Consumer Satisfaction: Cognitive and Affective Dimensions

Narasimhan (Han) Srinivasan, University of Connecticut
[ to cite ]:
Narasimhan (Han) Srinivasan (1995) ,"Consumer Satisfaction: Cognitive and Affective Dimensions", in NA - Advances in Consumer Research Volume 22, eds. Frank R. Kardes and Mita Sujan, Provo, UT : Association for Consumer Research, Pages: 543-544.

Advances in Consumer Research Volume 22, 1995      Pages 543-544

CONSUMER SATISFACTION: COGNITIVE AND AFFECTIVE DIMENSIONS

Narasimhan (Han) Srinivasan, University of Connecticut

Overall, I liked the papers and the creativity and rigor of thinking apparent in each of the papers. First, let me deal with how the three papers appear to fit in the same session. Then, the papers will be commented upon individually.

Droge and Mackoy's paper, "Postconsumption Competition: The Effects of Choice and Non-Choice Alternatives on Satisfaction Formation," sets an appropriately broad framework, removing the isolation of the chosen brand to measure satisfaction, after any decision-making context. Rightly, it has been recognized that the literature has touched upon this aspect previously: for example, cognitive dissonance theory and regret theory talk about non-chosen alternatives impacting the chosen alternative. An example of neutralizing dissonance is the wide use of accepting competitor's promotional incentives, such as doubling/tripling coupons and other special deals. Where trial is not costly, e.g. a store having a return policy with no questions asked, perhaps, no decision is ever final and it is always possible to return the chosen alternative in favor of a non-chosen alternative.

Gurhan and Creyer's work, "Exploring Consumers' Interpretations of a Product Related Illness," fits well in this broad framework in the sense that possible courses of action that a consumer could have followed is considered in evaluating the degree of responsibility/blame for the consequences of the decision an individual makes. Had a different course of action been followed (i.e. a different alternative chosen), there would be less cognitive dissonance and no regret, perhaps. Re-evaluating the wisdom of one's choice with hindsight happens commonly, particularly when there are potential legal consequences of product liability. e.g. smoking, toy guns, etc.

Raman, Chattopadhyay and Hoyer's manuscript, "Do Consumers Seek Emotional Situations: The Need for Emotion Scale," adds to this big picture in expanding the decision making context itself from its cognitive orientation bias. There are additional dimensions of choice that have not been as heavily researched as cognition, such as emotions. These could be momentary, being salient only at the time of decision making, or may linger for long because the product has a lot of emotional overtones. e.g. decision-making process involving the choice of a potential spouse. Do the emotional overtones associated with non-chosen alternatives (assuming that they exist, of course) remain time-capsuled (or get etched in memory after cognitive distortion), so that an "old flame" never dies? Sometimes, people remarry the same person, like Elizabeth Taylor; sometimes, the intensity of emotion may not diminish but the valence may change, particularly in hotly contested separations/divorces. In the above examples, strong emotions are aroused and could dominate the cognitive dimension.

In the three papers, we are essentially dealing with the conceptual complexity of satisfaction. Is it stationary, due to the constraints of operational definition? Is it dynamic? For instance, what happens to my satisfaction when I buy a personal computer and better models at cheaper prices keep appearing on the market? When I spend a lot of time learning how to operate certain software and then something else comes along, do I feel less satisfied with my own learning? How much of it is attributable to one's self or to the external environment? Can satisfaction be sub-divided into several components: fit with the current need, sense of accomplishment in solving a complex task, length of usefulness of chosen alternative given uncertainty. When does something become "sour grapes" and remain so? Under what conditions do "failures become stepping stones to success?" i.e. present dissatisfaction turns out to have been a good thing to happen because we learnt tremendously from the situation.

Now, let me take the papers individually:

1) The Gurhan and Creyer paper is interesting and particularly useful for trial lawyers perhaps. It found that there is a variation in (1) the willingness to award money to victims and (2) the amount of blame attributed to firms for product liability.

Perhaps because I just talked to a friend of mine, who testified in federal court recently about monopolistic competition in a seemingly fragmented market, I don't understand fully the manipulation of market share: 70% versus 30% for high and low market share respectively. Does it not depend upon market structure? In a very competitive market, 30% could be the share of the market leader. It is possible that nobody else in the market has more than 30% individually because the concentration ratio is very low. On the other hand, 70% denotes market domination, of course. Perhaps, you had pre-tested these levels and have not provided the figures in the manuscript. How many people could have made (in)correct assumptions about the relevant market, in terms of its structure? Vignettes are useful, I agree, but when the cognitive processing of information may be different due to differential comprehension of the stimuli, then some error could creep in.

If Mrs. Watson had the time to look for her regular brand in another store but still chose to buy Bomex, which was on promotion, what about the possible interaction due to promotion? What if Mrs. Watson did not have the energy or the inclination or did not want to make the effort and is just plain lazy, as is being assumed? How far is the other store? Round the block or across town? Time alone is not the only relevant cost.

Both the items tapping at the blame assignment relating to the manufacturer appear to have face validity. However, it is surprising that the correlation of the two items is only 0.64. The index for event foreseeability, on the other hand, does not appear to have face validity and it is not surprising that the correlation is only 0.56. Under high time pressure particularly, when the regular brand is not available, it appears quite rational to buy an alternative on promotion. What does this have to do with foreseeability of allergic reaction?

The significant impact (p<0.10) of market share could have an alternate explanation. High market share brands generally come from very reputable companies, which are usually large and hence can afford to make large payments in damages. Suppose there is environmental damage caused by oil spills and the company involved is not Exxon, but a very small company. Wouldn't the damages being awarded change? Exceptionality of occurrence aside, ability to pay also figures in the respondent's intuitive thinking and judgment, I feel.

In the second experiment, where the manipulation of market share has been improved, the significance vanishes when examining the blame assigned to the manufacturer! The culprit is guilty, whether rich or poor. On the other hand, when examining the blame assigned to Mrs. Watson, market share becomes significant and the expected interaction with promotion shows up.

I'm just curious that there is no analysis or even mention of the amount of money that Mrs. Watson seemed to merit under different scenarios. You have the opportunity to use a ratio level dependent variable instead of just nominal data.

2) The Droge and Mackoy paper is a conceptual paper, having a lot of merit in that it brings in greater realism to the context under investigation. I believe that the context can even be enhanced when we consider "new" alternatives not currently available in the market but which are expected. When decisions can be reversed, such as product returns, alternatives will continue to exert competitive pressure. For example, when deciding to purchase a personal computer and/or software, would not the market entry of better and cheaper alternatives impact the (dis)satisfaction with my purchase? I'd like to see some empirical follow-up because it is an exciting new avenue.

3) The Raman, Chattopadhyay and Hoyer paper reveals a good awareness of the related literature. The methodology for scale refinement is pretty standard and appears to have been followed. The potential applicability of the scale in terms of predictive validity remains to be shown.

The conceptual definition of the Need for Emotion (NFE) scale is given as "the tendency or propensity for individuals to seek out emotional situations, enjoy emotional stimuli and exhibit a preference to use emotion in interacting with the world." In hindsight, it appears that this has not been accomplished, though an avoidance of emotion is being tapped.

Supposedly, NFE taps mainly into short term emotion. How "short" is short-term? What is its stability? One of the examples cited in the Droge and Mackoy paper is visiting a university which we rejected attending or seeing someone driving another car we considered. Residual emotions are a factor to contend with and perhaps the emotion scale may help in measuring the declining intensity over time.

The authors do a splendid job of spelling out that there are affect intensity differences across individuals and show awareness of different scales which are available in the literature: Larsen's Affect Intensity Measure (AIM), Allen and Hamsher's Test of Emotional Styles (TES) and Booth-Butterfield and Booth-Butterfield (1990)'s Affective Orientation Scale (AOS). However, the use of the literature could have been better.

The conceptual distinction between Larsen's Affect Intensity Measure (AIM) and the NFE is given as follows: While AIM measures the intensity of response to an affective stimulus, NFE taps the "tendency to seek out affective stimuli and enjoy these emotional situations." On the surface, it appears that AIM would be applicable widely, whereas the NFE would try to isolate the self-selection behavior of individuals who seek out and enjoy emotional stimuli. Under what conditions would one wish to use the NFE? Hopefully, when further work is done on criterion validity, such questions would be answered. Also, the persuasiveness is less than strong when it is stated that low AIM individuals could have high NFE scores and vice versa. Why wasn't AIM used in the study? An empirical testing could have answered this speculation.

As the authors recognize, another scale in the literature called the Test of Emotional Styles (TES) is very close conceptually because it includes "attitudes toward emotional experiences and expressions." Actually, the paper states that it is "conceptually identical to NFE." Unfortunately, while re-inventing the wheel, "none of the NFE items were imported from previous scales." Why not in this case, if there is conceptual identity? While there could be methodological deficiencies, isn't it possible to convert the scaling procedure to being Likert-type? I'd suggest that reworking the response format of TES could be a worthwhile exercise.

The Affective Orientation Scale which taps the "degree to which individuals are conscious of affective cues and use these cues to guide decision-making processes" is used in this study, though this scale appears to have a cognitive bias commonly found in consumer decision-making models. Yet, the correlation of the AOS and NFE was 0.69, consistent with a moderate correlation between NFE and NFC (0.46 and 0.31 in the first and second samples respectively). The use of a cognitive oriented measuring instrument could have made a contribution, I suppose.

Some questions on the analysis:

The first factor (the general factor) appeared to be the "need for emotion" dimension, whereas the second factor appeared to reflect a "generation of emotions" dimension. What exactly is this second factor and why was it discarded? Is it possible to isolate those items which loaded on both factors (48 items-20 items for first factor-10 items for second factor = 18 items?) and use these to form a scale which includes the need and generation of emotions, because it is these items which appear to correspond to the seeking out and enjoyment of emotions? Being a full information analysis, the number of factors in a factor analysis depends on the type and number of items you throw in. Hence, it is possible that the common items, when a factor analysis is done just with these items, will all load on a single factor and the Cronbach alpha will be high.

In the second analysis, the NFE and AOS items were factor analyzed and four factors emerged. Was this as expected? What is the correlation between NFE and AOS in the second sample? How were the four dimensions related in the oblique rotation? Information is missing about the relationship of NFE with the other three factors.

It was found that all the items of NFE were reverse coded items i.e. subjects indicated avoidance of emotional situations. If this denotes a tendency to self-proclaim being non-emotive on the part of the respondents, perhaps it would be useful to test the scale for social desirability bias in future research. The gender difference found in NFE could just be a reflection of social values.

REFERENCES

Droge, Cornelia and Robert D. Mackoy (1994), "Postconsumption Competition: The Effects of Choice and Non-Choice Alternatives on Satisfaction Formation," ACR, Boston.

Gurhan, Zeynep and Elizabeth H. Creyer (1994), "Exploring Consumers' Interpretations of a Product Related Illness," ACR, Boston.

Raman, Niranjan V., Prithviraj Chattopadhyay and Wayne D. Hoyer (1994), "Do Consumers Seek Emotional Situations: The Need for Emotion Scale," ACR, Boston.

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