This article was downloaded by: [McMaster University] On: 14 February 2015, At: 15:44 Publisher: Routledge Informa Ltd Registered in England and Wales Registered Number: 1072954 Registered office: Mortimer House, 37-41 Mortimer Street, London W1T 3JH, UK

Multivariate Behavioral Research Publication details, including instructions for authors and subscription information: http://www.tandfonline.com/loi/hmbr20

Experimenting With Perceptual Change Strategies Douglas L. MacLachlan , Reza Moinpour & James M. McCullough Published online: 10 Jun 2010.

To cite this article: Douglas L. MacLachlan , Reza Moinpour & James M. McCullough (1977) Experimenting With Perceptual Change Strategies, Multivariate Behavioral Research, 12:4, 429-445, DOI: 10.1207/s15327906mbr1204_2 To link to this article: http://dx.doi.org/10.1207/s15327906mbr1204_2

PLEASE SCROLL DOWN FOR ARTICLE Taylor & Francis makes every effort to ensure the accuracy of all the information (the “Content”) contained in the publications on our platform. However, Taylor & Francis, our agents, and our licensors make no representations or warranties whatsoever as to the accuracy, completeness, or suitability for any purpose of the Content. Any opinions and views expressed in this publication are the opinions and views of the authors, and are not the views of or endorsed by Taylor & Francis. The accuracy of the Content should not be relied upon and should be independently verified with primary sources of information. Taylor and Francis shall not be liable for any losses, actions, claims, proceedings, demands, costs, expenses, damages, and other liabilities whatsoever or howsoever caused arising directly or indirectly in connection with, in relation to or arising out of the use of the Content.

Downloaded by [McMaster University] at 15:44 14 February 2015

This article may be used for research, teaching, and private study purposes. Any substantial or systematic reproduction, redistribution, reselling, loan, sub-licensing, systematic supply, or distribution in any form to anyone is expressly forbidden. Terms & Conditions of access and use can be found at http://www.tandfonline.com/page/terms-and-conditions

Xlfuitit*a7;nfeBeka.t.iorai Resenrck, 1977, 12, 429-4-5

EXPERl3IEKTIk'G WITH PERCEPTUAL GHAXGE STP,:ITEGIES

Downloaded by [McMaster University] at 15:44 14 February 2015

DOfJGEAP E. RTcLcL.?ICPILAX.REZX XOTKPOUII. University of W a s k i n m n

Konx~enicmuicidirneneiona1 scz+li:~gprocedcres are ernpioped in this espe1:rne:lta; eralaation of perceptual changes resdlnng f r o r differences in :he prese: tatio:. of ir::'om~ation about stirnuii. Lo:lg:tudinal 31DS techniques are fosntl to h e useful In monitorhg pe~ceptun: changes r~oducedb y v a d o - ~ s con~x--~iicatior~ ,r:l;ernatises. Commncicatior. s t ~ a t e g i e ssimcd at d x n g i n g the re:a~iorah'gs ,r:mor:g e t i ~ w l :s r c x n r c effective t h a i ~those i i e s i p c d cchzr.ge the perceptcal space structzre. F x t h e m o r e , pror~dingcomparktive izformstion aSo:,t groups of stimuli has more apparent impact than p r o ~ ~ d ~inr,g folx~stiona b o ~ at single stixulcs.

Sonmetric multidirnenslonal scaling (9IDS) methods have bee11 embracect entkusiastica:lg by researchers and practitioners in the field of rna~keticgbecame they provide ,z means of identifyxng the psycho:dsgicaliy competitive c?imensiorrs of markets while aiIoivrr~gcnstomers to choose the criteria pereeil-ecl to be reIer-ant i:? di~t~npnfshing among the product offerings. As explorato2';v devices, 51DS methods ka\-e served very adequateiy to describe markets. But, as Green (1975) has noted, except as crude diagnostic aids the :cchniques haye not offered much in the n-a:i of c l e a r cr:t strategic guidaxce. T ! ~ ~ o u g hp~oclrrct design znd promotion, m n r k e t i ~ gplanners nttemp",~ position their brand in the minds of potential czstomers so :hat i: 112s mzximarn oppor.trrr.lity for diffe~el~tial adra~ltage over competitive brancls. An initial step in the formulation of a pcsitioning strategy is to unclersrand ha\{- customers in the market perceire the brand relative t o all other brands in the product ciass. h v-ay to c'ao this is t o construct a product space TI-ith JiDS techniques which represents the customer^' perceptions of re1P.tl ve sirnihritiee among the brands, These rnethocfs are discrrssec! by Green slid Carmone (1970) and Green and Baa (1912),' If the I. F3r ax excellent sun-eq- of theoretical and mezhodolsgicd aspects of multidimensiar-a! scsling, see Sheparti, R o m n e ~Xerlo3-e , (1972 1.

QCTOBER, 1977

423

M. McCuliaugh brand occrrpies an u ~ f a r o r a b l eposition in the p ~ o d u c tspace, the p?amer must, seek wags to repositiom~it. Two major types of: bra& positioning strategies can be r'dectified once a pspehsisgical representatior, of a nza~ket(i.e., the perceived pr-oduct space) has been obtained. First, the planner can attempt to change the relative posr'tisns of brands in the prodtier space. Second, he can fry to alter the character om. number of the major dimel~sSsarsdefining the prodnet space. The former strategy can be termed a spatial chaq~gestrategy an6 tlze lcttter denoted a stmeh.ilraS change strategy. Either strategy could be implemented by modifying an existing brand in the space om. by changing cnatamers' perceptions of the characteristiex sf the existing brands thrmgkz eormzmnnimtion of new information. It should be possible to employ MDS techniques to monitor and experiment with shifts in spatial ar:d structural orr"entatiens of procluct spaces in 1cmgjtr;dinal studies, Most studies of attitude and perceptual change hare considered only unidirnensional changes QRiesEer,Collins, and Miler, 1969) a r have been based on factor analytic algorithms requiring prespecification sf reFes7ant attributes and rigorous measurement assumptions (Assael, 19'71; Vavm, 1972). MDS avoids ce priori clesignation of the aktributes underlying a product space and makes only very weak aissurrrptjosis about the data employed in cozstt.uetion of phgsica': representations of the space. Validation of procedures for longitridinal compar.ison of MDS output should permit monitsriag of perceptual shifts ausecl by purposefnl communiatioa strategies. Experimentation should precede impEementation of cornmurrication strategies in the marketplace to asswe that the antielpated pereep tual shifts will occur. There are difficulties invsived with Iong.E"tudina1measurement of perceptual changes with multidimensiond scaling (Eeeler. 1974; Jones and Young, 1972; I\&oinpour and MacLachlsm, 3973). The primary probJems stem from being unable, in some cases, to separate spatial from structural shifts, to assure proper dimensionality, and to provide a resonable standard against which to judge change. Ful*themore, since little is known about the sampling Ceistveribu5sn of MBS output, i t is not easy to assess the significance of changes observed.

Downloaded by [McMaster University] at 15:44 14 February 2015

Douglas k. Maciachiar;, Reza Mainpour and Sanres

This strrcig was designed to demonstrate that intent;iona;lIg created shifts in perceptcast spaces can be evaluated both visualIy sncl statistically. By carefully predetermining the salient percep430

MUkTlVARIATE BEHAVIORAL RESEARCH

L. MacLacRian, Rezz M o i n ~ o d rand James tvl h?cCuiiotlgh tual dimensions of a :+IDS-generated product space ai:d es7)e?i-

Downloaded by [McMaster University] at 15:44 14 February 2015

Doilg!as

rnezltinp with perceptaal-change strategies of each ki~:d fieparat-ely, it is poss1bIe to anticipate ckacges in the prorIuct space (i.e., flypothesbe the directions acd extent of efiznges) and secure e ~ - i decce of resulting changes, if xnk- (i.e., test the hypotheses). Before discrassing the experimental design used in the pre5e::t study: several general propositions ca:] be advanced. .!. dl~clticlirneazsio71nlscali?ig caq be '~tsejit o C O ? ~ i 3Pr r~~f f.t-,!(tt; F l y stable configz~rntiorsof 6r.nnds (proc?t,ct rpcrct7s/ re;?tescl f i n g t a ~ p e t?nnri;~tperceptions of byand p0sif10?i. The stcbllit:; !:ere refers both to stability over- time and across horncger:eo:~s snnq:les of the target market. Vario~zs studies of product classes t s v e turned up evidence of constancy regarding the snlient dimel~eions of product spaces in the absence of strong exrcrnal s:irnu:t,s azid the same is t m e of perceptuaI eonfiguratiorrs of br.a:?cls across samples of persons from the same markets. -4 cornHary is rkat it will be possible to identify the major perceptual c',ime~:siccs sf a pm'odact space, thereby providing the basis far ex;ler;rne~rtation ~ - i t hchange strategies. It is necessary for s:jch identificarion ko precede exper?rnentation so that a standa~dm-iK be a~ai!able r\gair,st15-hich to evaluate changes. 2. Spntinl chmzges in pi.oefucf. spaces ctre ~nz'iel 10 E,! feet : ~ ' i f h csmrrt~.z~~zr'catio stvmfegiss thnrr a r e et7.itcf.,c~ctl chat.gcs. Spatial

brand positions rely primarily 01:. experience n-ith tke brl;lzc!:s and hov-ledge about brand possession of salient attributes, Therefo:-e. these can 1 ~ ereadily infiuenced by ~ t t r i b r ~ It ne f o ~ ~ a t I othat n n-as not available before about the brznds. For mast; p!m?uer classes, or?, the other harld, the salient dimecsior:~comprising the strnctr~reof product spaces are based on rather ftrnda:rie~~talvalces regarding dimension importance des~eIopedthrozgh long esperier.,ce n-it5 sscia1 norms, reference grofip behavior, and cultural inlreritcr~:ces. One possible proeed~refor changing the ~~~~~~~~e of t i e p~od'.:ct space would be to corrrince people that all brands \%-erealike on some previously salfent c?irnecsio7-.This ~i-i-o:rlde l i ~ r n a t etke diner,sion as R saIien t discriminator. 3. 41~farma~tis~z. ccmtaaru.v~icated abozct a single bmzcd 2citE be Less eb~'ectit:e,in c r e n t k g spatial shifts than will I R farm4tion about more t1~n.i~ crre tsrn?td. The impact of Enfo~mationabout s single brand will depend partly on the product class arid koi\- sstartlicg is the revelation abouk the brand. It is assumect, hon-ever, that perceptions regarding brands are kekd in relation to one another; OCTOBER, 1937

431

Douglas b. Maciachlsn, Reza Mo~npeurand James M

McCuliough

Downloaded by [McMaster University] at 15:44 14 February 2015

hence esmpar.&tt::zleinfol-anatisn about two or more brands oaght to create a greater perceptiral shift than s i ~ p l yinformation about a s h g j e bral~d,

The prodnct category for the experiment was selected 6 t h considerable care so there vivouId be reasonable assul-an@@ that the dimensiondity of the product space would be simple and u7eU understood m i a advance of t h e experiment. It had been found in a namber of studies (Bass and Talarzyk, 1972; Bass and FF'iikie, '1973: Gohen and -.ahtoIa>1971: Wa'iey, 1968; Kakantshi and Beth man, 1974: She% and Talarzyk, 1972) th2t leading toothpaste brands were perceived on only a couple of major dimensions, the most salt'rlent of s ~ h i c hecsrtdd be Sooselg- labeled decay preverttion C associated with presence of fluoride) and whitening {associated with the physical qrxality of abrasiveness). Other dime~sionssuch as flavor, appmrance, and price were found to be iess important perceptual discriminators for the brands, Eight brands were selected thzt had relatkie'ly high brand awareness,

Subjects for the experiment were students in sections of the basic undergraduate marketing co~rrseat t h e Z'niversity of Washi n g % ~Two ~ . types sf control gronps were selected from two regnl a y s e c t i s ~ ~ofa the marketing course and other groups receicirmg randomly assigned experimental treatrien ta were cjnia sections of a large-ieeture version of the same course. There were thus eight g ~ * o c gof s subjects Err the experinlent: two control groups, three treatment groups for the spatial phase, and three treatment groups for the structura!. phase of the experiment.

The experiment co~sistedof obtaining cornpaz-ative perceptual jrrdgments (dissimilarities ratings) f o r the eight toothpaste brands after providing each group w-ith particular. insta-~ctionsand, in some cases, different inf errnative messages, Subjecia were asked to proride d i ~ s i ~ i i a r i tj y~ d g m e n t sfor all pdrs of brands on a scale af l-similar to 9-dissimilar. 432

MUhTlVARkATE BEHAVIORAL RESEARCH

Downloaded by [McMaster University] at 15:44 14 February 2015

Douglas L. Maclachlan, Weza Moinpour and James Ait, h4cCullougk

To establish that the anticipated dimensionali~yof the space eoulcl be obhined, a eontrol group (group 1) n-as designated to receive a standard set of instructions askizp subjects to employ only the criteria of decay prevention and abrasiveness in their judgment of brand dissimilarities (a]! treatment groups received this same set sf standar-8 instrr~ctionsdnriz~gt h e first week sf the experiment). A second control group i g ~ o u p2 ) a~--asg i ~ e n na instructions regarding the criteria to employ En making brand similarity jndgmen ts. Both con troI groaps recei~edunchanged instructions oser the three weeks in order to detect any modificai ions of the p ~ o d u c tspaces due to the passage of time (envir.o:lmenhB, testing and maturation effects), During the third week, group 2 rated a!l brands on a series of atiributes f 7-point scales) : cleaning abijit-y, decay prevention sbility, price, fluoride contenc, whitens o~ brightens, abrasiveness, and taste. This ~~t-sas done to aid tke interpretation of the product? space dimensions. Groups 3, 4 and 5 constitutecl the s ~ b j e c t sfor the spatial change strategy pkase of the experiment, After girir-ig 211 gror:ps the stanndard instructions and obtsining dissimilarities data during the %?st week: different communication messages were gil-en to each group in the second ancl third ~veeks,Groups 3 and 5 recelved a message in he second vreek r e g s r d i ~ ~one g brand's advantage with respect to decay prevention. Grotips 4 and 5 recei~ed a message incllrdirng physical rnexsurernertts of abrasireness for all brands a t the third week session. Both messages swre ssid ts come from highly erdibie sources to assure believabiiity arid irnpact.2 Srrbjects ti-ere reminded in each period of the standard instructions given to them at the beg-inning of the study, that is, to judge the brands in terms of the tx-o criteria, decay prevention and abrasiveness, The portion of the experiment cbealing with the struchxral change strategy consisted of groups 6, 7 and 8. Each group was given the standard irastruetions during the first week session and dissimilarity jrtdgments were obhinecf as in all other groups during period. During the next periods, howe~er,subjects 7%-ere instructed to consider similarities among brmds along the two criteria taken one a t a time. Groups 6 2nd 8 were asked to judge 2. The message on decay prevention properties of sodium fluoride was preparect by the experimenters and n5 cIaini of authenticity is made. Ifa~vewr, the data f o r the sbrasiveness message n a r e taken droni the Joz~mzal of the American Dental dsso&tdan (Xovember, 197Gf. OCTOBER, 1977

4.33

Douglas k, MactachOan, Reza Mainpour and James M. McCuilowgh

Downloaded by [McMaster University] at 15:44 14 February 2015

the brands only in terms of the decay prevention criterion during the second week session, while groups 7 and 8 vFere instruetec% to use only the abrasiveness criterion in the third period. The idea here was to see if specific, absolute instructions to use only a single criterion for perceptr~djndgmeats would be sufficient to eliminate the alternate criterjon.

INDSCAL soIutioras were obtained for groups 1 and 2 far

each of the three uyeeks.Vn spite of the fact that group 2 received no instmctions to employ only decay prevention and abrasiveness as judgment criteria while g o n p f had received these instructions, the product spaces SOP the two groups were remarkably similar over the thl-ee-weeIc p e r i d . Pairwise correlations be= tween aII six solutions"inte.erpoint distances averaged 0.96, and C-BXhTCH goodness-of-fit indices were, in every case, Isrger than the correlation values.* More e~~ideace of the abhility of perceptions between control groups and over time is given in Table 1. Mean saliences (dimension weights) across subjects do not differ significantly between ta~eeksfor each group, nor do they differ significantly between grceups.V"PI dimensions are here given IabeIs of "'decay p r e ~ ~ e n tisn" and "whiteningy' for reasons that wilI be discrrssed shortly, Table 1 indicates that group 2 apparently employed approximately the same criteria as group 1, at Beast according .to the dimension weights given. Although there was a alight tendency for group 2 subjects to: treat &xis P as less saEierlt than did group 5 , the differences were not statiaticaHIg significant. 8. See Carroll and Chang (Note 2, 1990) far computer algorithms and Ceiscussion of individual differences scaling (INDSCAL), The INDSGAE procedure results in unique axes (Wish and Carra11, 1973), The uniqueness of the axes ha8 a mibthematical fonaciadon; whether they are 81% psychoIogicnily meaningful depends at a minimum on the assnmptian that the subjects share a consistent frame of referenee in maEng t3eir judgments. 4. Sea Cliff (1966) and PenneIl and Yourag (1967) regarding the cornbgu:~&on matching procedure known as C-MATCH.

ti. The first step of the MANOVA is Box's test of equality of dispersions (multhariate covarimce matrices for all grou~ps).If the latter cannot be rejected, it is appropriate to use an FF-statistic ksed on Wilks' Lambda test for equality of centroids. See Cooley and Loh2es (1971, 228-288), A recent stucly (Olson, 1974) found this test to be fairly robust to deviations from no~malityprovided the group dispersions were not heterogeneous.

43%

MULTIVARIATE BEHAVIORAL RESEARCH

Douglas b. htacliachlan, Rezs Moinpavr and James M. N&uBTsugR --

- ---- --

Table 1 ControIGroups - MAKOVA - -for ---. -

a -

Bimenerions:

A

--

-.-

Nean Saliences Axis I Axis I1 "Decay Prevention" "Whi eening" - ---

Sta~edard Mec~snga Group 1: Week 1 Week 2 Week X

-

--

50

.38

34

.46

.35 .36 .86

41 33 43

Downloaded by [McMaster University] at 15:44 14 February 2015

No Msssaepr Group 2: Week I Week 2 Reek 3

.48 .42

Pooled Groups

Estimated B.D.

.22 .20 Univariate F3,,JI 1.16 ,023 p-level B.S. n.s. F o r equality of dispersions, MANOVA pl;,,, = .84 (n.s.1 Par emaalitp of centroids. MAXOVA Floe .7T(n.s ) -

- --

-

- - -

--

=,

- - - --

-- --

The product space coordinates s h o in ~ ~ Table ~ 2 svere @omputed as a composite solution for group I, meks f through 3, and groups 3, 4 and 5 for week 1. Each of these groups received

the same Enstruetjons and provided ~ e r t ~ d the l y same product space configurations. For convenience, this configuration was used as the basis for making comparisons of treatment effeets. A MANOVA was condt~ctedfor the dimension saliences of these gl.oups. Neither the hypothesis of equality of dispersions noz. eqtastlity of centroids could be rejected,

Tnbie 2

Coordinates of the Common Product Space for Groups Receiving Gtamcisrd lnetmctions ---- - - ---- -- -. --- - - - - -- ----- -Stimulus f Brand) Axis I Ards II . - ----- -. - - -- - - - -A

Closeup

-.934

Colgak

.696

Crest Gleem 11 Epana Macleans Pepsdent

.'SO6

.343 -.0&7 -.414 .302

Ultra BriB

-.756 - - - - -

OCTOBER, f 977

273 .045 .I 39 .288

-.515 -.091 -.ZOO .087

4-35

Downloaded by [McMaster University] at 15:44 14 February 2015

Doughs b. Machachlan, Reza Moinpour and James M, McCuliough

In order to judge the valictity of the dimension labels, m a n attribute ratings (obtzined from group 2, week 3) were employed to fit vectors into the product space with m&ximnm "r" regreasionz."he results shown in Table 3 tend to substantiate labeling axis I as "'decay prevenastEon." The mean attribute ratings are given in Table 3, along with an index of discrimination, carrelations of the attribute means with brand coordinates on the two axes, and the PROFIT rho for fit of each vector. As can be seen from the discr3mination index (a f u ~ e t i o nsf the standard deviations of the means across brands), fluoride, decay prevention, and cleaning ability are criteria along which subjects are able to differentiate the Srar;ds best. As e x p ~ t e d high : csrrelrttions t~-ereobtained between mean attribute ratings and axis 3 for decay prevention and fluoride, the presence of the latter being wideIg associated with causing the former. Subjects appear to have Iess info~mationupon w-hich t0 base judgments for axis TI. Another indication of the lesser impo~-t-tanee of axis I1 as a j~dganentcriterion is the smaller dimension saliences observed in Table X for that axis. Although "'price" was highly correlated with axis 11, there was very low dispersion of brand means on this attribrate idiscrirnination index = .43), and the correlation cor~ldthus be Inflated spuriously. "1Tkritct.n~or brigtrtens9h:as s;Tso associated with axis IH, although it, too, had a somewhat low discrimination index, TIme multiple correlation of the brand projeetr"ons an the "'w~hitensa r brightens" vector with the attribute mean scores myasalso lev. The physical property "degree of abrasiveness" is associated with ability to get teeth white, Since the instrractions were given to w e "abrasirenese9'ans one perceptual criterion, it was hoped that "abrasiveness" would shard out elearly as the dimension mark for axis PI and be strongly correlated with, "whitens sr brightens." There is actuaTIy eonsiderablp higher correlation sf this attribute with axis I. Recall, however, that these ratings were obtained from groirp 2 which did not receive instructions to use the abrwslveness criterion and hence could not have been expected to use the criterion expEicitlg, In any event, the attribute "ab~*asi~-eness" has a strong componert in the direction of axis II, Since both groups 1 and 2 provided the same perceptual configwations, it was decided that abra~ivenesswas a significant enough character6, The PROFIT aIgox+th for performilrg maximu% "p.'"egressior! is discussed by Czrro1I and Chang (h'ote 1): bcilding Gpon their earlier article

(1964).

436

MLliTIVARfATE BEHAVIORAL RESEARCH

-

Downloaded by [McMaster University] at 15:44 14 February 2015

Douglas t. MacLachlan, Weza Moinpour and James

OCTOBER, 1977

-

.,A

M McCuliough

Da~sglas6. Macbechlen, Weaa Msinpcrinr and James hlt. McCullough

istic with which brand repositioning might be attempted through the provision sf additional information along this criterion to subSeek frr experimental groups. The drastic changes in brand position along axis 11 that are shown in the nex? section attezPt to the proper labeling of that axis as ^'whitening,"' with ""abrasive ness" at least as a major component.

Downloaded by [McMaster University] at 15:44 14 February 2015

Spatial AmEysG This wats the heart of the stndy and the section in which the most interesting results occrrrred. Two spatial change strategiea were rmdertaken in this phase. As the first strategy, a message wrasgiven the subjects in groups 3 and 6 (week two) seeking to improve the decay prevention perception of one brand, Glwm PI. As seen in Table 2, Gleem I3 was already perceived as having good decay prevention ability relative to most of the other brands, Therefore it is not sur-prisiny that the brand did not rnak7e much in the product space a a aesuSt of the message (@-MATCH index 09 fit was .96 between configurations). On the other hand, some change was observed in the dimension saliences as a result of the message (see Table 4). Group 3 appeared to increase the weight given the '%ecay prevention'~dimensionand to decrease the vr~eighk given the "'whitening" dimension, No significant change as observed for Group 5 (except for an anomalous, but insignificant, increase in salience for the "whr'tening" dimension). The most startling and significant changes observed in the experiment were found as the result of proa~iding'&he subjects in groups f and 5 (week 3) with physical measures GE the relative abrasiveness af all toothpaste brands in the study. EIere* as anticip a M , large movements in b r a l ~ dposition were obtained w d drastic increases in the salience of the ""alzjtening" dimension were recorded. The latter can be seen in Table 61 between weeks I and 3 for grorrps 4 and 5. These differences are Ialyely responsible for the significance of the F-statistic f a r equality of centraids (see also the p-IeveI for the unirariate analysis of variance for the "whitenirtg" dimension saliences). The C-MATCH prwedure was used to best fit the changed space for group 4 (week 3) t o the common product space configuration. As was expedecl, the C-MATCH index of fit was only .GI. The distance between points on the s~erticalaxis was increaseds indicating the iaereased iinpodnce sf that dimension as a perceptr~attcfiscrirninator for the brands, A compal.ison of the: order 4%

MULTIVARIATE BEHAVIORAL RESEARCH

Douglas b. Macbachlan, Rezs Moirapotar and James M, kfcCulisugh

Table 4 NANOVA For Spatial Group A- n a I ~ sls --- - --- --Dimensions: -Sta&+~d Message Week 1: Group 8

@map 4 Gmup 5

Downloaded by [McMaster University] at 15:44 14 February 2015

-

Mean Saliences Axis II Axis 5 6Wkitaning" "Decay prevention" -

Flw1.irEe infoimrrtances in Caars ~ n f p rResenrch. T'oL I . Asso~iatlo~:f o r Conwmer ReeczrcE, 1974,

Downloaded by [McMaster University] at 15:44 14 February 2015

134-165.

Shepard, Roger h'., Romney, A. Kimball and Seriove, Sara 3.,ieds.1 h1vlf:dhens;onal Smlhig : Tdceorg mid ilppl'icntie% i.rl the B e ~ t c v i o lr ~ Sdmces. Vol. I, Sea- York: Academic Press, lb?2. Sheth, Jagdish X, and Talarzvk, W, K a m e . Perceived ins;~.lmlenta!ity and value importance as dete~w~issnks a? attitudes. Jo2in:ot o f N a r F e f l n g Research, 1972, 9, 6-9, T'arra, Temy G. Factor arialvsis of perceptual change. Jov ma: o f Jiarketing Research. 1972,9, 193-199. Wisk. I\.Ip-on end Carroil, J. DougIas. App!lcatinns of ISDSCAL to stndies o":~umzc percepCor and jadgmext, i?: Carterette, E. C. and Frie&man. &I. P., reds.) Eiandbook cf Percepf:on, Val, I . Seqv York: fcaderric P ~ e s s1973, ,

OCTOBER, 3937

Experimenting With Perceptual Change Strategies.

Nonmetric multidimensional scaling procedures are employed in this experimental evaluation of perceptual changes resulting from differences in the pre...
566B Sizes 0 Downloads 13 Views