Use LEFT and RIGHT arrow keys to navigate between flashcards;
Use UP and DOWN arrow keys to flip the card;
H to show hint;
A reads text to speech;
48 Cards in this Set
- Front
- Back
Statistics & Research Design
Glossary Alpha (Level of Signifance) |
Alpha refers to the probabilty of rejecting the null hypothesis when it is true; i.e., the probablity of making a Type I error. The value of alpha is set by an exprimenter prior to collecting or analyzing th data. N psycholgcl rsrch, alpha is cmmnly set at ethr .01 or .05.
|
|
Statistics & Research Design
Glossary ANCOVA (Analysis of Covariance) |
A vrsn of th ANOVA usd 2 ncres th effcncy of th anlys by ststcly rmvng varblty n th DV tht is du 2 an xtranus varbl. Whn usng th ANCOVA, ea prsn's scor on th DV is adjustd on th bsis of prsn's scor on th xtranus varibl.
|
|
Statistics & Research Design
Glossary Autocorrelation |
A dsadvntg of th tim-seres nd othr w/in-sbjcts dsgns is tht th anlys of th dta cn b confnded by autocrrltn, wch ocurs whn sbjcts' prfrmnc on th post-tsts is lkly 2 crrlt w/thr prfrmnc on th pre-tsts. Autocrrltn cn nflat th valu of th nfrntl ststc (e.g., th t or F), thrby rsltng n an ncresd prbablty of a Type I error.
|
|
Statistics & Research Design
Glossary Between-Groups Designs |
Stdys n wch th effcts of th dffrnt lvls of one or mre IVs r cmprd by admnstrng ea lvl or cmbntn of lvls to a dffrnt grp of prtcpnts.
|
|
Statistics & Research Design
Glossary Central Limit Theorem |
Th theorm derivd from probablty thry tht prdcts tht th smplng dstrbtn of th mean (1) wil apprch a nrml shap as th smpl sz ncreses, rgrdlss of th shp of th popultn dstrbtn of scors; (2) has a mean = 2 th popultn means; (3) hs a std dvtn =2 th popultn std dv dvided by th squr root of th smpl sz.
|
|
Statistics & Research Design
Glossary Chi-Square Tests/Single-Sample and Multiple Sample |
Nfrntl ststcl tsts usd whn th dta 2 b anlyzd rprsnt a nomnl scal. Th sngl-smpl chi-squr tst is usd whn th stdy ncluds one varibl; th multi-smpl chi-squr tst whn it ncluds two or mre varibles.
|
|
Statistics & Research Design
Glossary Cluster Sampling |
N cntrst 2 othr forms of smpling (wch nvolv selctng prsns from the popultn), clustr smplng ntails selctng units or grps (clusters) of prsns frm th popultn (e.g., schools, hospitals, clinics.)
|
|
Statistics & Research Design
Glossary Correlation Coefficients/Pearson r, Point Biserial, Biserial, ETA |
Th crrltn cffcnts is a numrcl ndex of th r/s (dgre of assctn) btwn two or mre varbls. Th mgntud of th cffcnt ndcats th strngh of th r/s; its sign ndcats th drctn (pstv or ngtv). Th Pearson r is usd whn dta on both varibls reprsnt a cntnus scal; th point biserial whn one varible is a tru dichtmy and th othr is cntnus; th biserial whn one varible is an artficial dichtomy and th othr is cntnus; nd eta whn th varibl r cntnus but hv a nonlinear r/s.
|
|
Statistics & Research Design
Glossary Counterbalanced Design |
A rsrch dsgn usd 2 cntrl crryovr (ordr) effcts; nvolvs admnstrng th dffrnt lvls of th IV to dffrnt sbjcts or grps of sbjcts n a dffrnt ordr. Th Latin squr dsgn is a type of counterbalanced design.
|
|
Statistics & Research Design
Glossary Cross Validation/Shrinkage |
Rfrs 2 valdtng a crrltn cffcnt (e.g., a critrn-rltd valdty cffcnt) on a nu smpl. B/c sam chnc fctrs oprtng n an orignl smpl r not oprtng n th sbsqnt smpl, th crrltn cffcnt tnds 2 "shrnk" on crss-vldtn. N trms of th mltpl crrltn cffcnt (R), shrnkg is grtst whn th orignl smpl is smll nd th # of prdctrs is lrg.
|
|
Statistics & Research Design
Glossary Demand Characteristics |
Q's n th xprmntl situatn tht nfrm rsrch prtcpnts of hw thy r xpctd 2 bhv durng th cors of th stdy. Dmnd chrctrstcs cn thretn a stdy's ntrnl nd xtrnl vldty.
|
|
Statistics & Research Design
Glossary Discriminant Function Analysis |
Th mltivarit tchnq usd whn ther r 2 or mre cntnus prdctrs nd 1 dscret (nomnl) critrn. Rfrd 2 as mltpl dscrmnt fnctn anlys whn th critrn hs mre thn 2 ctgres.
|
|
Statistics & Research Design
Glossary Experimental Research/True and Quasi-Experimental |
Xprmntl rsrch nvolvs cndctng a stdy 2 tst hypthses abt th r/s btwn IVs nd DVs. A tru xprmntl stdy prmts grtr cntrl ovr th xprmntl situatn - its "hlmrk" is rndm assgnmnt 2 grps. A qusi-xprmntl prmts lss cntrl.
|
|
Statistics & Research Design
Glossary Experimentwise Error Rate |
Rfrs 2 th prblty of mkng Type I eror. As th # of ststcl cprsns n a stdy ncreses, th xprmntws eror rate als ncreses.
|
|
Statistics & Research Design
Glossary External Valdity |
Th dgre 2 wch a stdy's rslts cn b gnrlyzd 2 othr peop, setngs, cndtns, etc.
|
|
Statistics & Research Design
Glossary Factorial ANOVA |
Th typ of ANOVA usd whn a stdy ncluds 2 or mre IVs (i.e., whn th stdy hs usd a fctrl dsgn). Als rfrd 2 as a 2-way ANOVA, 3-way ANOVA, etc./ w/th wrds "two' nd "three" frng 2 th # of IVs.
|
|
Statistics & Research Design
Glossary Factorial Design/Main Effects and Interaction |
Th nam gvn 2 any rsrch dsgn tht ncluds 2 or mre "fctrs" (IVs). Fctrl dsgns prmt anlys of main nd ntrctn effcts. A main effct is th effct of a sngl IV on th Dv. An ntrctn ocurs whn th mpct of an IV dffrs @ dffrnt lvls of anthr IV.
|
|
Statistics & Research Design
Glossary Indpendent and Dependent Variables |
Th ndpndnt (xprmntl) vribl is mnpultd n a rsrch stdy 4 th prpos of dtrmng its effcts on th dpndnt varbl; th varbl tht is blevd 2 hv an effct on th DV. Ea IV n a stdy must alwys hv at lst 2 lvls. Th DV is obsrvd nd measrd n a stdy nd is blevd 2 b affctd by th IV.
|
|
Statistics & Research Design
Glossary Internal Validity - Maturation, History, Selection |
Th dgre 2 wch a rsrch stdy allws an nvstgtr 2 cnclud tht obsrvd variblty n a DV is du 2 th IV rthr thn 2 othr fctrs. Maturtn is 1 thret 2 ntrnl valdty. It ocurs whn a physcl or psychlgcl prcss or evnt ocurs as th rslt of passg of tim (e.g., ncresng fatig, dcresng motvtn) nd hs a systmtc effct on sbjct's status on th DV. Anthr potntl thret is hx, wch rfrs 2 an evnt tht is xtrnl 2 a rsrch stdy nd tht is nt rlvnt 2 th rsrch hypths but tht affcts sbjct's prfrmnc on th DV n a systmtc wy nd thrby cnfnds th rsalts of th stdy. Slctn thrtns ntrnl valdty whn prtcpnts n dffrnt tx grps r initlly dffrnt nd, ther4 wud dffr @ th end of th stdy evn if no tx hd bn appld. Slctn is a thrt whn prtcpnts r not rndmly assgnd 2 grps.
|
|
Statistics & Research Design
Glossary Interval Recording / Event Sampling |
Ntrvl rcrdng is a mthd of bhvrl smplng tht nvolvs dvidng a period of tim n2 dscret ntrvls nd rcrdng whthr th bhvr ocurs n ea ntrvl. It is prtculrly usfl whn 4 bhvrs tht hv no clr bginng or end. Event smplng is a mthd of bhvrl smplng tht is usfl 4 bhvrs tht r rare or tht lev a prmnt prdct. I nvolvs recrdng ea occurnc of a bhvr durng a predfnd or preslctd evnt.
|
|
Statistics & Research Design
Glossary LISREL |
A causl (strctral equtn) mdlng tchnq usd 2 vrfy a predfnd causl modl or thry. Mre cmplx thn path anlys; it allws two-wy (non-rcusv) paths nd tks n2 accnt obsrvd varbls, th latnt traits thy r blevd 2 meas, nd th effcts of msmnt eror.
|
|
Statistics & Research Design
Glossary MANOVA |
A form of th ANOVA usd whn a stdy ncluds 1 or mre IVs nd 2 or mre DVs, ea of wch is measrd on an ntrvl or rtio scal. Use of th MANOVA hlps rduc th xprmntws eror rate nd ncreses pwr by anlyzng th effcts of th IV on all DVs simultneusly.
|
|
Statistics & Research Design
Glossary Measures of Central Tendency Mean, Median, Mode |
Th mean is th arthmetc avrg of a set of scors. Th mean cn b usd whn scors r measrd on an ntrvl rtio scal. Th median is th mddle scor n a dstrbtn of scors whn scors hv bn ordrd from lwst 2 hghst. It is usd w/ordnl dta (nd w/ntrvl nd rtio dta whn th dstrbtn is skwd or cntains 1 or a fw outliers). Th mode is th valu (scor or ctgry) tht ocurs most frqntly n a dstrbtn of nomnl ctgres.
|
|
Statistics & Research Design
Glossary Mixed Designs |
Rsrch dsgns n wch both btwn-grps nd w/in sbjcts cmprsns cn b made.
|
|
Statistics & Research Design
Glossary Multiple Regression |
Th mltivrit tchnq usd 4 prdctng a scor on a cntnus critrn bsed on prfrmnc on 2 or mre cntnus nd/or dscret prdctrs. Prdctrs ncluded n a mltpl rgrssn equtn wil hv lo crrltns w/es othr nd hi crrltns w/th critrn. (Hi crrltns btwn prdctrs is rfrd 2 as "multicollinearity.") Th outpt of mltpl rgrssn is th mltpl crrltn cffcnt nd a mltpl rgrssn equtn.
|
|
Statistics & Research Design
Glossary Normal Curve - Areas Under the Normal Curve |
A symtrcl bel-shpd dstrbtn tht is dfnd by a spcfc mthmtcl frmula. Whn scors on a varbl r nrmly-dstrbtd, it is pssbl 2 cnclud tht a spcfc # of obsrvtns fall wthn crtn areas of tht dstrbtn tht r dfnd by th std dv: in a nrml dstrbtn, abt 68% of obsrvtns fal btwn th scors tht r + nd - one std dv from th mean, 95% btwn th scors tht r + nd - two std dv from th meas, nd 99% btwn th scors tht r + nd - three std dv from the mean.
|
|
Statistics & Research Design
Glossary Null and Alternative Hypotheses |
N xprmntl rsrch, th nvstgtr tsts a vrbl rsrch hypthses by simultneusly tstng 2 cmpetng ststcl hypthses. Th frst, th null hypthsis, is statd n a way tht mplies th th IV duz NOT hv an effct on th DV. Th scnd ststcl hypthsis, th altrntv hypthsis, states th oppsit of th null hypthsis. It is xprssd n a way tht mplies tht th IV duz hv an effct.
|
|
Statistics & Research Design
Glossary One-Way ANOVA / F-Ratio |
A parmtrc ststcl tst usd 2 cmpr th means of two or mre grps whn a stdy ncluds 1 IV ns 1 DV tht is measrd on an ntrvl or rtios scal. It is prfrbl 2 multpl t-tsts whn a stdy nvolvs mre thn three grps b/c it hlps cntrl th xprmntws eror rate. Th one-wy ANOVA ylds an F-Rtio tht ndcats if any grp means r sgnfcntly dffrnt. Th F-rtio rprsnts a meas of tx effcts + eror dvided by a meas of eror only (MSB/MSW). Whn th tx hs an effct, th F-ratio is lrgr thn 1.0.
|
|
Statistics & Research Design
Glossary Parametric and Nonparametric Tests |
Parmtrc tsts r nfrntl ststcl tsts tht r usd whn dta 2 b anlyzd rprsnt an ntrvl or rtio scal nd whn crtn asmptns abt th popultn dstrbtn(s) hv bn met: i.e., whn scors on th varibl of n trst r nrmly dstrbtd nd whn ther is homoscdstcity (popultn varnces r +). An advntg of th parmtrc tsts is thtb thy r mre "pwrfl" thn th nonpparmtrc tsts. Nonparmtrc tsts r nfrntl ststcl tsts usd whn th dta 2 b anlyzd rprsnt ethr an ordnl or nomnl scal or whn th asumptns 4 a paramtrc tst hv not bn met. Nclud th chi-squr tsts, th Mann-Whitney U, nd th Wilcoxon matchd-pairs tst.
|
|
Statistics & Research Design
Glossary Path Analysis |
A causal modlng tchnq usd 2 vrfy a predfnd causl mdl ot thry. Nvolvs trnsltng th thry n2 a path diagrm, cllctng dta, on th varibls of ntrst (th obsrvd varbls), nd calcultng nd ntrprtng path cffcnts.
|
|
Statistics & Research Design
Glossary Protocol Analysis |
Tchnq usd by cgntv psychlgst 2 ID th cgntns undrlyng prblm-slvng nd dcsn-mkng. Nvolvs hvng an ndvdul "thnk aloud" whl wrkng nd then anlyzng th rcrd (protocol) of th ndvdl's vrblztns.
|
|
Statistics & Research Design
Glossary Random Assignment |
Rfrs 2 a mthd of assgng sbjcts 2 tx grps usng a rndm mthd; smtms rfrd 2 as "rndmiztn." Cnsdrd th "hlmrk" of tru xprmntl rsrch b/c it nabls an nvstgtr 2 cnclud tht any obsrvd effct of an IV is du 2 th IV rthr thn 2 eror. (B crful not 2 cnfus rndm assgnmnt nd rndm slctn/smplng.)
|
|
Statistics & Research Design
Glossary Random Error |
Eror tht is unprdctbl (rndm). Smplng eror is a typ of rndm eror.
|
|
Statistics & Research Design
Glossary Randomized Block Factorial Anova |
A vrsn of th ANOVA tht is appropriat whn blocking has bn usd as a mthd 4 cntrllng an xtranus varibl. Allws an nvstgtr 2 ststclly anlyz th main and ntractn effcts of th xtraneus varibl (wch is bng tx'd as an addtnl IV).
|
|
Statistics & Research Design
Glossary Regression Analysis |
Aststcl tchnq usd 2 prdct a scor on a critrn bsed on th prsn's obtained scor on prdctr. Nvolvs th idntifctn of a regrssn line ("line of best fit") nd th us of th equtn 4 tht line, th rgrssn equtn.
|
|
Statistics & Research Design
Glossary Rejection and Retention Regions |
Th rjctn rgn of a smplng dstrbtn cntns thos smpl valus (e.g., means) tht r unlkly 2 b obtnd smply as th rslt of smplng eror. Whn an nfrntl ststcl tst ndicats tht th obtnd smpl valu fals in th rjctn rgn, th nul hypthsis is rjctd nd th altrntv hypthsis is rtained. Th sz of th rjctn rgn is dfnd by alpha. Th rtntn rgn is th rgn of a smplng dstrbtn tht cntns thos valus tht r lkly 2 b obtnd smply as th rslt of smplng eror. Whn an nfrntl ststcl tst ndicats tht an ontnd smpl valu is n th rtntn rgn, th nul hypthsis is rtand nd th altrntv hypthsis is rjctd. Th rtntn rgn is = to 1 - aplha.
|
|
Statistics & Research Design
Glossary Sampling Distribution of the Mean - Standard Error of the Mean |
Th dstrbtn of smpl means tht wud b obtnd if an nfinte # of =sz smpls wer rndmly slctd from th pop nd th mean of ea smpl calcultd. Th smplng dstrbtn is nrmly-shpd, its means is =2 th pop mean, nd tht its std dv (th std eror of th means) is =2 th pop std dvtn dvided by th squr root of th smpl sz. Usd n nfrntl ststcs 2 dtrmn hw lkly it is 2 obtn a prtclr smpl mean gvn th popultn mean, th popltn std dvtn, th smpl sz nd th lvl of sgnfcnc.
|
|
Statistics & Research Design
Glossary Scales of Measurement |
A mthd of ctgrznf th arus wys 2 measr varbls. Ther r four scals of msmnt tht dffr n trms of mthmtcl "sphstctn." From least 2 most sphstctd thy r nomnl,ordnl, ntrvl, nd rtio. A nomnl scal ylds "frqncy dta"; tht is, th frqncy of obsrvts n ea nomnl ctgry. Ordnl, ntrvl nd rtio yld scl valus of scors.
|
|
Statistics & Research Design
Glossary Shared Variability |
A crrltn cffcnt cn b squrd 2 obtn a measr of varblty n Y is shrd w/ (or is accntd 4 by) varblty n X.
|
|
Statistics & Research Design
Glossary Single-Subject Design / AB, Reversal, Mutliple Baseline |
Sngl-sbjct dsgns all ctn at lst one A (baseline) nd one B (tx) phz nd ncluds mltpl memnt of th DV @ rgulr ntrvls durng ea phz. Th AB dsgn ncluds a sngl bsln (A) phz nd a sngl Tx (B) phz. Th rvrsl dsgn ncluds, at a mnmum, two bsln phzes nd one tx (B) phz (e.g., an ABA or ABAB dsgn). Th tx is wthdrwn ("rvrsd") durng th scnd nd th sbsqnt bsln phzes. Use of th mltpl-bsln dsgn nvolvs seqntlly aplyng a tx 2 dffrnt "bslns" (e.g., to dffrnt bhvrs, sttngs, or sbjcts).
|
|
Statistics & Research Design
Glossary Skewed Distributions |
Asymtrcl dstrbtns n wch th mjrty of scors r loctd on 1 side of the dstrbtn. N a pstvly skwd dstrbtn, most scors r n th lo side but a fw scors r n th hi (pstv) side of th dstrbtn. N a ngtvly skwd dstrbtn, th mjrty of scors r n th hi side of thdstrbtn, but a fw r n th lo (ngtv) side. Th tail tels it.
|
|
Statistics & Research Design
Glossary Standard Deviation |
A meas of dsprsn (varblty) of scors arond th mean of th dstrbtn. Calcultd by dvidng th sum of th squrd dvtn scors by N (or N-1) nd tkng th squr root of th rslt. Th squr root of th variance.
|
|
Statistics & Research Design
Glossary Statistical Power |
Power rfrs 2 th prblty of rjctng a fals null hypthsis. Pwr cnnt b dirctly cntrlld but cn b ncresd by ncludng a lrg smpl, mxmzng th effcts of th IV, ncresng th sz alpha, nd rducng eror.
|
|
Statistics & Research Design
Glossary Systematic Error / Extraneous Varibles |
Systmtc eror is prdctbl eror. Xtranus (cnfndng) varbls r a sorc of systmtc eror.
|
|
Statistics & Research Design
Glossary T-Test/Singls Sample, Independent Samples, Correlated Samples |
Paramtrc tst usd 2 cmpr 2 means. th sngl-smpl t-tst is usd 2 cmpr a sngl obtnd smpl mean 2 a knwn or hypthszd pop mean. Th t-tst 4 ndpndnt smpls is usd 2 cmpr means from 2 ndpndnt smpls. Th t-tst 4 crrlts smpls is usd 2 cmpr 2 smpl means whn subjcts n th 2 grps r rltd n som wys (e.g., b/c thy wer mtchd on an xtrnus varbl or b/c a sngl-grp pretest/postest dsgn ws usd).
|
|
Statistics & Research Design
Glossary trend Analysis |
A typ of anlys of varinces (ANOVA) usd 2 asses linear nd nonlinear trnds whn th ndpndnt varibls is quntatv.
|
|
Statistics & Research Design
Glossary Type I & Type II Errors |
A typ I eror ocurs whn a tru nul hypths is rjctd. Th prblty of mkng a typ I eror is = 2 alpha. A typ II eror ocurs whn a fals nul hypths is rtnd. Th prblty of mkng a typ II eror is = 2 beta (wch is usuly unknwn).
|
|
Statistics & Research Design
Glossary Within-Subjects Designs |
An xprmntl dsgn n wch es prtcpnt recvs, at dffrnt tmes, ea lvl of th IV (or cmbntns of th IV) so tht cmprsns on th DV are mad w/in prtcpnts rthr thn btwn grps.
|