Control Foundations: Rationales and Approaches

• Author: Shari Seidman Diamond
• Pages: 201-216

201

9

Control Foundations:

Rationales and Approaches

By Shari Seidman Diamond

INTRODUCTION

The Lanham Actexplicitlyspecifiesthatatrademarkownerwho is

claiming infringementmust provethatconsumers arelikelytobe

influenced bythe actionsof the alleged infringer.Thatis,the trademark

ownermust showthatthe infringer’s use“islikelytocauseconfusion, or

tocausemistake, ortodeceive. ...”1Similarly,the statuterequiresthe

plaintiff alleging deceptiveadvertisingtoshowmisrepresentation likely

todamage the plaintiff.2Analogous language appears in the statute

concerning dilution: the plaintiff must showthatthe use“islikelyto

causedilution. ...” 3Thus,causation isan unambiguous requirementin

each instance. Itisperhapssurprising, therefore, thatearlytrademark

surveys did notrecognizeand address thisstatutory requirement.

1. 15 USC §1125(a)(1)(A)(2011).

2.15USC §1125(a)(1)(B)(2011). Assessing consumerreaction isnecessary in

determining whethermisrepresentation hasoccurred, thatis,thatthe consumerislikely

tobedeceived bythe advertising message, unless the message isliterallyfalseor

necessarilyfalsebyimplication. B. Sanfield, Inc. v.FinlayFine Jewelry Corp., 168F.3d

967,971(7thCir.1999).

3.15USC §1125(c)(1)(2011).

The authorisgratefultoRaquel C. RodriguezatBrinksHoferGilson &Lione for

the research assistance she provided on thisprojectwhen she wasalawstudentat

NorthwesternUniversity.

Section IV

202

Itisnotasif the methodologysuitable forconducting survey-experiments that

could assess causation wasunknown, oreven obscureorimpossible toimplement.

The underpinningsof the method go back atleast tothe publication of The Design of

Experiments in 1935.4Methodsof experimental design havebeen the standardfarein

bothlaboratory and field settingsformanyyears,used in areasasdiverseasmedical

research (doesadrugorothermedical treatmentproduce improvements in health?)and

agricultural studies(isafertilizereffectiveinpromoting plantgrowth?). Yetparties

and experts submitting surveys asevidence in trademarkand deceptiveadvertising

cases,perhapsfailing torecognizethatthe relevantsurveyquestion wasacausal

one, regularlyfailed toinclude the controlsthatarecrucial elements in experimental

designs. 5Courts in turnadmitted and relied on results from thosecontrol-absent

surveys.Astrademarkand deceptiveadvertising surveys haveevolved in the past

20 years,and particularlyin the wake of growing federal court sophistication about

surveys,the tide hasturned.6Although afewcourts in recentyears havebeen willing

toadmitand relyon surveys thatlacked controls,7the absence of acontrol ismore

often recognized asafatal weakness.8

Asimilarneed forastandardforcomparison in the formofacontrol cell or

cellsorin the formofacontrol stimulus orstimuli arisesin surveys assessing

genericismorsecondary meaning. We need toassess whetherconsumerresponses

in the surveyreflectthe constructof interest (e.g., forsecondary meaning, thatthe

consumerbelievesthatthe markindicatesthatthe productcomesfrom asingle

source;forgenericism, thatthe consumerbelievesthatthe markasapplied tothe

productindicatesthatitisabrand ratherthan a“common”name) orwhetherthey

havebeen produced bynoiseduetoguessing orsome othersource (e.g., abelief by

some consumers thatall vodka ortomatosauce ismade orauthorized bythe same

company).

The second section of thischapterbeginsbyreviewing in detail the threats to

valid inference posed when acontrol isabsent.The thirdsection then shows how

including an appropriatecontrol cell can eliminatenearlyall competing explanations

4. The authorwasBritishstatistician Ronald Fisher,called “the fatherof modernstatisticsand

experimental design”byRichardDawkinsatwww.edge.org(2010).

5. Foran earlydiscussion of thisneglectin casesof deceptiveadvertising, see ShariSeidman

Diamond, Using Psychology to Control Law: From Deceptive Advertising to Criminal Sentencing, 13

L&H. B.239–52(1989).

6.ShariSeidman Diamond, Reference Guide on Survey Research 400, in R M 

S E (3ded., Fed. Jud. Ctr.2011) (documenting growthofsurveys withcontrol groups

in Lanham Actcases).

7.See, e.g., Ironclad v.Poly-America, 2000 U.S. Dist.LEXIS 10728*23–25(N.D. Tex.2000)

(although the court viewed the lack of acontrol as“the most significantchallenge tothe surveyevidence”

and discounted its weight,the court “yetstill gives[the survey]results some consideration”).

8. See, e.g., Procter&Gamble Pharm., Inc. v.Hoffman-LaRoche, Inc. 2206 U.S. Dist.LEXIS 64363

*25(S.D.N.Y. 2006);Simon Property GroupL.P. v.mySimon,Inc., 104F.Supp. 2d1003,1045–51 (S.D.

Ind. 2000);National Football LeagueProperties,Inc. v.ProStyle,Inc., 57F. Supp. 2d665, 668(E.D.Wis.

1999);P&G v.Ultreo, Inc., 574F.Supp. 2d339, 351–52(S.D.N.Y. 2008);Bracco Diagnostics,Inc. v.

Amersham Health, Inc., 627 F. Supp. 2d384, 448 (D. N.J. 2009).