ill allo) me to ta.e a more analytical approac( to in+ormation I encounter and see it in a more mat(ematical li*(t& It )as no accident t(at I only (ad to ta.e t)o mat( classes t(rou*(out my under*raduate studies: num,ers (ave never ,een a topic t(at interests me at all& =+ter I *raduated- t(ou*(- I too. a jo, at an or*aniLation t(at depends (eavily on statistics& 5(ile I my position contri,utes (eavily to t(e *at(erin* o+ t(e in+ormation t(at is later analyLed- I am rarely involved in t(e actual compilin* and ;num,er crunc(in*< side o+ t(e or*aniLation& =+ter ta.in* t(is class and studyin* t(e included concepts as in dept( as )e did- I +eel muc( more .no)led*a,le a,out ,ot( sides o+ t(e )or. my or*aniLation does and am more con+ident to ta.e part in t(e analysis portion& 2ne o+ t(e most practical applications o+ t(is class to me is not t(e actual proper and t(rorou*( statistical analyses- ,ut instead just ,ein* a,le to ma.e 3uic.- in+ormal conclusions a,out a sample or population& I .no) t(at suc( conclusions cannot al)ays ,e relied on +or accurate in+ormation- ,ut in my )or. I +ind mysel+ (avin* to ma.e 3uic. decisions a,out *roups o+ people ,ased on limit in+ormation& I am a case mana*er +or (omeless men- and recently- +or e%ample- I (ad to develop a ne) pro*ram ,ased on )(at I (ave *at(ered as t(e most prominent needs o+ t(e population )it( )(ic( I )or.& I *at(ered in+ormation t(rou*( o,servations and intervie)s and made conclusions a,out t(e *eneral population o+ (omeless men in Salt Ma.e Bity ,ased on t(e numerous intervie)s and o,servations I conducted& 9or suc( real0li+e situations- it is not al)ays possi,leKor necessaryKto de+ine actual statistical varia,les li.e mean and standard deviation /t(e mean num,er o+ men dealin* )it( unmedicated sc(iLop(renia is %- )it( a standard deviation o+ y&&&1 ,ut I am a,le to *et a *eneral idea o+ t(e prominence o+ certain issues and actually trust my conclusions a,out (o) many ot(er individuals mi*(t ,e dealin* )it( t(e same issues& =s I said ,e+ore- t(e s.ills and concepts I (ave learned in t(e class )ill undou,tedly (elp in +uture studies and in my current and +uture )or.& =lt(ou*( I sometimes stru**le )it( t(e mat(ematical side o+ statistics- t(e real )orld applications are e%tremely relevant to my interests and allo) me to understand t(e )or. I am doin* in muc( *reater detail&
Recent Developments in Clustering and Data Analysis: Développements Récents en Classification Automatique et Analyse des Données: Proceedings of the Japanese-French Scientific Seminar March 24–26, 1987