Special Session Summary Developing Better Marketing Measure: in Search of Nomological Validity

Citation:

Heather Honea and Cristel Antonia Russell (2001) ,"Special Session Summary Developing Better Marketing Measure: in Search of Nomological Validity", in E - European Advances in Consumer Research Volume 5, eds. Andrea Groeppel-Klien and Frank-Rudolf Esch, Provo, UT : Association for Consumer Research, Pages: 193-194.

European Advances in Consumer Research Volume 5, 2001      Pages 193-194

SPECIAL SESSION SUMMARY

DEVELOPING BETTER MARKETING MEASURE: IN SEARCH OF NOMOLOGICAL VALIDITY

Heather Honea, San Diego State University, U.S.A.

Cristel Antonia Russell, San Diego State University, U.S.A.

This session addresses the issues of scale development, focusing on nomological validity testing as a means of developing better marketing measures. The first paper illustrates nomological validity issues that may emerge later in the "life" of a measure, suggesting that the context in which a measure is administered can compromise its validity. The other two papers explore remedies for such problems by conducting nomological validity tests at the scale development stage.

 

"USING NESTED MODELS TO EXAMINE PERCEPTIONS OF EASY-DIFFICULT"

Mark Leach, Loyola Marymount University

Michael Hennessy, The University of Pennsylvania

Martin Fishbein, The University of Pennsylvania

Leach, Hennessy, and Fishbein discuss nomological validity issues pertaining to the attitude construct. The purpose of their study is to investigate whether or not perceived difficulty items (i.e., easy-difficult) should be considered a measure of attitude or a measure of self-efficacy. The relationship between ease-difficulty and the concepts of attitude toward the act and self-efficacy are particularly interesting given the use of difficulty items in the measurement of both constructs. When investigating perceptions of difficulty, insight may be gained by paying attention to what "difficult" means to respondents in various behavioral situations. Fishbein and colleagues (Fishbein 1993; Chan & Fishbein 1993) poit out that there are difficult behaviors that are nonetheless under volitional control. Thus, there are likely to be difficult behaviors that people feel confident they can perform well. For these types of behaviors, how do respondents evaluate items asking them to rate their perception of difficulty? When perceived volitional control is low, asking whether this goal or behavior is easy or difficult may reflect one’s perception of control. When confidence or perceived competence is low, assessments of perceived difficulty may tap self-efficacy perceptions. When both perceptions of control and self-efficacy are high, then something that is perceived to be easy may be liked and something that is hard may be disliked. In this case, assessments of difficulty will reflect one’s attitude toward the behavior. This situation illustrates the possibility that the context of a study may impact the relationship of an indicant to the construct it is purported to measure.

A sequence of nested confirmatory factor analysis models (James, Mulaik & Brett 1982; Anderson & Gerbing, 1988) are employed to systematically test three alternative explanations: first, the easy- difficult item measures the attitude construct, second, it measures self-efficacy, third, it poorly discriminates between both constructs. The models are fitted using data from two large health intervention programs on condom use behavior. Their results indicate that perceptions of difficulty or ease can be highly related to either or both attitude and self-efficacy for different populations and behaviors. For male participants reporting condom use with their main sex partner, the easy-difficult item loads highly with other measures of attitude. For females (and to some extent males with occasional sex partners), the easy-difficult item appears to tap both attitude and self-efficacy dimensions. Females may be less likely to view condom use as a behavior and instead consider it a goal; particularly when measurement items specifically identify the behavior as getting a sex partner to use condoms rather than as using condoms. Thus, when evaluating a behavior that is considered a goal, responses to easy-difficult items appear to be partially based on feelings of confidence that the goal can be reached. In these situations, the use of easy-difficult type items is likely to confound measures of attitude with self-efficacy.

By showing that easy-difficult items act differently depending on the type of behavior investigated, this paper illustrates the existence of a moderator that shifts the relationship between an indicator and a construct. Specifically, this research demonstrates the moderating role of perceived volitional control in assessments of perceived difficulty, which in turn affects the construct tapped by the item. These findings illustrate that even a widely accepted theoretical construct is not exempt from measurement concerns, further highlighting the need to conduct appropriate nomological validity test as early possible in the scale development process.

 

"THE CONSUMPTION OF REGULAR TELEVISION PROGRAMMING: DEVELOPMENT AND VALIDATION OF THE CONNECTEDNESS SCALE"

Cristel A. Russell, San Diego State University

Andrew Norman, Iowa State University

Russell and Norman present nomological validity tests of their television connectedness scale. The connectedness scale is a multi-factor measure that captures the complex set of relationships that individuals develop with "their" television shows. Their presentation illustrates the need to complement traditional reliability and validity tests with tests of nomological validity by checking the psychometric properties of individual difference measures in multiple administration contexts.

As a follow-up to the Russell and Puto (1999) exploratory paper which determined that individuals differ in the way they relate to television shows, a scale was developed to assess individuals’ level of connectedness with television shows. The process of developing the connectedness scale was two-phased. In the first phase, a battery of 67 items generated in three focus groups was administered to 175 student subjects, which generated data for an exploratory factor analysis of the connectedness measure. In the second phase, additional data from other student samples (N=265) were collected to confirm the factor structure that emerged in Phase 1. This process resulted in an internally valid and reliable 24-item measure of connectedness.

Connectedness thus emerged as a multi-faceted construct capturing the complex set of relationships that individuals develop with "their" television shows. While some factors reflect the vertical relationship formed between an individual and a television program; others illustrate the horizontal linkages among audience members (networks of co-consumers); and yet others capture the connections formed between the individual and the characters in the show. The vertical dimension is of particular relevance to nomological validity testing as it is theoretically related to the construct of involvement. It is comprised of four factors. The "Must see" factor indicates how dedicated an individual is to watching the program every time it is on. The "Quality" factor refers to the inherent quality of the entertainment experience. The "Escape" factor is the cathartic element of television shows, which allows viewers to relax and forget about their problems. Finally, the "Paraphernalia" factor describes the fact that people collect items that are related to the programs.

In an attempt to investigate nomological validity, the relationships between connectedness and established constructs were tested across different scale administration contexts. In particular, a series of studies included the administration of the connectedness scale and Zaichowsky’s Enduring Involvement scale (1985) in a variety of applications and contexts. In line with predictions that involvement is more situational than connectedness, the results demonstrate that viewing a segment of the TV show prior to the administration of the measures increases the level of correlation between connectedness and involvement scores, thereby suggesting that prompting the show moderates the hypothesized relationship between the two constructs. The authors conclude by emphasizing the importance of being able to empirically discriminate between the construct of involvement and that of connectedness. They argue that these nomological validity tests help support the claim that connectedness extends beyond the mere viewing experience and depicts a show’s contribution to its viewers’ self-definition.

 

"'I GOT A DEAL !?’ DEFINING THE AFFECTIVE DIMENSION OF PROMOTIONAL PURCHASES"

Darren W. Dahl, University of Manitoba

Heather Honea, San Diego State University

This article examines the psychometric properties of consumer affect that is produced by a promotional offer: e.g., price discount, coupon, free gift, etc. Building from the existing sales promotion and emotion literature, we propose a general model that identifies the affective response resulting from a promotional purchase. The proposed model is operationalized and tested with a new scale instrument that measures affective response to promotion. Classic scale development procedures, as well as specific efforts in nomological validation, are undertaken to test the properties of the scale instrument and refine it through its use in a variety of promotional settings. The scale instrument, the Promotional Affect Thermometer (PAT), can be used by both researchers and marketing practitioners to examine the affect elicited by a promotional offer.

We propose that promotional affect is a multi-dimensional construct with both positive and negative dimensions, ranging in their degree of specificity. Thus, while there are generic positive and generic negative feelings of happiness and sadness due to a promotional purchase, there are also more specific valenced feelings along these positive and negative dimensions. The positive dimensions identified are feelings of happiness and activation, feelings of being a smart shopper, feelings of gratitude, and feelings of being lucky. The negative dimensions identified are feeling of unhappiness, feelings of self-consciousness, feelings of being deceived and annoyed, and feelings of uncertainty.

Three phases of data collection are reported that systematically examine the content validity, criterion validity, and construct validity of the proposed scale instrument. Results indicate that PAT is a good measurement tool of the affect produced by a promotion. In terms of content (face) validity, variables such as the type of promotion and product category impact the sub-dimensions of the affect scale in the expected direction. In addition, affect is shown to mediate the route to promotion evaluation and purchase intentions. The scale is also able to discriminate between people who have positive, negative, and neutral thoughts regarding the offer as well as discriminate between those individuals who choose to buy and not to buy a product on promotion; all evidence of criterion validity.

While the issues of content and criterion validity are important to the acceptance of PAT as a measurement tool, the construct validation process undertaken in this research serves to solidify the value of this scale. Drawing on the proposed affect response model the paper focuses on certain components of construct validation, specifically trait and nomological validity.

This research contributes towards a better theoretical understanding of the consumer affect resulting from promotional purchases, and provides a readily usable measurement instrument for theoretical researchers and marketing practitioners. It contributes to the sales promotion literature by suggesting that different types of promotions, products, consumers, and contexts arouse differential levels and dimensions of affect. A better understanding of the antecedents of promotion evaluation and purchase intentions can help managers to better design and implement their promotional efforts. In addition, this paper contributes to the development of "better marketing measures" and continues to advance the general process of scale development by providing concrete examples of how individual difference variables and experimental manipulation can be used to support the construct validation process before a scale is ever put into use.

REFERENCES

Anderson J. & Gerbing D. (1988), "Structural Equation Modeling In Practice: A Review And Recommended Two Step Approach," Psychological Bulletin, 103 (3), 411-423.

Chan, D. and Martin Fishbein (1993), " Determinants Of College Women’s Intentions To Tell Their Partners To Use Condoms," Journal of Applied Social Psychology, 23, 1455-1470.

Fishbein, Martin (1993), Introduction. In D. J. Terry, C. Gallois, and M. McCamish (Eds.), The Theory of Reasoned Action: Its Application to AIDS-Preventative Behavior, Oxford: Pergamon, 15-25.

Jacoby, Jacob (1978), "Consumer Research: A State of the Art Review," Journal of Marketing, 42 (April), 87-96.

James L., Mulaik S., & Brett J. (1982), Causal Analysis: Assumptions, Models, and Data, Sage: Beverly Hills.

Russell, Cristel A. and Christopher P. Puto (1999), "Rethinking Television Audience Measures: An Exploration Into The Construct Of Audience Connectedness," Marketing Letters, 10 (4), 387-401.

Zaichkowsky, Judith Lynne (1985), "Measuring the Involvement Construct," Journal of Consumer Research, 12 (December), 341-352.

----------------------------------------