An Assessment of Chronic Regulatory Focus Measures

*Kelly L. Haws is Assistant Professor of Marketing, Mays BusinessSchool, Texas A&M University (e-mail: [email protected]). UtpalM. Dholakia is William S. Mackey and Verne F. Simons DistinguishedAssociate Professor of Management, Rice University (e-mail: [email protected]). William O. Bearden is Distinguished Professor Emeritus, Bankof America Professor in Business Emeritus, Moore School of Business,University of South Carolina (e-mail: [email protected]). The firstauthor acknowledges the Mays Business School for its financial support ofthis research and Jeff Conant for his never-ending encouragement and sup-port. Chris Janiszewski served as associate editor for this article.

KELLY L. HAWS, UTPAL M. DHOLAKIA, and WILLIAM O. BEARDEN*

Prior consumer research has demonstrated the ability of promotionand prevention regulatory orientations to moderate a variety of consumerand marketing phenomena but also has used several different methodsto measure chronic regulatory focus. This article assesses five chronicregulatory focus measures using the criteria of theoretical coverage,internal consistency, homogeneity, stability, and predictive ability. Theresults reveal a lack of convergence among the measures and variationin their performance along these criteria. The authors provide specificguidance for choosing a particular measure in regulatory focus researchand suggest a composite measure.

Keywords: regulatory focus, composite regulatory focus scale, promotion,prevention, goal orientation, chronic traits, measurement

An Assessment of Chronic Regulatory FocusMeasures

© 2010, American Marketing AssociationISSN: 0022-2437 (print), 1547-7193 (electronic) 967

Journal of Marketing ResearchVol. XLVII (October 2010), 967–982

Regulatory focus theory (Higgins 1997, 1998) elaborateson the means people employ for self-regulation during goalpursuit and distinguishes between two regulatory orienta-tions: a promotion focus and a prevention focus. A promo-tion focus emphasizes the “ideal” self, as reflected in theperson’s hopes and aspirations, and favors strategic meansthat are eagerness oriented. In contrast, a prevention focusemphasizes the “ought” self, as reflected in the person’sduties and obligations, and supports strategic means that arevigilance oriented. Thus, a promotion focus emphasizes thepresence of positive outcomes and minimizing errors ofomission (e.g., missing opportunities to make progress),whereas a prevention focus favors the absence of negativeoutcomes and minimizing errors of commission (e.g., doingsomething that turns out to be a mistake; Higgins 1997;Higgins and Spiegel 2004). Regulatory focus theory isincreasingly used in consumer research to explain a widerange of consumer phenomena (for a review, see Pham andHiggins 2005).

As the body of regulatory focus research continues togrow, an issue with the potential to muddle our understand-ing pertains to the diverse operationalizations of people’schronic regulatory focus. Researchers have used several dif-ferent approaches to measure chronic regulatory focus. Weidentify and study five distinct measures in this article. Typi-cally, when conducting a study on chronic regulatory focus,researchers must choose one of these measures. However,they have little guidance on which to base their choicebecause prior studies have not considered how these meas-ures overlap or vary in theoretical content, compared theirpsychometric properties, or explained which measure maybe more suitable in one or another study context.We attempt to shed light on this issue. Through a series

of studies, we investigate the convergence, psychometricproperties, and predictive power of the measures used toassess chronic regulatory focus. We also examine the theo-retical domain covered by these measures with regard to theprimary postulates of regulatory focus theory. Through thisanalysis, we offer researchers useful information and guid-ance regarding which measures they should use when con-ducting research involving a chronic conceptualization ofregulatory focus. Most of the detailed analyses that supportour conclusions are available in the Web Appendix (seehttp://www.marketingpower.com/jmroct10).

REGULATORY ORIENTATIONS OF CONSUMERS

Regulatory focus theory has gained influence in psychol-ogy, marketing, and beyond because of its ability to explainand predict a variety of psychological processes and behav-

http://www.marketingpower.com/jmroct10

968 JOURNAL OF MARKETING RESEARCH, OCTOBER 2010

examining regulatory fit and its consequences have employedchronic as well as situational operationalizations of con-sumers’ regulatory focus.

The Operationalization of Chronic Regulatory Focus

Recent consumer research studies have employed achronic operationalization of regulatory focus (e.g., Aakerand Lee 2001; Avnet and Higgins 2006; Herzenstein,Posavac, and Brakus 2007; Pham and Avnet 2004), with atleast five different measurement approaches for chronicregulatory focus assessments. Table 1 provides a summaryof the measures, along with examples of prior consumerresearch that uses each scale. Despite some overlap in themethods of administration and content across the measures,there are also significant differences. We introduce andbriefly describe each measure and then proceed to conductan initial test of the relationships among them in Study 1.Regulatory focus questionnaire. Perhaps the most fre-

quently employed scale to measure chronic regulatory focusis the 11-item regulatory focus questionnaire (RFQ) thatHiggins and colleagues (2001) developed. The RFQ con-tains six reflective items that assess chronic promotionfocus and five reflective items that assess chronic preven-tion focus. In the RFQ design, Higgins (2002) conceptual-izes promotion and prevention success as promotion prideand prevention pride, respectively, or anticipatory reactionsto new task goals, derived from the person’s subjective his-tory of past success in promotion and prevention goal attain-ment. The two RFQ subscales generally have a low correla-tion, indicating independence.BIS/BAS scale. A second means to assess chronic regula-

tory focus is the BIS/BAS scale that Carver and White(1994) developed to assess Gray’s (1990) regulatory sys-tems. These systems include the behavioral inhibition sys-tem (BIS), which is sensitive to negative outcomes, and thebehavioral activation system (BAS), which is responsive topositive outcomes. The seven BIS items (prevention focus)reflect concerns about the possibility of a negative occur-rence or sensitivity to such events when they occur. The fiveBAS items reflect responsiveness to rewards, measured asexcitement tendencies or arousal associated with positiveexperiences.1Selves questionnaire. The third chronic regulatory focus

measure employs Higgins and colleagues’ (1986) selvesquestionnaire. It assesses the degrees of congruency amongpeople’s actual selves, their ideal selves, and their oughtselves. These three selves represent the attributes peoplebelieve they possess, the attributes they aspire to possess,and the attributes they perceive as their duty to possess,respectively (Pham and Avnet 2004). A promotion focus isrepresented by the degree of congruence between theiractual and ideal selves, whereas a prevention focus reflectsthe convergence between actual and ought selves (Brockneret al. 2002).In the most commonly used approach to administer the

selves questionnaire, study participants provide six differentattributes that they either would like to possess (ideal attrib-

iors. A key assumption underlying regulatory focus theoryis that though some people are chronically more promotionoriented and others are more prevention oriented, these ori-entations also can be activated by situational demands, suchas through experimental promotion versus prevention fram-ing (e.g., Avnet and Higgins 2006; Wang and Lee 2006;Zhou and Pham 2004). In experimental studies, tasks suchas essay writing, reflecting on past experiences, reading per-suasive information framed in gain/nongain versus loss/nonloss terms, and describing aspirations or duties havebeen successfully used to activate participants’ situationalregulatory orientations. Previous research also providesample evidence that orientations can be chronically (trait)or temporarily (situationally) more accessible to a person(e.g., Avnet and Higgins 2006; Higgins 1997, 1998, 2000).Although we recognize the importance of a consumer’s sit-uationally accessible regulatory focus, we limit our assess-ment to chronic regulatory focus measures.The chronic form of regulatory focus largely derives from

a person’s developmental history (Higgins and Silberman1998). Higgins (2002) explains that regulatory orientationsreflect distinct and enduring concerns about how to getalong in the world, stemming from how the person wassocialized in his or her childhood. The social regulatorystyle of parents, for example, can emphasize either nurtu-rance or security. Nurturing social regulation engenders apromotion focus in the child, marked by an enduring con-cern with advancement and accomplishments; security-based social regulation instead fosters a prevention focuscentered on protection and responsibilities (Higgins 2002).Furthermore, Higgins (2002) posits that promotion and pre-vention concerns are not the endpoints of a unidimensional,bipolar construct; rather, the promotion or prevention systemmay be chronically more accessible than the other systemfor a person, but both systems also coexist independently.

Regulatory Focus in Marketing Research

Regulatory focus theory has generated significant con-sumer and marketing research. For example, Lee and Aaker(2004) find that promotion-focused consumers are morepersuaded by appeals framed in terms of gains, whereasprevention-focused participants are persuaded to a greaterdegree by appeals framed in terms of losses. Jain and col-leagues (2007) determine that promotion- and prevention-focused participants react differently to negative versuspositive frames in comparative brand advertising, and Zhaoand Pechmann (2007) find that antismoking advertisementsare most persuasive when the viewers’ regulatory focus andthe message’s valence (positive or negative) are congruent.Studies have found that the consumer’s regulatory focusplays important roles in processes related to persuasion,self-regulation, categorization, judgment, and choice.In other recent consumer research, applications of regula-

tory focus theory have emphasized the importance of regu-latory fit (e.g., Avnet and Higgins 2006; Hong and Lee2008; Wang and Lee 2006; Zhou and Pham 2004). Regula-tory fit occurs when people derive value from using strate-gic means during goal pursuit that align with their under-lying regulatory orientations. Specifically, when consumersengage in activities that are consistent with their regulatoryorientation, they experience heightened motivation and asense that “it just feels right” (Aaker and Lee 2006). Studies

1In the original BIS/BAS scale (Carver and White 1994), BAS had threedimensions: reward responsiveness, drive, and fun seeking. However, con-sistent with previous regulatory focus theory research (e.g., Dholakia et al.2006), we use the BAS reward responsiveness measure.

Chronic Regulatory Focus Measures 969

scores of ideal and actual ratings are summed to obtain themeasure of promotion focus, and difference scores of oughtand actual ratings are summed to obtain the preventionfocus measure (see Brockner et al. 2002; Higgins, Shah, andFriedman 1997; Pham and Avnet 2004). We employed thisassessment approach in our studies.Using a somewhat different approach, Avnet and Higgins

(2006) compute two scores based on the match betweenactual and ideal attributes and between actual and ought

utes) or think they should possess (ought attributes), in ran-dom order. After each ideal attribute, participants indicatethe extent to which they would like to possess the attribute(ideal rating), and after each ought attribute, they reveal theextent to which they should possess the attribute (ought rat-ing). Then, for all six attributes, participants indicate thedegree to which they currently possess that attribute (actualrating). Differences between actual ratings and either idealor ought ratings are calculated for each attribute. Difference

Table 1MEASURES OF CHRONIC REGULATORY FOCUS IN CONSUMER RESEARCH

Measure Items/Description Used by

RFQ (Higgins et al.2001)

1. Compared to most people, are you typically unable to get what you want out of life? (R)a2. Growing up, would you ever “cross the line” by doing things that your parents would not tolerate? (R)3. How often have you accomplished things that got you “psyched”‘ to work even harder?a4. Did you get on your parents’ nerves often when you were growing up? (R)5. How often did you obey rules and regulations that were established by your parents?6. Growing up, did you ever act in ways that your parents thought were objectionable? (R)7. Do you often do well at different things that you try?a8. Not being careful enough has gotten me into trouble at times. (R)9. When it comes to achieving things that are important to me, I find that I don’t perform as well as I

ideally would like to do. (R)a10. I feel like I have made progress toward being successful in my life.a11. I have found very few hobbies or activities in my life that capture my interest or motivate me to put

effort into them. (R)a

Louro, Pieters, andZeelenberg(2005); Hongand Lee (2008)

BIS/BAS scales(Carver and White1994)

1. If something unpleasant is going to happen I usually get pretty “worked up.”2. I worry about making mistakes.3. Criticism or scolding hurts me quite a bit.4. I feel pretty worried or upset when I think or know somebody is angry at me.5. Even if something bad is about to happen to me, I rarely experience fear or nervousness. (R)6. I feel worried when I think I have done poorly at something.7. I have very few fears compared to my friends. (R)8. When I get something I want, I feel excited and energized.a9. When I’m doing well at something, I love to keep at it.a10. When good things happen to me, it affects me strongly.a11. It would excite me to win a contest.a12. When I see an opportunity for something I like, I get excited right away.a

Dholakia, Gopinath,and Bagozzi(2005); Dholakiaet al. (2006)

Selves questionnaire(Brockner et al.2002; Higgins etal. 1986; Shah,Higgins, andFriedman 1998)

In the following pages, you will be asked to list the attributes of the type of person that YOU believe youactually are, you would ideally like to be, and you ought to be:

Your actual self: Your beliefs concerning the attributes or characteristics you think you actually possessnow.

Your ideal self: Your beliefs concerning the attributes or characteristics you would like ideally topossess; the type of person you wish, desire, or hope to be.

Your ought self: Your beliefs concerning the attributes or characteristics you believe you should orought to possess; the type of person you believe it is your duty, obligation, orresponsibility to be.

Pham and Avnet(2004); Avnetand Higgins(2006); Jain etal. (2007)

Lockwood scale(Lockwood,Jordan, and Kunda2002)

1. In general, I am focused on preventing negative evets in my life.2. I am anxious that I will fall short of my responsibilities and obligations.3. I frequently imagine how I will achieve my hopes and aspirations.a4. I often think about the person I am afraid I might become in the future.5. I often think about the person I would ideally like to be in the future.a6. I typically focus on the success I hope to achieve in the future.a7. I often worry that I will fail to accomplish my academic goals.b8. I often think about how I will achieve academic success.a, b9. I often imagine myself experiencing bad things that I fear might happen to me.10. I frequently think about how I can prevent failures in my life.11. I am more oriented toward preventing losses than I am toward achieving gains.12. My major goal in school right now is to achieve my academic ambitions.a, b13. My major goal in school right now is to avoid becoming an academic failure.b14. I see myself as someone who is primarily striving to reach my “ideal self”—to fulfill my hopes, wishes,

and aspirations.a15. I see myself as someone who is primarily striving to become the self I “ought” to be—fulfill my duties,

responsibilities, and obligations.16. In general, I am focused on achieving positive outcomes in my life.a17. I often imagine myself experiencing good things that I hope will happen to me.a18. Overall, I am more oriented toward achieving success than preventing failure.a

Yeo and Park(2006); Zhao andPechmann(2007)

aBAS or promotion item.bItem was not used with adult samples.Notes: (R) = reverse scored.


measures is indicated by more negative correlations. We cal-culated the response latency measures as described previ-ously; in addition, because the latency distributions werepositively skewed, we performed natural log transforma-tions on the latencies before summing them. This tactic isconsistent with the approach followed by other researcherswho have used this measure (e.g., Shah, Higgins, and Fried-man 1998).

Results

Factor analyses. Separate exploratory factor analyses con-ducted for the 11-item RFQ, the 12-item BIS/BAS scale, andthe 18-item Lockwood scale revealed clean, two-factor load-ing patterns in each case, with each item demonstrating asimple structure on the appropriate subscale. Detailed resultsfor the exploratory factor analyses are available in the WebAppendix (http://www.marketingpower.com/jmroct10). Wesummarize the descriptive statistics (means and standarddeviations) and Cronbach’s α values for all the measures inthe top panel of Table 2. The promotion focus subscale ofthe selves questionnaire had lower-than-acceptable (<.60)reliability, whereas the other subscales achieved adequatereliability. In addition, we performed confirmatory factoranalysis (CFA) on these data to assess the factor structure.The results of this analysis, as provided in the Web Appen-dix, are consistent with an orthogonal, two-dimensionalconceptualization of regulatory focus orientations (Higgins2002).Correlations among regulatory focus measures. Table 2

also provides the bivariate correlations between subscales.Notably, the correlations between the corresponding sub-scales achieved minimal convergence (Campbell and Fiske1959). Most correlations are less than .30 in magnitude. Theonly correlations greater than this magnitude are +.45between the selves promotion subscale and the selves pre-vention subscale, +.80 between the self-guide strength pro-motion and prevention measures, +.38 between the BASscale and the Lockwood promotion subscale, and +.36between the BIS scale and the Lockwood prevention sub-scale. Therefore, we find minimal correlations among thescales intended to assess a promotion focus and among thescales designed to assess a prevention focus. Furthermore,the low to nonexistent correlations between the self-guidestrength measures and the other scales are unsurprising,considering their fundamentally different methods (i.e.,response latencies for the former versus self-reported rat-ings for the remaining scales). These low correlations indi-cate different underlying processes for implicit versusexplicit measurement approaches (Cunningham, Preacher,and Banaji 2001).To confirm the surprising lack of convergence among the

different subscales, we replicated this study with a secondstudent sample (N = 232). The results in the bottom panelof Table 2 show that the reliabilities of both selves subscalesare below the conventional .60 threshold. The same fourcorrelations as in the first sample are greater than +.30. Thecorrelation between the selves promotion and preventionsubscales is +.44, and that between the self-guide strengthpromotion and prevention subscales is +.46. The Lockwoodpromotion subscale and the BAS scale exhibit a correlationof +.51, and the Lockwood prevention subscale and the BISscale have a correlation of +.39. In contrast with the first

attributes, as scored by independent raters. Study partici-pants have a promotion orientation if their actual–ideal con-gruency score is higher than their actual–ought congruencyscore and a prevention orientation otherwise. This approachcollapses promotion and prevention focus into a single bipo-lar dimension.Self-guide strength measure. The fourth approach meas-

ures the chronic accessibility of the participants’ ideals andoughts, employing latencies of their responses in the selvesquestionnaire as measures of strength of their respectiveregulatory focus (Shah, Higgins, and Friedman 1998). Par-ticipants complete the tasks of eliciting ideal and oughtattributes, followed by ideal or ought ratings and actual rat-ings, on a computer equipped with the ability to recordresponse times. The time taken by each participant to pro-duce each attribute and rate it is recorded. Attribute elicita-tion latencies and ratings latencies are summed across thethree ideal attributes and across the three ought attributes,separately, resulting in one total ideal strength assessmentand one total ought strength assessment. In this way, it ispossible to assess both the selves questionnaire and the self-guide strength measures in a single computer-based session.We employed this approach in all our studies.Lockwood, Jordan, and Kunda’s (2002) measure. The

fifth measure of regulatory focus, as described by Lock-wood, Jordan, and Kunda (2002), was developed to studymotivation instigated by role models. The 18 items theydescribe incorporate key regulatory focus theory terms andconsider success and failure in goal pursuit, with 9 reflec-tive items each for promotion and prevention focus. Similarto the RFQ measure, Lockwood, Jordan, and Kunda’s meas-ure (which we refer to hereinafter as the Lockwood scale)achieves only a modest correlation of +.17 between the pro-motion and the prevention focus subscales.

STUDY 1

Prior research has not examined the degree to which themeasures of chronic regulatory focus converge in theirassessment. Accordingly, the primary objective of Study 1was to conduct correlation tests to investigate convergenceof the promotion and prevention focus scales. We also repli-cated Study 1 using a second sample.

Method

We gathered the data from 174 students who receivedcourse credit for participation. These participants completedthe RFQ, BIS/BAS, selves (including response latencies toassess the self-guide strength measure), and Lockwoodmeasures. We counterbalanced the administration of thescales across participants, and short filler tasks, such as rat-ing artwork and solving simple anagrams, appearedbetween each scale to provide cognitive separation.For the RFQ, BIS/BAS, and Lockwood measures, we cal-

culated averages of the respective items. For the selves pro-motion focus measure, we calculated the sum of the differ-ence scores of the three ideal ratings and the threecorresponding actual ratings. Lower scores reflect morecongruence between actual and ideal selves and, thus, astronger promotion focus (Brockner et al. 2002). A similarprocedure served to compute the selves prevention focusmeasure. Greater convergence between these measures andthe corresponding RFQ, BIS/BAS, and Lockwood scale



Table2

STUDY1:VARIABLE

MEANS,STA

NDARDDEVIATIONS,RELIABILITIES

,ANDINTERCORRELATIONS

MSD

Alpha

12

34

56

78

910

Study1(N=174)

1.RFQ

promotion

5.30

.79

.64

—2.

RFQ

prevention

4.74

1.11

.80

.19*

—3.

BAS(promotion)

6.14

.84

.88

.21**

–.00

—4.

BIS(prevention)

4.89

.97

.80

–.12

.20*

.25**

—5.

Selvespromotion

5.25

2.08

.58

–.13

–.07

.06

.19*

—6.

Selvesprevention

4.02

2.05

.64

–.08

–.13

.08

.20**

.45**

—7.

Self-guidestrengthpromotion

3.17

3.27

.76

–.00

.08

.01

–.01

.14

.07

—8.

Self-guidestrengthprevention

3.11

3.27

.73

.02

.09

.05

–.00

.11

.06

.80**

—9.

Lockwoodpromotion

5.70

.85

.85

.25**

.07*

.38**

.26**

–.07

–.08

.02

.03

—10.Lockwoodprevention

4.12

1.00

.77

–.28**

–.14

.05

.36**

.03

.05

–.09

–.15*

.27**

—

ReplicationSample(N=232)

1.RFQ

promotion

3.94

.53

.64

—2.

RFQ

prevention

3.52

.77

.80

.13

—3.

BAS(promotion)

3.62

.39

.77

.21**

.12

—4.

BIS(prevention)

3.94

.51

.78

–.19**

.12

.24**

—5.

Selvespromotion

5.33

1.99

.58

–.19**

.01

.02

.18**

—6.

Selvesprevention

4.32

2.24

.54

–.20**

.00

–.09

.04

.44**

—7.

Self-guidestrengthpromotion

4.56

.30

.65

.01

.03

–.02

.02

.06

.02

—8.

Self-guidestrengthprevention

4.53

.34

.62

.05

.10

–.01

–.04

–.04

.01

.46**

—9.

Lockwoodpromotion

7.42

1.00

.78

.27**

.07

.51**

.09

.01

–.12

.06

–.06

—10.Lockwoodprevention

5.30

1.37

.85

–.48**

–.18**

.04

.39**

.16*

.08

.02

–.19**

.12

—

*p<.05.

**p<.01.


sample, the correlation between RFQ promotion and Lock-wood prevention is –.48.

Discussion

The main result from these correlation tests reveals thatexcept for the moderate convergence of the BIS/BAS andLockwood subscales, the scales demonstrated minimal con-vergence. This result was consistent across both samples,even though the scales were administered consecutively inthe same experimental session and were intended to meas-ure the same underlying constructs. The conservative natureof this test—that is, a consecutive counterbalanced adminis-tration of the measures that should have accentuated exist-ing relations—makes these results even more intriguing andraises pressing questions about the differences among thescales, as well as which scales consumer researchers shoulduse when conducting research involving a chronic concep-tualization of regulatory focus. By carefully assessing themeasures, we seek to answer these questions.

ASSESSMENT OF CHRONIC REGULATORY FOCUSMEASURES

To understand the reasons for the lack of convergence anddetermine which chronic regulatory focus measure to use,we evaluated the measures according to four criteria thatpersonality literature has identified as comprehensive forscale assessments: (1) representativeness, (2) internal con-sistency and homogeneity, (3) stability, and (4) predictivevalidity (Simms and Watson 2007). First, representativenessrefers to the scale’s theoretical content. A scale is represen-tative to the extent to which its items cover all the importanttheoretical facets of the target construct. A representativescale should have items that reflect different manifestationsand levels of the underlying trait for which the reliablemeasurement is desired (Simms and Watson 2007). Thisparticular criterion can be evaluated only with a thoroughunderstanding of the theoretical underpinnings of the meas-ures and an examination of their specific content.Given the findings of Study 1 and our subsequent studies,

each regulatory focus scale appears to have unique strengthsand weaknesses, and the measures appear to emphasize dif-ferent aspects of regulatory focus theory. Therefore, themeasures appear to be differentially representative of theunderlying tenets of regulatory focus theory. We provide adetailed examination of the representativeness of the scalesin the Web Appendix (http://www.marketingpower.com/jmroct10) and a summary of this analysis in Table 3. There-fore, Table 3 provides a summary of our analyses regardinghow well the measures incorporate the key tenets of regula-tory focus theory, item-specific issues with each measure,and suggested uses according to our analyses. We demar-cate our theoretical analysis into three sections: the scale’sprimary theoretical focus, how well it captures strategicaspects of goal pursuit and approach/avoidance motivations,and affective and cognitive elements of regulatory focustheory.Second, the scale should have high internal consistency

or coherence, which includes its reliability, as measured byindexes such as Cronbach’s α, and homogeneity, or theextent to which all items in the scale tap a single dimension.Study 1 revealed that the subscales of the selves question-naire (promotion in both, prevention in one sample) had

lower reliabilities than the conventional .60 threshold,whereas the other scales had higher reliabilities. Throughboth exploratory analysis and CFA, Study 1 also providedpreliminary evidence that the RFQ, BIS/BAS, and Lock-wood scales were homogeneous. We test scale homogeneityfurther by conducting additional CFAs in Study 2.Third, the scale should have stability, such that the meas-

ure remains stable for the same participant across measure-ments at different times. We examine test–retest reliabilityin Study 2.2Fourth, the scale should have predictive validity and be

able to match and/or predict relevant behavioral manifesta-tions postulated by theory. To meet this criterion, partici-pants who score high on the chronic promotion focus shouldbehave in ways that are consistent with a promotion focus,and those scoring high on the chronic prevention focusshould show behaviors that theory suggests are consistentwith such a focus. The same should be true for medium andlow levels of promotion and prevention focus. We test thepredictive validity of the scales in Studies 3a and 3b. Inaddition to these four criteria, we assess the scales for logis-tical and administrative issues (see the Web Appendix,http://www.marketingpower.com/jmroct10).

STUDY 2

Study 2 was computer based and designed to evaluate thehomogeneity and stability of the five promotion and preven-tion focus measures using a standard test–retest reliabilityexperimental design (Peter 1979). Homogeneity is theextent to which the items in a subscale tap into a singledimension, whereas stability is the scale’s ability to obtainconsistent responses across multiple administrations.

Method

A total of 197 undergraduate students participated in atwo-part study involving regulatory focus measurement aspart of a class assignment. Participants received a link to anonline survey at Time 1 and again at Time 2, with fiveweeks between the two measurements. Although partici-pants had four days each time to complete the study, theaverage time between administrations was five weeks. Eachversion of the study included all regulatory focus scales(along with response latencies gathered in conjunction withthe selves questionnaire), collected in a random order foreach participant and separated by brief, unrelated filler tasksto provide cognitive separation between the measures.To assess the homogeneity of the RFQ, BIS/BAS, and

Lockwood scales, we used CFAs in the LISREL 8.80 pro-gram (Jöreskog and Sörböm 1999). These three scales aresummated scales, whereas the selves questionnaire and self-guide strength measures are cumulative scales (Pedhazurand Schmelkin 1991); consequently, CFA is not applicableto the latter two scales. We assessed each model’s goodness-of-fit with the root mean square error of approximation(RMSEA), standardized root mean square residual(SRMR), nonnormed fit index (NNFI), and comparative fitindex (CFI). Bentler (1990) and Marsh, Balla, and Hau

2Another way to gauge the scale’s stability is to examine the extent towhich it changes after a situational manipulation of the factor. We providethe details of such a test in Study 2b in the Web Appendix (http://www.marketingpower.com/jmroct10).







Table3

THEORETICALCONTENTANDSUGGESTEDUSESOFCHRONIC

REGULATO

RYFOCUSMEASURES

RelationtoTenetsofRegulatoryFocusTheory

RegulatoryFocus

Measure

PrimaryTheoretical

Emphasis

StrategyDuringGoal

PursuitandApproach/

AvoidanceMotivations

Affective/Cognitive

RegulatoryFocusTheory

Elements

CompositionofItemsa

Promotion-Specific

Issues

Prevention-Specific

Issues

SuggestedUses

RFQ

Anticipatorygoal

reactionsbasedonpast

successusing

promotion/prevention

orientationsandthe

resultingpridefrom

those

successes.

Distinguishesbetween

approachandavoidance

withineachofthetwo

regulatoryorientations.

Approachpositive

outcomes,avoid“crossing

theline”orbreakingthe

rules.

Doesnotfocuson

affectiveaspectsof

regulatoryfocus.More

emphasisonoutcomes

andcognitionsabout

outcomes.

Uniqueamongscalesin

focusingonpastevents.

Widevarietyincluding

moregeneralandmore

specificinstances.

Achievingsuccess;

accomplishingwhat

youwantto.

Adherencetorules

andspecifically

parentalrules;may

engenderasenseof

rebelliousness

castingpositivelight

onpreventionfocus.

Mostgenerallyapplicable

andsuitablescalefor

testingtheory.Unsuitable

foroperationalizing

emotionalcomponentsof

regulatoryfocus.

BIS/BASscale

Basedonregulating

aversive/appetitive

motivationsand

approach/avoidance

regulatorysystemsin

whichonewishesto

approachsuccessor

pleasantoutcomesand

inhibitfailureor

unpleasantoutcomes.

Emphasizesemotional

reactionsofrespondents.

BASitemscapture

approachbehaviors

mostlythroughemotions;

BISaddressesthe

unpleasantemotionsfrom

negativeoutcomes.Does

notdistinguishbetween


withineachregulatory

orientation.

Stronglyemphasizes

affectiveaspects:

Emphasisonworry,

nervousness,hurt,fear

(BIS)andexcitement

(BAS).Addresses

motivationalaspectsof

goalachievement.

Cognitiveelementsof

regulatoryfocusarenot

measuredbythesescales.

Scalestapintoaffective

responsestoeventsand

outcomes.Reactionsof

respondentstofearof

failure(prevention)orthe

joyofsuccess

(promotion).

Excitementor

positiveenergy

generatedfrom

good

outcomes.

Dreadornegative

energyinresponseto

potentialfailureor

negativeoutcomes.

Mostsuitableforstudying

questionsinvolving

emotion-basedself-

regulationofconsumers.

Weakestofallthe

measuresexaminedalong

therepresentativeness

criterion.M

aynotbe

suitableforstudying

questionsinvolving

cognitiveprocessing,

judgmentanddecision

making.

Selvesquestionnaire

Basedonself-discrepancy

theory,thisscalemeasures

thelevelofmatch

betweenactualand

ideal/oughtselves.

Emphasisisonfailingin

eitherpromotionor

preventionself-regulation

andthereforeworkingto

achievemoreineither

dimension.

Assumesdiscrepancies

representselvesweare

strivingfortoachieve

idealsorachieveoughts.

Presumablylessemphasis

onavoidanceofnegative

outcomes.

Dependsonaffectelicited

bythechosenattribute.

Lessdiscrepancybetween

actualandidealoractual

andoughtisviewedas

producingpositiveaffect,

andviceversa.

Relatedtocharacteristics

peoplebelievesthatthey

have,shouldhave,and

wouldliketohave.Could

leadtowidelydifferent

definitionsfordifferent

respondents.Ideal

attributeforoneperson

couldbeoughtattribute

foranotherone.

Differencescores

obfuscateactuallevel

ofidealattributes

desired,suchas

strengthofapproach

motivation.

Sameissueas

promotion,strength

ofavoidance

motivationisnot

capturedbecauseof

differentscoring

methodology.

Suitableasadependent

measureofregulatory

focusand/ormeansof

manipulatingregulatory

focus.Lessusefulasa

dispositionalmeasureof

chronicregulatoryfocus

exceptwhereself-

discrepancytheoryis

specificallybeingtested.

Item-SpecificIssues


Table3

CONTINUED

a Theoreticaldimensionsandmixofantecedents,indicators,andconsequences.

Notes:N.A.=notapplicable.

RelationtoTenetsofRegulatoryFocusTheory

RegulatoryFocus

Measure

PrimaryTheoretical

Emphasis

StrategyDuringGoal

PursuitandApproach/

AvoidanceMotivations

Affective/Cognitive

RegulatoryFocusTheory

Elements

CompositionofItemsa

Promotion-Specific

Issues

Prevention-Specific

Issues

SuggestedUses

Self-guidestrength

measures

Basedontheideaof

attitudeaccessibility

throughresponse

latencies.

Quickresponsesfor

attributesthataremore

important.Doesnotmake

distinctionbetween


self-regulation.

Reliesonanassociative

processcharacterization

ofregulatoryorientations.

Affectiscapturedthrough

implicitprocessingmodel.

Responselatenciesforthe

attributeslistedforselves

measure.Sameissuesof

diversedefinitionsand

attributedissimilarity

acrossparticipantsasthe

selvesquestionnaire.

N.A.

N.A.

Mostsuitablefortheory-

testingforresearch

questionsinvolving

unconsciousand

automaticprocessing.

Mayalsobeeffective

whensocialdesirability

biasisapossibility.

Lockwoodscale

Intendedtobeameasure

oftheprimarytenetsof

regulatoryfocustheory.

Incorporateskeyconcepts

inwordingofmeasures.

Presentandfuturefocus

onthinkingabout

achievingsuccess/ideals

orthinkingabouthow

topreventfailures.Less

emphasisonavoiding

nongainsorapproaching

nonlosses.

Includesmeasureofhope

forideals;otherwisemore

negativeemotionfocuson

anxiety,worry,andfear.

Cognitiveresponsesto

potentialfutureoutcomes

areincluded.

Mixofcognitive,

affective,andbehavioral

responses.Itemsare

mostlyabstract.Some

itemsseem

narrowly

suitedtostudents(e.g.,7,

8,12,13).Focusison

respondingtocurrentand

futureeventsratherthan

pastgoalpursuit.Includes

items(e.g.,11,18)that

capturedifference

betweenpromotionand

preventionwithinasingle

item.

Keytermsinclude

ideal,achieve,

success,hopes,

wishes,and

aspirations.

Keytermsused

includeought,avoid,

prevent,duties,

responsibility,and

obligations.

Canbeusedasan

alternativetotheRFQ

for

theory-testing.Couldbe

usedasafirstchoice

whenresearchersare

specificallyinterestedin

studyingresearch

questionsinvolving

academicperformance

andachievement.

Item-SpecificIssues


for the RFQ subscales, the BIS/BAS scales, and the Lock-wood subscales. The results also indicate that the selvesquestionnaire subscales and the self-guide strength sub-scales are relatively unstable over a longer period.

STUDY 3A

In Study 3a, we examine the predictive validity of thescales—that is, their ability to predict salient outcomes, ashypothesized by regulatory focus theory. Study participantsresponded to dependent measures that elicited their adop-tion of innovative products, preference for a secure versus ahigh-paying job, preference for investing an inheritance in arisky business venture versus a safer mutual fund, and thepersuasive impact of public service announcement (PSA)advertisements, along with assessments of their affectiveresponses to and the substantive content in these advertise-ments. Prior research has shown or hypothesized that thesedependent measures are influenced by a person’s regulatoryorientations (Pham and Higgins 2005).

Method

Participants were 367 undergraduate students participat-ing in a behavioral laboratory research session in exchangefor course credit. The study contained two main parts (scaleassessment and dependent measure collection), along with a15-minute unrelated task between the parts to provide tempo-ral separation. Half the participants completed the measure-ment portion first, and the remainder completed the depend-ent measure portion first. The order of responses was notsignificant in the analysis, so we do not consider it further.In the scale assessment part, participants responded to

each of the regulatory focus scales, counterbalanced acrossparticipants and separated by brief filler tasks, as in the pre-vious studies. For the dependent measure part, participantswere instructed to answer a series of questions. The first setof questions asked them to indicate whether they hadadopted the following innovative products: Nintendo Wiigame console, Apple iPhone, Internet telephone (VOIP),PlayStation 3 game console, a GPS navigation device, anApple MacBook computer, satellite radio, a plasma televi-sion, high-definition television service, a Blu-ray DVDplayer, the Netflix movie rental service, or teeth whitening

(1996) provide details regarding these indexes. Satisfactorymodel fits are indicated by RMSEA and SRMR values lessthan or equal to .08 and NNFI and CFI values of at least .90.To evaluate stability, we summarized each scale using the

same procedures as in Study 1. We then calculated the cor-relations or stability coefficients between the scales fromthe first and second measurements. Higher stability coeffi-cients indicate greater scale stability (test–retest reliability).

Results

Replication of convergence results. We calculated thebivariate correlations between the subscales at each meas-urement and obtained similar results to those in Study 1,including the low to moderate correlations among the dif-ferent measures. We provide the correlation matrices and adiscussion of these results in the Web Appendix (http://www.marketingpower.com/jmroct10).Homogeneity of regulatory focus measures. Table 4 pro-

vides the goodness-of-fit measures for the CFAs for theRFQ, BIS/BAS, and Lockwood scales (recall that the selvesand self-guide strength measures are not appropriate for thisanalysis), completed across both sessions. The measurementmodels fit the scale measures reasonably well in all sixcases, indicating that the three scales have adequate homo-geneity. The phi coefficients (measures of correlation, attenu-ated for measurement error) between the promotion and theprevention factors in Table 4 also indicate that the promo-tion and prevention factors of the RFQ and BIS/BAS scaleshave moderately positive correlations, whereas the Lock-wood scale factors are uncorrelated. These results are con-sistent with an orthogonal, two-dimensional conceptualiza-tion of the regulatory orientations (Higgins 2002).Stability of regulatory focus measures. The last two rows

of Table 4 include the stability coefficients (and, therefore,the test–retest reliability), which are the highest for the RFQmeasures (.69 and .75 for promotion and prevention sub-scales, respectively), followed by the Lockwood (.67 and.62) and BIS/BAS (.55 and .69) scales. The selves measuresreveal moderate coefficients (.47 and .40), and the self-guide strength measures have the lowest stability coeffi-cients (.14 and .25) across time intervals. Therefore, Study2 provides evidence of adequate homogeneity and stability

Table 4STUDY 2: GOODNESS-OF-FIT MEASURES FOR CFAs OF RFQ, BIS/BAS, AND LOCKWOOD SCALES AND STABILITY COEFFICIENTS

Goodness-of-Fit At Time 1 At Time 2

Measures RFQ Scale BIS/BAS Scale Lockwood Scale RFQ Scale BIS/BAS Scale Lockwood Scale

χ2 χ2(43) = 75 χ2(53) = 133 χ2(134) = 414 χ2(43) = 99 χ2(53) = 170 χ2(134) = 316p = .003 p = .007 p ≈ .00 p ≈ .00 p ≈ .00 p ≈ .00

RMSEA .06 .08 .09 .08 .05 .05SRMR .06 .06 .08 .07 .09 .05NNFI .93 .88 .91 .91 .87 .97CFI .92 .91 .87 .90 .89 .98Φ-valuea .48 .29 .02 .30 .42 .04

Stability Coefficients

RFQ BIS/BAS Lockwood Selves Strength

Promotion subscale .69** .55** .67** .47** .14*Prevention subscale .75** .69** .62** .40** .25

*p < .05.**p < .001.aΦ-value provides the correlation between the promotion factor and the prevention factor in the CFA model, attenuated for measurement error.



strips. The first dependent measure was the sum of innova-tive products adopted by each participant.3 We reasoned thata promotion focus should be a positive predictor and pre-vention focus should be a negative predictor of new productadoption (Herzenstein, Posavac, and Brakus 2007).The second dependent measure involved the respondent’s

selection of a job. Participants evaluated two jobs: Job Aoffered a high salary with low job security, and Job B hadan average salary with high job security. Participants pro-vided their relative preference for the jobs on a scale from–5 (“I strongly prefer Job A”) to +5 (“I strongly prefer JobB”). In line with prior research, we expected a promotionfocus to be negatively related and a prevention focus to bepositively related to a greater preference for Job B (Phamand Higgins 2005; Zhou and Pham 2004).The third dependent variable involved making an invest-

ment decision. Participants were told to “Imagine that youreceived $25,000 as an inheritance recently that is availablefor investment.” They had two investment opportunities toinvest this money: “(1) A business venture that has a 70%chance of earning an 18% return and a 30% chance of los-ing 11% in a year, and (2) a mutual fund that has a 90%chance of earning a 10% return and a 10% chance of losing5% in a year.” Participants indicated how much of the$25,000 they would invest in the risky business venture.The latter two dependent variables were based on recent

findings from Pham and Avnet (2004), who found that forpersuasive communications, a promotion focus increasesconsumers’ reliance on subjective affective responses to theadvertising relative to the message’s substance; the oppositeoccurs for a prevention focus. We employed an approachsimilar to Pham and Avnet’s Study 4 and asked participantsto read and evaluate three print PSAs, one at a time, accord-ing to their affective responses, substantive assessments,and the advertisement’s persuasive impact.4 We regressedpersuasive impact on affective response, substantive assess-ment, promotion and prevention focus, and the interactionterms to test the differential impacts of promotion and pre-vention focus.

Results

We summarized each scale in the same way as in the pre-vious studies. For this study, we also reverse-scored theselves questionnaires and response latencies to make theresults of the analyses easier to compare across the fiveregulatory focus scales. The interscale correlations, which

replicated our prior findings, appear in the Web Appendix(http://www.marketingpower.com/jmroct10).To examine the scales’ predictive validity, we analyzed

the dependent variables in two ways. First, we ran separateregressions for each dependent variable on each chronicpromotion and prevention focus subscale. The results, in thetop panel of Table 5, show considerable variability in thepredictive power of the different scales. At least one compo-nent (i.e., either promotion or prevention) of the RFQ pre-dicted four of the five dependent variables significantly. Forthe assessment of the persuasive impact of the PSAs, we ranseparate regressions for promotion and prevention focus.The regression equation for the former was as follows:

(1) Persuasive impact = β0 + βAffectX1 + βSubstanceX2

+ βAffect × SubstanceX3 + βPromX4

+ βProm ×AffectX5 + βProm × SubstanceX6.

Second, we ran a similar regression for prevention focus.The key coefficients that explain the differential impact ofthe consumer’s regulatory orientations are those of the inter-actions between each regulatory focus and affectiveresponse and substantive assessment. According to Table 5,both coefficients were significant for the promotion and pre-vention regressions for the RFQ (marginally, at the p < .10level for the prevention focus × substantive content inter-action). Only the inheritance amount allocated to the riskyventure investment was not predicted successfully by eitherRFQ subscale.The predictive performance of the other four scales was

markedly inferior to that of the RFQ. The BIS/BAS and theself-guide strength scales predicted two of the five variablessignificantly. The prevention subscale of the BIS/BAS pre-dicted the allocation of inheritance and relative preferencefor jobs, whereas the self-guide strength scales were effec-tive in explaining the differential effects of promotion andprevention on the persuasive impact of the PSA advertise-ments. The selves measures only predicted one of the fiveoutcomes (relative preference for jobs), as did the Lock-wood scale (allocation of inheritance) in our study. Exceptfor the interaction regressions, the promotion and preven-tion coefficients were not simultaneously significant in theregressions for any of the five scales.5

STUDY 3B

Collecting multiple measurements of the same constructat the same time provides the best opportunity to demon-strate correlations between measures. The results reportedin the studies thus far (and in the Web Appendix, http://www.marketingpower.com/jmroct10) therefore provideconservative tests of the relationships among the variousscales. However, to reduce concerns about the drawbacks ofadministering the regulatory focus scales within the sameexperimental session, we conducted Study 3b, in whicheach measurement was spaced over four different sessions


3A pretest conducted with 166 participants from the same subject poolindicated that current adoption rates for these products ranged from 23.6%for the Netflix movie rental service to 71.5% for the Wii game console. Inmost cases, adoption rates ranged from 40% to 60%. The adoption rates forthe study sample appear in the WebAppendix (http://www.marketingpower.com/jmroct10).4For each advertisement, participants (1) reported their subjective affec-

tive responses using three seven-point measures anchored with “catchy/notcatchy,” “appeals/doesn’t appeal to me,” and “excites/doesn’t excite me”(α = .88); (2) provided their substantive assessments of the advertisementswith seven-point measures anchored with “gives me/doesn’t give me addi-tional information about [issue],” “explains/doesn’t explain the linkbetween [issue and its consequences],” and “stimulates/doesn’t stimulatemy thoughts about [issue]” (α = .81); and (3) rated the advertisement’s per-suasive impact using three measures anchored with “influences/doesn’tinfluence my opinion about [issue],” “changed/didn’t change my attitudetoward [issue],” and “the ad will influence my/other people’s [issue]habits” (α = .85).

5We also ran the analysis for Studies 3a and 3b using the methodologicalapproach employed by many regulatory focus researchers, which classifiesparticipants as either promotion or prevention focused. The results largelyremained the same, but minor differences emerged, as we detail in the WebAppendix (http://www.marketingpower.com/jmroct10).








in the course of a three-month period. We again examinedscale convergence.Study 3b also had a second objective: to further evaluate

the predictive validity of the scales using two additionaldependent measures. In a final session, we (1) evaluatedparticipants’ creativity by having them generate creative andunusual uses for a brick (Friedman and Förster 2001) and(2) obtained the size of their consideration set (i.e., theoptions they seriously considered) by soliciting the numberof movies they considered before selecting movie rentals(Pham and Change 2008). Prior research has shown thatthese measures are influenced by a person’s regulatoryorientations.

Method

Study participants were undergraduate students enrolledin a multiple-phase study involving several visits to a behav-ioral laboratory throughout the duration of a semester inexchange for course credit. Although 205 participants beganthis five-phase study, the inherent difficulties of getting mul-tiple repeat appearances from the same participants reducedthe sample size to 180 participants (87.8%). The analysesreported herein are based on this sample. The primary pur-pose of the study was unrelated to the current research;however, from an administration perspective, the studystructure provided an excellent opportunity to assess regula-tory focus measures over time.Accordingly, the first task in each of the first four ses-

sions was to complete one of the measures. Because of the

difficulty of tracking, the order in which the scales were col-lected was the same for all participants, but the two- to four-week separation between each scale made this point a mini-mal concern. Scales were administered in the followingorder: BIS/BAS, RFQ, Lockwood, and selves questionnaire.Response latencies for the self-guide strength measures wereagain recorded in conjunction with the selves questionnaire.In the fifth and final study phase, participants completed

two dependent measures. For the creativity task, they gener-ated as many unusual uses for a brick as possible, excludingordinary or impossible uses. They received examples ofunusual uses for paper clips to help them understand thetask. The responses were analyzed by a coder, who evalu-ated their answers for suitability and counted the number ofcreative uses provided by each participant. In the secondtask, participants were told that the researchers were inter-ested in their movie preferences. They were asked to visitthe Netflix Web site and to spend time looking at the avail-able selections, with the goal of choosing two movies torent. After they finished the browsing task, they listed all themovies they seriously considered renting and then providedtheir choices. To enhance involvement in the task, two par-ticipants were randomly selected to receive their chosenDVD. The participants’ consideration set (i.e., the numberof movies they seriously considered) was the seconddependent measure in the analysis. Our reasoning, follow-ing Pham and Chang (2008), was that a promotion focusshould be a positive predictor and a prevention focus should

Table 5STUDIES 3A AND 3B: REGRESSIONS OF OUTCOMES ON REGULATORY FOCUS MEASURES

Chronic Regulatory Focus Measure RFQ BIS/BAS Selves Strength Lockwood

Study 3a (N = 367)Number of Products AdoptedPromotion .03 .05 –.06 .04 .06Prevention –.13** –.02 –.04 –.01 .07

Relative Preference for JobsPromotion .02 –.04 –.12* .05 –.01Prevention .19*** .13** .02 –.09 .04

Amount of Inheritance Invested inBusiness VenturePromotion –.01 .09 .02 .07 .11*Prevention –.02 –.14* –.04 –.01 –.07

Persuasive Impact of Advertisement:Interaction with AffectPromotion .30* .15 .05 .87** .26†Prevention –.23* –.08 –.08 –.48* –.10

Persuasive Impact of Advertisement:Interaction with SubstancePromotion –.22* –.06 –.04 –.51* .00Prevention .17† .11 .05 .54* –.01

Study 3b (N = 180)Number of Unusual Uses for a BrickPromotion .03 .16* .05 .12 .14*Prevention –.07 –.08 –.02 –.12 –.16*

Number of Movies ConsideredPromotion .16* .03 .01 .12 .01Prevention –.08 –.08 –.15* –.26** –.02

*p < .05.**p < .01.***p < .001.†p < .10.Notes: All values reported are standardized regression coefficients.

be a negative predictor of consideration set size (see alsoPham and Higgins 2005).

Results

We summarized the scales in the same way as in Study3a. The results largely replicate the findings of the previousstudies by revealing low correlations between the subscaleswhen they were administered weeks apart, which strength-ens the support for a lack of convergence.As we report in the bottom of Table 5, the regressions

reveal that the number of creative uses for a brick is pre-dicted significantly by two scales: the BIS/BAS scale, forwhich the coefficient is negative for the BIS, and the Lock-wood scale, in which the promotion focus scale has a posi-tive coefficient and the prevention focus scale has a negativecoefficient. The RFQ, selves questionnaire, and self-guidestrength measure do not predict respondents’ creativity.However, the three scales are all successful in predicting thesize of the movie consideration set provided by the respon-dents, whereas the BIS/BAS and Lockwood scales are not.

Discussion

Studies 3a and 3b confirm the results from our previousstudies regarding low convergence among the measures.Furthermore, Study 3b provides a strict test of convergenceby spacing the measures over time, thereby addressingpotential concerns about measure collection during thesame experimental session. Most important, the two studiesassess the predictive validity of the five measures using aslate of dependent measures suggested by previousresearch. Such a direct comparison of the measures isrevealing and shows considerable variation in the measures’performance. Across the studies, the results of the assess-ment indicate that the RFQ performs the best in predicting

outcomes. That is, at least one dimension of the RFQ sig-nificantly predicted five of the seven outcomes examined inTable 6. The self-guide strength measure predicted four out-comes, the BIS/BAS and Lockwood scales were each ableto predict three outcomes, and the selves questionnaire onlypredicted two of the seven outcomes.Notably, both subscales were rarely significant predictors

at the same time. Coupled with the low to moderate correla-tions between the subscales of these measures, this findingprovides a further glimpse into the complexity of the opera-tions of people’s regulatory orientations. The pattern of find-ings suggests the need for more careful theory development toaccount for the orthogonal nature of the two regulatory orien-tations. In many cases, it may not be sufficient to hypothesizethat if a promotion focus operates one way, a prevention focuswill operate in the opposite way in terms of influencing agiven outcome. Again, we refer to our discussion of represen-tativeness in theWebAppendix (http://www.marketingpower.com/jmroct10) for additional analyses of the theoreticalunderpinnings of regulatory focus and related measures.

GENERAL DISCUSSION

Regulatory focus theory has become extremely influen-tial in psychology, marketing, and beyond and will likelycontinue to have a strong presence in the future. Observinga variety of approaches to measure chronic regulatory focus,we attempted to evaluate, conceptually and empirically, fivecommonly used measurement approaches. Through theseefforts, we aimed to bring greater clarity to the key aspectsof regulatory focus theory itself but also deeper knowledgeof the extent to which these measures overlap, how wellthey perform on various psychometric criteria, and theirtheoretical strengths and weaknesses. Our goal was to pro-vide guidance to researchers regarding the appropriate use


Table 6SUMMARY OF MEASURE PERFORMANCE USING SIMMS AND WATSON’S (2007) CRITERIA

Regulatory FocusMeasure Representativeness

InternalConsistency Homogeneity Stability Predictive Validity

RFQ Best performing. Lacks itemsmeasuring emotional reactions, andpresent- or future-oriented items.

Adequate Adequate. The twosubscales havemoderate positivecorrelation.

Adequate. Bestperforming ofthe fivemeasures

Best performing of thefive measures. Five ofseven significantpredictions.

BIS/BAS scale Inadequate. Only approach-avoidanceissues covered in the items.

Adequate Adequate. The twosubscales havemoderate positivecorrelations.

Adequate Three of sevensignificant predictions.

Selves questionnaire Adequate in covering self-discrepancy.Concerns about using differencescores to represent regulatory focus

theory adequately.

Promotionsubscale isinadequate

Not applicable Inadequate Poorest performance ofthe five measures. Twoof seven significant

predictions.

Self-guide strengthmeasures

Adequate for unconscious andautomatic processing research

questions.

Adequate Not applicable Inadequate Four of sevensignificant predictions.

Lockwood scale Adequate. Could be used as analternative to the RFQ, especially tostudy academic performance and

achievement.

Adequate Adequate. The twoscales are uncorrelated

Adequate Three of sevensignificant predictions.

Composite regulatoryfocus scale

Proposed as a means of addressinglimitations of the RFQ. Includesemotional content and blends past-,present-, and future-oriented items.

Adequate Adequate Adequate Comparable to the RFQin performance. Four of

five significantpredictions (Study 3a

data only).




of these scales and their integration into a variety ofresearch questions involving regulatory focus theory. In thissection, we succinctly summarize our current findings andpresent recommendations for further research.

Key Findings

Our series of studies suggest a primary conclusion: Thevarious measurement approaches for chronic regulatoryfocus orientation are very different empirically, with littletheoretical or empirical overlap. Table 6 provides a sum-mary of the five measures’ performance according to Simmsand Watson’s (2007) criteria.As Table 6 shows, the RFQ emerges as the most suitable

measure for general purpose theory testing. Of the fivemeasures, it alone is adequate in internal consistency, homo-geneity, and stability, and it performs the best in terms ofpredictive validity and representativeness. Notably, it is theonly scale to distinguish explicitly between approach andavoidance in each regulatory focus and to encapsulate thekey tenets of regulatory focus theory.Although they have more weaknesses than the RFQ, the

other measures may have niche uses. The BIS/BAS scalecan be used to examine research questions involving emo-tion and when the subtleties of regulatory focus theory,beyond the approach–avoidance motivation distinction, arenot paramount. The selves questionnaire would be appropri-

ate to study questions about consumers’ self-discrepancies,and the self-guide strength measures are apt for unconsciousand automatic processing issues or when social desirabilitymay make an association-based assessment more accuratethan an explicit, attitude-based assessment. The Lockwoodscale seems designed to study questions of academic per-formance and achievement.

A Composite Scale of Chronic Regulatory Focus

Despite the RFQ’s superior comparative performance, asnoted in our analysis of its representativeness (see the WebAppendix, http://www.marketingpower.com/jmroct10), itsuffers from two limitations: (1) an absence of emotionalcontent in its items and (2) items that are entirely past ori-ented. Using the large amount of data collected in ourresearch endeavor, we offer insights regarding a compositescale that uses measures from the RFQ, BIS/BAS, andLockwood scales to address these limitations. We providethis composite scale as a useful alternative for researchersinterested in the impact of differences in chronic regulatoryfocus. It comprises ten items (five each for promotion andprevention focus) and includes four items from the RFQ(two promotion, two prevention), two items from theBIS/BAS scale (one promotion, one prevention), and fouritems from the Lockwood scale (two promotion, two pre-vention). Table 7 provides this composite chronic regulatory

Table 7COMPOSITE REGULATORY FOCUS SCALE

Promotion Focus (5 Measures)1. When it comes to achieving things that are important to me, I find that I don’t perform as well as I would ideally like to do. (R)a2. I feel like I have made progress toward being successful in my life.a3. When I see an opportunity for something I like, I get excited right away.b4. I frequently imagine how I will achieve my hopes and aspirations.c5. I see myself as someone who is primarily striving to reach my “ideal self”—to fulfill my hopes, wishes, and aspirations.c

Prevention Focus (5 Measures)1. I usually obeyed rules and regulations that were established by my parents.a, d2. Not being careful enough has gotten me into trouble at times. (R)a3. I worry about making mistakes.b4. I frequently think about how I can prevent failures in my life.c5. I see myself as someone who is primarily striving to become the self I “ought” to be—fulfill my duties, responsibilities and obligations.c

Psychometric Properties of the Composite Scale (Study 3a Sample, N = 367)Scale ReliabilityPromotion focus Cronbach’s α = .79 (ranges from .69 to .84 in other study samples)Prevention focus Cronbach’s α = .74 (ranges from .67 to .77 in other study samples)

CFAχ2(34) = 124.99, p < .001, RMSEA = .07, NNFI = .92, CFI = .91, Φ-value = .29

Test–Retest Reliability (Study 2 Sample, N = 197)Promotion focus coefficient of stability = .67***Prevention focus coefficient of stability = .64***

Predictive Validity for the Dependent Variables of Study 3a

Percentage of $25K Persuasive Impact Persuasive ImpactNumber of Innovative Relative Inheritance Invested of Advertisement: of Advertisement:Products Adopted Preference for Jobs in Risky Venture Promotion Regressions Prevention Regressions

βProm = .07 βProm = –.09* βProm = .00 βProm ×Affect = .33* βPrev ×Affect = –.16βPrev = –.11* βPrev = .11* βPrev = –.03 βProm × Substance = –.21* βPrev × Substance = .23**

*p < .05.**p < .01.***p < .001.aRFQ measure.bBIS/BAS scale measure.cLockwood scale measure.dThis item was reworded from the original question, “How often did you obey rules and regulations that were established by your parents?” so that the ten

items could all use with the same “strongly disagree” to “strongly agree” response scale.


focus scale, as well as the results of tests of its psychomet-ric properties.The items in the composite scale meet the following rep-

resentativeness criteria: (1) They cover the key tenets ofregulatory focus theory, including distinguishing betweenapproach and avoidance within each regulatory focus (RFQitems), describing key concepts such as ideal and oughtselves in the measures (Lockwood items), and measuringself-regulation for promotion and prevention focus; (2) theyinclude both cognitive and emotional measures (BIS/BASitems) to provide a broad-based assessment; and (3) theyuse items that are past, present, and future oriented. Conse-quently, from a scale representativeness perspective, thecomposite scale can overcome the limitations from whichthe RFQ, as well as the other scales, individually suffer.We also examined the psychometric properties of the

composite scale using the Study 3a sample to assess its reli-ability, homogeneity through CFA, and predictive validity.The Study 2 sample provided a means to assess test–retestreliability. In both samples, we assessed the five measuresat the same time in counterbalanced order, so that all meas-ures included in the composite scale were assessed duringthe same experimental session. As Table 7 shows, the com-posite scale performs well on all tests. The reliability of thepromotion focus subscale is .79 and that for the preventionfocus subscale is .74. A CFA reveals good fit to the two-factormodel and a phi coefficient of .29, indicating a moderatepositive correlation between the promotion and preventionfactors. The test–retest reliability analysis with the Study 2sample reveals a coefficient of stability of .67 for the pro-motion focus subscale and .64 for the prevention focus sub-scale. Finally, the predictive validity results are also strong,with significant results along at least one dimension for fourof the five dependent measures.It is worth noting that in both Studies 2 and 3a, the meas-

ures of the composite scale were dispersed within theexperimental session, in that they were collected along withtheir respective parent scale items. Therefore, the psycho-metric test results of the composite scale are consideredconservative; if all the measures are assessed consecutively,we expect even stronger reliability, stability, and test–retestreliability results. On the basis of these findings, we canconclude that the composite scale offers a viable alternativeto the RFQ for assessing consumers’ chronic regulatoryfocus, and we recommend its use in research that examinesthe impact of differences in regulatory focus.Our findings also provide some useful guidelines regard-

ing how to operationalize chronic regulatory focus meas-ures. When used as independent variables, researchers oftendiscretize chronic regulatory focus measures. Many studieshave used difference scores to classify participants as pro-motion or prevention focused, according to a median splitof the sample (e.g., Avnet and Higgins 2006; Louro, Pieters,and Zeelenberg 2005). Although this approach collapsespromotion and prevention focus into a single dimension andis convenient for exposition purposes, the current researchhighlights its limitations. The low to nonsignificant correla-tions we find between promotion and prevention focus(even within the same measurement approach) imply thatsuch divisions may be tenuous. Because there is little or norelationship between a person’s chronic promotion andchronic prevention orientation, we recommend that promo-

tion and prevention orientations should be considered sepa-rately. Another limitation of collapsing two scales that usedifference scores is that doing so goes against the very grainof regulatory focus theory, which postulates that the twoorientations are orthogonal. Moreover, there are severalmethodological concerns with discretizing the chronic regu-latory focus measures into high and low categories, such asa significant loss of statistical power and increased likeli-hood of confounding (Fitzsimons 2008).

Further Research and Conclusions

This research provides detailed theoretical and empiricalanalyses of the measurement approaches to regulatoryfocus. The results and actionable recommendations canguide researchers in their endeavors. We consider two gen-eral research areas particularly worthwhile: (1) the exami-nation of potential interactions between chronic promotionand prevention orientations, and (2) further investigations ofthe role of regulatory fit in decision making.First, we believe that studying the chronic orientations of

regulatory focus in conjunction with widely used situationalmanipulations is an important topic for research. Empiricalinsights that we did not report herein highlight the potentialfor interactions between these chronic and situationalinstances of regulatory focus. Researchers should stronglyconsider including our composite regulatory focus measure(or other measures best suited to their research purpose)when using regulatory focus manipulations, at least to con-trol for these differences. In addition, a careful evaluation ofmanipulation approaches would be useful to researchers try-ing to choose among the many alternatives available. Someexamples of potential alternatives include essay writing,self-reported reflections on past experiences, brand nameanagram solution scoring, descriptions of aspirations orimportant duties, varying types and attributes of productsused in evaluations (e.g., retirement investment accountsinduce a prevention focus; stock trading investmentaccounts induce a promotion focus; Zhou and Pham 2004),and decision frames that emphasize gains versus the avoid-ance of losses.Second, regulatory fit occurs when consumers use strate-

gies and engage in activities that are consistent with theirgoal orientation, and as a result, they experience heightenedmotivation and a sense that “it just feels right” (Aaker andLee 2006). This fit can be achieved through matches withtask framing and chronic regulatory focus or situationallymanipulated focus. To understand such fit effects fully andhow they lead to enhanced value perceptions and other posi-tive outcomes (e.g., Avnet and Higgins 2006; Wang and Lee2006; Zhou and Pham 2004), researchers should account fordifferences in chronic regulatory focus, as our research rec-ommends. Regulatory fit will continue to be an importanttheoretical perspective, and the inclusion of a chronic per-spective on regulatory focus is essential.In conclusion, we recommend that researchers carefully

consider the purposes of their research when selecting amethod to assess chronic differences in regulatory focus.Although both theoretically and empirically the RFQ (Hig-gins et al. 2001) is the most widely appropriate of the exist-ing measures, we also offer a composite measure as an alter-native that has broad applicability across research contextsand captures the fundamental theoretical components of



regulatory focus theory. The use of any measures shouldavoid collapsing promotion and prevention orientations intoa single dimension. We also recommend assessing chronicregulatory focus when using situational manipulations ofpromotion and prevention orientations.

REFERENCES

Aaker, Jennifer L. and Angela Y. Lee (2001), “‘I’ Seek Pleasuresand ‘We’ Avoid Pains: The Role of Self-Regulatory Goals inInformation Processing and Persuasion,” Journal of ConsumerResearch, 28 (June), 33–49.

——— and ——— (2006), “Understanding Regulatory Fit,” Jour-nal of Marketing Research, 38 (February), 15–19.

Avnet, Tamar and E. Tory Higgins (2006), “How Regulatory FitAffects Value in Consumer Choices and Opinions,” Journal ofMarketing Research, 43 (February), 1–10.

Bentler, Peter M. (1990), “Comparative Fit Indexes in StructuralModels,” Psychological Bulletin, 107 (2), 238–46.

Brockner, Joel, Srikanth Paruchuri, Lorraine Chen Idson, and E.Tory Higgins (2002), “Regulatory Focus and the ProbabilityEstimates of Conjunctive and Disjunctive Events,” Organiza-tional Behavior and Human Decision Processes, 87 (January),5–24.

Campbell, Donald T. and Donald W. Fiske (1959), “Convergentand Discriminant Validity by the Multitrait–MultimethodMatrix,” Psychological Bulletin, 56 (2), 81–105.

Carver, Charles S. and Teri L. White (1994), “Behavioral Inhibi-tion, Behavioral Activation, and Affective Responses to Impend-ing Reward and Punishment: The BIS/BAS Scales,” Journal ofPersonality and Social Psychology, 67 (2), 319–33.

Cunningham, William A., Kristopher J. Preacher, and Mahzarin R.Banaji (2001), “Implicit Attitude Measures: Consistency, Stabil-ity, and Convergent Validity,” Psychological Science, 12 (2),163–70.

Dholakia, Utpal M., Mahesh Gopinath, and Richard P. Bagozzi(2005), “The Role of Desires in Sequential Impulsive Choices,”Organizational Behavior and Human Decision Processes, 98(2), 179–94.

———, ———, ———, and Rajan Nataraajan (2006), “The Roleof Regulatory Focus in the Experience and Self-Control ofDesire for Temptations,” Journal of Consumer Psychology, 16(2), 165–77.

Fitzsimons, Gavan J. (2008), “Death to Dichotomizing,” Journalof Consumer Research, 35 (1), 5–8.

Friedman, Ronald S. and Jens Förster (2001), “The Effects of Pro-motion and Prevention Cues on Creativity,” Journal of Person-ality and Social Psychology, 81 (6), 1001–1013.

Gray, Jeffrey A. (1990), “Brain Systems that Mediate Both Emo-tion and Cognition,” Cognition and Emotion, 4 (3), 268–88.

Herzenstein, Michal, Steven S. Posavac, and J. Jokso Brakus(2007), “Adoption of New and Really New Products: TheEffects of Self-Regulation Systems and Risk Salience,” Journalof Marketing Research, 44 (May), 251–60.

Higgins, E. Tory (1997), “Beyond Pleasure and Pain,” AmericanPsychologist, 52 (December), 1280–1300.

——— (1998), “Promotion and Prevention: Regulatory Focus as aMotivational Principle,” in Advances in Experimental Psycho-logy, Vol. 30, M.P. Zanna, ed. NewYork: Academic Press, 1–46.

——— (2000), “Does Personality Provide Unique Explanationsfor Behaviour? Personality as Cross-Person Variability in Gen-eral Principles,” European Journal of Personality, 14 (January),391–407.

——— (2002), “How Self-Regulation Creates Distinct Values:The Case of Promotion and Prevention Decision Making,” Jour-nal of Consumer Psychology, 12 (3), 177–92.

——, Ronald N. Bond, Ruth Klein, and Timothy Strauman (1986),“Self-Discrepancies and Emotional Vulnerability: How Magni-tude, Accessibility, and Type of Discrepancy Influence Affect,”Journal of Personality and Social Psychology, 51 (July), 5–15.

———, Ronald S. Friedman, Robert E. Harlow, Lorraine C. Idson,Ozlem N. Ayduk, and Amy Taylor (2001), “Achievement Orien-tations from Subjective Histories of Success: Promotion PrideVersus Prevention Pride,” European Journal of Social Psychol-ogy, 31 (1), 3–23.

———, James Shah, and Ronald Friedman (1997), “EmotionalResponses to Goal Attainment: Strength of Regulatory Focus asModerator,” Journal of Personality and Social Psychology, 72(March), 515–25.

——— and Israela Silberman (1998), “Development of Regula-tory Focus: Promotion and Prevention as Ways of Living,” inMotivation and Self-Regulation Across the Life Span, JuttaHeckhausen and Carol S. Dweck, eds. Cambridge, UK: Cam-bridge University Press, 78–113.

——— and Scott Spiegel (2004), “Promotion and PreventionStrategies for Self-Regulation: A Motivated Cognition Perspec-tive,” in Handbook of Self-Regulation: Research, Theory, andApplications, Roy F. Baumeister and Kathleen D. Vohs, eds.NewYork: Guilford Press, 171–87.

Hong, Jiewen andAngela Lee (2008), “Be Fit and Be Strong: Mas-tering Self-Regulation Through Regulatory Fit,” Journal ofConsumer Research, 34 (February), 682–95.

Jain, Shailendra Pratap, Charles Lindsey, Nidhi Agrawal, andDurairaj Maheswaran (2007), “For Better or For Worse?Valenced Comparative Frames and Regulatory Focus,” Journalof Consumer Research, 34 (1), 57–65.

Jöreskog, Karl and Dag Sörböm (1999), LISREL 8: User’s Refer-ence Guide, 2d ed. Chicago: Scientific Software International.

Lee, Angela Y. and Jennifer L. Aaker (2004), “Bringing the Frameinto Focus: The Influence of Regulatory Fit on Processing Flu-ency and Persuasion,” Journal of Personality and Social Psy-chology, 86 (2), 205–218.

Lockwood, Penelope, Christian H. Jordon, and Ziva Kunda (2002),“Motivation by Positive or Negative Role Models: RegulatoryFocus Determines Who Will Best Inspire Us,” Journal of Per-sonality and Social Psychology, 83 (4), 854–64.

Louro, Maria J., Rik Pieters, and Marcel Zeelenberg (2005),“Negative Returns on Positive Emotions: The Influence of Prideand Self-Regulatory Goals on Repurchase Intentions,” Journalof Consumer Research, 31 (March), 833–40.

Marsh, Herbert W., John R. Balla, and Kit T. Hau (1996), “AnEvaluation of Incremental Fit Indices. A Clarification of Mathe-matical and Empirical Properties,” in Advanced StructuralEquation Modeling: Issues and Techniques, George A. Mar-coulides and Randall E. Schumacker, eds. Mahwah, NJ:Lawrence ErlbaumAssociates, 315–53.

Pedhazur, Elazar J. and Liora Pedhazur Schmelkin (1991), Meas-urement, Design, and Analysis: An Integrated Approach. NewYork: Lawrence ErlbaumAssociates.

Peter, J. Paul (1979), “Reliability: A Review of PsychometricBases and Recent Marketing Practices,” Journal of MarketingResearch, 16 (February), 6–17.

Pham, Michel Tuan and Tamar Avnet (2004), “Ideals and Oughtsand the Reliance on Affect Versus Substance in Persuasion,”Journal of Consumer Research, 30 (March), 503–518.

——— and Hannah Chang (2008), “Regulatory Focus and Regu-latory Fit in Consumer Search and Consideration of Alterna-tives,” working paper, Graduate School of Business, ColumbiaUniversity.

——— and E. Tory Higgins (2005), “Promotion and Prevention inConsumer Decision Making: The State of the Art of TheoreticalPropositions,” in Inside Consumption: Consumer Motives,Goals, and Desires, S. Ratneshwar and David G. Mick, eds.London: Routledge, 8–43.

Shah, James, E. Tory Higgins, and Ronald S. Friedman (1998),“Performance Incentives and Means: How Regulatory FocusInfluences Goal Attainment,” Journal of Personality and SocialPsychology, 74 (2), 285–93.

Simms, Leonard J. and David Watson (2007), “The Construct Vali-dation Approach to Personality Scale Construction,” in Hand-book of Research Methods in Personality Psychology, RichardW. Robins Jr., Chris Fraley, and Robert F. Krueger, eds. NewYork: Guilford Press, 240–58.

Wang, Jing and Angela Y. Lee (2006), “The Role of RegulatoryFocus in Preference Construction,” Journal of MarketingResearch, 43 (February), 28–38.

Yeo, Junsang and Jongwon Park (2006), “Effects of Parent Exten-sion Similarity and Self Regulatory Focus on Evaluations ofBrand Extensions,” Journal of Consumer Psychology, 16 (3),272–82.

Zhao, Guangzi and Cornelia Pechmann (2007), “The Impact ofRegulatory Focus on Adolescents’ Response to AntismokingAdvertising Campaigns,” Journal of Marketing Research, 44(November), 671–87.

Zhou, Rongrong and Michel Tuan Pham (2004), “Promotion andPrevention Across Mental Accounts: When Financial ProductsDictate Consumers’ Investment Goals,” Journal of ConsumerResearch, 31 (June), 125–35.


Copyright of Journal of Marketing Research (JMR) is the property of American Marketing Association and its

content may not be copied or emailed to multiple sites or posted to a listserv without the copyright holder's

express written permission. However, users may print, download, or email articles for individual use.

Documents

An Assessment of Chronic Regulatory Focus Measures