372596014

II

liur Delphi Mady ul Ułih N'ur»*tił Amu-iila* th. i-ais Lleiii:nUof ifiiuitif Cie> ta Ih-UimI

vxfKB$*i«a"*pcct»lłf$f in thc !tcUl (Wodnwn I9SV>, In a siudy oi cyidcfwe bzmi nurswg pr«»*m«. f rcnch <ei al l 2002} <ucci mity dehncd cKpert sUt&»4« thc j<iotip test plated in th* orpanisaJkin to rcpott ón tlie mbjert' (p. 26?). Ihu defimitan Undudes ihe idooUlkatłfin ul thc »e*pun«k*nis *vhu *v<*uUI be wtilłng »nd able tu re^łutid to ih« widu* roumk <ńiheiumy. aawdl as. łi4Ving litr rci,u)r<\lcspwkiKc lit thi finał ruumi oi thc l>dplwwirvcv, fhccxpcn datur, of the respondent;? vnw«l js-icturał to thc v*Jtdnv of thc coostomuathtcml (<»rccn et .U I9vy.)

2.5 Sarapling

Ihr wrmplr fją# Iot raeh nwnd ul tlić I ktlphi mit w.y appcnrs U) bc 3ii btcsaęi scicncr. wjth repom ol pand «W* rangjng front ftve Iftccch I99V) lo l .665 (lawrcnccetal 1983), AMhoogh łl ts gcnerally agiced that i ll/gcc satnpfc iłxe proJuc n Iwftrr tt&ulK tiw teggestłón tliai having marę reapauttents i* bet Iw lor MAchmga coTłseitMii uppCAt* tu Hf Kił bont nnpirk.il tnimdation (hluiphy cl al 1998; Grcen <ft *1 1999; Maison et *1 2000; fWell 2003> In ądditmn. thc usenf proUibilin <-r n»n fmthdbittfy lampie*3ppeai530 te-ft dchatdhfc thc majonty ot Mudie* rcYiewed used non'random purposrve wmpiing t miuc* (Imich ctal ?W2>. The nseof non probabillty tiliripfiltg may rcsłift In bias due to tćmplmg enror as. thc csptirtsdw*en Uu not dursen) may nr tniiy W bc ftfprcscutulhte ul the specjalna łuuwledge reqaired (McKiim.m and    2000), Ai»>tlu r major

profilem etteouMcred In twmjl thc Delphi k\:hntqi«:i)i thc lnv> ofsuhjectt łrran c5ńh rouiul ol ihru ,Mirvt‘y; thc guAtcr thc mm)ln.-i of runmts. tltt n»wc HMy that ropojwients >vtll bc Iojj, rcsiilnng m a non-responsc hi«s,

IhcgtiUtdmes areliimted tegardtnghow much attnuon a acctptsblc front thc Drsl to ihc last rouiut. ht>w«.*ver.

It ha-* b* en RiigprWd ihat respouse rate^ frmn Ctteh roumJ o( lite study should »cmaiłi it 70 pcrccnt (H^uiinn ct al 2000) thł* tr<fuirr& «ha< mdht»d< lt«sh ai IUllman'* (200<)) 'Total I h^lgn Mctlmd' bc uvcd u> cnwuc high returny ot MłTYoyjc ttattt «tdl Of thc mumii.

2.6 Consensus and Data Analysis

The data colkctcd may mjłłirt qua!itativc or i{Uimtitati^analj !>^. dej>cndmx on włtdltcr a clawk.il nr modifird I >clpht tcvhnique \\ umuI In thc ursr nuind of thc *:hvwi ul Ddphi masy, optin-t odcd queffiom aur iwcd, whidi can be aoalyęcd w«ngcont^nt analymn 11 tmon ti al 2000) Ihc ttu&raiuee prov(dcs tew guidelinw for mea$urtng conienui*nr tor ditermining when ćonsemus U readicd MrKcnna t l‘W4) usrit 51 pcixcnt .ij-ieement amung rc«|>oiidatU M an irulicatw ot con^ensiw, brmwr, this ‘thn-duild ha* hccn iriticw-J tur heing k» than »k‘ci4ivc{<»r«cn ct al I^W). A Ddphi vtinł> noinnniug ihe mtormatinn ic^uirrnients *•» <ilNuxctl rrtpunw raić cntcnalrom thc socwl Ntwnccs and m1 tousenvsiv ar m<) ptreenr sgrccmcm/disagrecmenr w uh Matemcms lit cuch roumi of thc Mirvtjr (Grccn ct al I999). MtwJ of tlw rcscarcli findmgs report a ttciiiiiikm of comeńsiiJr Uiat t» LtauiilU arbitran and p«wt hoc (WUilaios and Wlcbb 1994). Ib u rcvicw of quanUtative «pplkdfion» to thc Delphi technii^te. Wuudenbcrg (ł9»ł|)Mtggcste*l that conscłiMU ivarhtcved mauily by gmup ptmutc »«> ajnftinn and thof ihh jv mcdian-d byth# ^afusitcaJ gunrp rc^punse* that arc pvcn as leedback te uli rerpondtcuii. Kmally, an additional method for cuyunng consensus »Kubmtumg rhe jftnal rcsults (8flerallonaly«s has beeo to łhr paneli lik toc I treli agytrcmrnt on Ihc rcscarcher* eorttfcrmw .słatfmenls In ihc hnal analysis. it W Oh* parte!* pcripfćtłve uf amsetMua that i* aóugjhil. nut that of tlić n ‘carchcr

jKiłwccjuent rouncU of <hc Delphi mt analyscd uung.a varinv oi dmnpthrcsianstics, uwaJJy rriaimp jo meaaures ot central tcruleney (mran. medutil and rncasurt» I v«rtahHltv (standard tlcYtotton. rangt. Inter *{iurtllc range), Measurcs of central tcrulency ilim* f«>r * Mngle tmraum of tlić groupś lUAkiniu. whcrcai. mcałtuT* of >atiability (cn|Ht\ lali* thc *ttndiinJ dcsutwin) display thc !rvdof agrcemcnt nr «Iifc*j(rccnn*nt that occtni eround a rnem scorc < McKams 190*1, ( incatore* an<l pextei 2000. Masson ct al 2000 Hfuuiuhcid jud Iłumphnł 2tx>l), fitrjtotct and Da^cr J2000) usnl a combuncinn of itcm grapfiR* ronntain geaplis and th^edtary grapb.' lolmck tbo«pininna.uid lrvrUnt śgnrrment ovn r*:h Migcoftllc Ddf4d »urrpy ll isargued Ilia! thk approadt «mviłkr¥ *vhut happen* bclwren thc I3dpł»l «jt*vy nmiutv thcrrfonr, H oUotci litr re«Mnlicr to uudcr.umd Ihepwwt that nccnr» m ca»-h roumi pnor to thc mprmdent* or eepert* frnally rcochmg cwu-rwis,

RCS1 LIRRARY