An Experimental Investigation of Situational Effects on Risk Perception

Mark Vincent, Oklahoma State University
William G. Zikmund, Oklahoma State University
ABSTRACT - This exploratory study was designed to determine the effects of different buying situations on several dimensions of perceived risk. Statistical analysis of generated data indicated significant differences in responses to selected risk dimensions due to both situational main effects and situational-interaction effects. These results indicate that buying situations should be considered as partial determinants of perceived risk.
[ to cite ]:
Mark Vincent and William G. Zikmund (1976) ,"An Experimental Investigation of Situational Effects on Risk Perception", in NA - Advances in Consumer Research Volume 03, eds. Beverlee B. Anderson, Cincinnati, OH : Association for Consumer Research, Pages: 125-129.

Advances in Consumer Research Volume 3, 1976      Pages 125-129

AN EXPERIMENTAL INVESTIGATION OF SITUATIONAL EFFECTS ON RISK PERCEPTION

Mark Vincent, Oklahoma State University

William G. Zikmund, Oklahoma State University

ABSTRACT -

This exploratory study was designed to determine the effects of different buying situations on several dimensions of perceived risk. Statistical analysis of generated data indicated significant differences in responses to selected risk dimensions due to both situational main effects and situational-interaction effects. These results indicate that buying situations should be considered as partial determinants of perceived risk.

Marketing researchers frequently have suggested that their results have been influenced by "situation-specific" variables. While it has long been held that situational factors would aid substantially in explaining the variability which exists in buyer behavior, only recently have consumer psychologists attempted to systematically investigate the influence situational variables have on brand or product choice.

Most recently, Belk (1974) found that situational variables had a significant influence on consumer preference toward meat and snack product classes. Lutz and Kakkar (1975) applied the Mehrabian-Russell Emotional Mediators Instrument to analyze for psychological differences in the same buying situations which Belk considered. Their results basically confirmed the previous results of situational effects and situational-product interaction effects. In Sandell's (1968) research, situational factors such as "for breakfast," "with lunch," "at a party," explained a larger portion of total variance and beverage preference behavior than did Brand. Miller (1975) found a situational multi-attribute attitude model outperformed the traditional (non-situational) model when used to predict overall choice. Thus, the evidence is relatively clear that situational factors, at least marginally, improve prediction of brand choice for specific product categories.

By introducing situational variables into the research design, consumer psychologists should be able to account for more variance than by ignoring or controlling for them (Belk, 1975). However, Lutz and Kakkar hold the percent explained by situation is likely to be of the same magnitude as personality, i.e., "the situation in and of itself is not a powerful predictor of consumer behavior."

The objective of this research is to isolate which cognitive variables situational influence mediates. More specifically, rather than studying "situation" as an isolated construct to explain consumer behavior, our focus is to examine the relationship between situation factors and perceived risk. Hopefully, establishing experimental verification for the theoretical notion that buying situations may have a systematic and demonstratable effect on dimensions of perceived risk.

THE BUYING SITUATION

There is no general agreement concerning the definition of a situation. (See Lutz and Kakkar 1975 for a review). A locus of time and space is one of the simplest definitions. Belk (1975) argues that a rich and meaningful use of the concept "situation" includes both the antecedent conditions for the momentary individual states which the person brings to a given time and place and the physical features which he finds there. An objective situation may be externally verified without measuring intrapersonal factors.

"... objective descriptions may include the existence of external facts and events which bear upon current behavior even though they are not themselves physically a part of that situation."

"For example, if it can be determined that the stimulus object ... is an item to be served at an upcoming dinner party for the new boss, these objective descriptions can be included in the situational specification directly rather than attempt to measure whether or not the situation is regarded to be threatening by the individual." (Belk, 1975, p. 429)

However, several researchers tend to agree that there is a need to adopt a subjective view of this situation in order to understand how situational factors influence consumer behavior. Holding that individuals transform situational input into behavioral output, Lutz and Kakkar offer a definition that includes explicit consideration of perceptual processes:

The psychological situation, ... may be defined as an individual's internal response to, or interpretation of, all factors particular to a time and place of observation which are not stable intra-individual characteristics or stable environmental characteristics, and which have a demonstratable and systematic effect on the individual's psychological processes and/or his overt behavior.

This definition complements rather than competes with Belk's viewpoint and is adopted for purposes of this research.

PERCEIVED RISK AND BUYING SITUATIONS

Few addresses to the American Marketing Association have stimulated as much empirical research as Bauer's 1960 paper on perceived risk. Perceived risk, a subjective evaluation process, has been associated with new product adoption, brand and store loyalty, mode of shopping as well as word of mouth communications and opinion leader- ship. (For an excellent review, see Ross, 1974). Studies by Cox (1964) and Spence, Engle, and Blackwell (1970) indicated that perceived risk is affected by factors other than product. Each of these studies found that the place of shopping, a situational factor, influenced perceived risk. Each, however, was concerned with the physical surroundings of buying situation, "where to buy." Buying situations may also be concerned with dimensions such as the interpersonal surroundings (group opinion), time factors, goal structure (cf. Miller, 1975 Lutz and Kakkar, 1975) and may have a profound effect upon perceived risk associated with the product. It is therefore, the intention of this study to test if specific buying situations, not related to physical surroundings, are related to the various dimensions of perceived risk.

PERCEIVED RISK

The initial empirical studies related to perceived risk utilized the uni-dimensional measures of overall risk associated with certain products or brands. (e.g., Arndt, 1967, Cunningham, 1965). More recently consumer behavior theorists have recognized the need to identify risk as a multi-dimensional phenomena. General risk has been variously identified into dimensions such as social, performance, physical, and financial risk (e.g., Roselius, 1971; Perry and Hamm, 1969; Zikmund and Scott, 1973). This research views perceived risk as a multidimensional phenomena and utilizes the various types of risk as dependent variables.

EXPERIMENTAL METHODOLOGY

A two-by-two factorial experiment was designed to measure the influence of situational factors and store on perceived risk. The experiment was a factorial design with repeated measures on store factors.

Situation * Store

Figure 1 is a representation of this design. In Situation 1, respondents were asked to consider the purchase of an electric knife for home use. In Situation 2 they were asked to consider the purchase of an electric knife to give as a wedding present to a close friend. Store 1 was identified as Montgomery Ward. Store 2 was identified as Gibson's Discount Store.

FIGURE 1

FACTORIAL DESIGN OF EXPERIMENT

Each member of the sample taken was randomly assigned to a situational treatment and responded to both store types. While extensive pretesting indicated that differences may exist across these situations, the use of these two stores was based upon findings by Hisrich and others (1972) that differences in perceived risk may exist in store selection.

Stores were selected on the basis of availability to members of the sample rather than on any specific a priori distinction. In other words, two general department stores were selected, rather than identifying the stores as "an unknown store" versus a "well known store" or some other type of manipulation.

The order of appearance of the two types of stores was randomized within the measurement instrument. The home usage situation and the gift situation were intended to elicit differences which may exist between factors considered in buying for one's self and buying for another person i.e., the situation dimensions, goal structure and interpersonal opinion were manipulated (cf. Lutz and Kakkar, 1975 p.3).

This experiment was structured to provide information concerning situational effects, store type effects, and the interaction effects between store type and situation on perceived risk. It is concerned with a relatively inexpensive consumer durable, a home usage situation, a wedding gift situation, and two types of stores which are both recognized by and available to the intended sample. In this respect, the experiment could be considered to realistically reflect the realm of experience of the intended sample, i.e., it meets the requirement that the situation be encountered frequently (Miller, 1975). Regardless of whether an electric knife had ever been purchased for either of these reasons and/ or from either of these store types, it is not unreasonable to assume that such considerations may have been made or could be made.

The selection of situations for both experiments was based upon a pretest of thirty-four respondents, seventeen from Ponca City, Oklahoma, and seventeen from Midwest City, Oklahoma, which indicated that significant situational effects might exist.

Hypothesis

Just as individual risk dimensions are expected to influence the usage of selected information sources, we expected that certain types of buying situations would be more prone to influence specific dimensions of perceived risk. For example, situations varying in degrees of interpersonal relations would be expected to mediate the amount of social risk perceived by consumers. The general null hypothesis is as follows: There is no relationship between the type of buying situation and the dimensions of perceived risk. Another hypothesis related to stores is as follows: There is no relationship between the type of store which a person perceives and the dimensions of perceived risk.

Sampling Procedure

A sample of forty-four housewives and wives working outside the home was taken on a voluntary basis through personal contact and door-to-door solicitation. Twenty-two of these subjects were residents of Midwest City, Oklahoma, a suburb of Oklahoma City, and twenty-two of these subjects were residents of Stillwater, Oklahoma. T-tests indicated no significant differences in the overall manner in which the subjects of these two communities responded to the experimental measurement device.

Data Collection

In order to generate experimental data, each of the forty-four respondents received a self-administered questionnaire which contained one situation with both store types. The assignment of situational treatments, as well as the order of presentation of store type, were randomized so that as many respondents as possible received different orderings and situational treatment pairs.

Each page of the questionnaire contained a brief statement of the buying situation including a specific brand or store. Following each buying situation were eight questions concerning four different risk producing dimensions. Based on the conceptual work of the Harvard group (Cox, 1967), each type of risk (social, etc.) was assumed to have an uncertainty component and a consequence component. Thus, two types of risk measures were required for each type of risk. The respondents were asked to indicate on a seven point scale their impressions of the likelihood of occurrence and how important it would be to avoid the consequences associated with each of the four risk producing dimensions, within the context of the given situation. Following is an example of the type of questions and scales used.

SITUATION: You are considering buying a new brand of electric knife at Gibson's Discount Store to give as a wedding present to a close friend.

How likely is it that others would think less of you if you were to give this product?

1-Extremely Unlikely    2-Quite Unlikely     3-Somewhat Unlikely    4-Can't Say     5-Somewhat Likely    6-Quite Likely     7-Extremely Likely

How Important is it for you to avoid causing others to think less of you?

1-Extremely Unimportant    2-Quite Unimportant     3-Somewhat Unimportant    4-Can't Say     5-Somewhat Important    6-Quite Important     7-Extremely Important

The total response to any given risk producing dimension is considered to be the product of the numerical representations of the two questions. These measures of dimensional perceived risk will be considered as dependent variables for purposes of statistical analysis.

The four risk producing dimensions which were chosen for study were: social risk, financial risk, performance risk, and physical risk. Table 1 contains the operational definitions for each of these dimensions.

TABLE 1

OPERATIONAL DEFINITIONS OF RISK DIMENSIONS

Statistical Procedure

The statistical analysis for the experiment was an analysis of variance with repeated measures. This type of analysis separates the respondents into groups according to the number of situations in the experiment. Since each respondent was presented only one situation and each received individual identification numbers, it was possible to separate the residual term into two distinct parts, between variations and within variations.

The structuring of F-statistics for an evaluation of brand or store effects was accomplished through division of the mean square due to brand or store effects by the mean square due to within variations. In a similar manner, the F-statistics for the evaluation of brand situational interaction or store situation interaction was accomplished through division of their respective mean squares by the mean square due to within variations. The F-statistics for the evaluation of situational effects was accomplished through division of the mean squares due to situational effects by the mean square due to between variations. A further discussion on repeated measures may be found in statistical principles in experimental design by B. J. Winer (1962).

RESULTS

The experiment concerns the purchase of an electric knife either for one's home use or as a wedding gift to a close friend. Each of forty-four individuals was given one of these two situations and responded to measures of performance risk, social risk, physical risk, and financial risk for two stores.

Table II contains the mean scores obtained for each dependent variable of the experiment and should be used in conjunction with the analysis of variance tables to indicate where differences exist and the direction of those differences. The higher the score, the greater is the level of perceived risk. Tables III through VI are the analysis of variance tables for each dependent variable or risk producing dimension. The percent contribution column was obtained utilizing an Omega Square statistic described by Hays (1964) and is intended to reflect the percentage of the total variation from the dependent variable which can be explained by each independent variable. This statistic has been calculated for only those effects which entered the analysis at a .05 or higher level of significance.

TABLE 2

TABLE OF MEANS PERCEIVED RISK IN ELECTRIC KNIFE PURCHASE

TABLE 3

ANALYSIS OF VARIANCE FOR SOCIAL RISK

TABLE 4

ANALYSIS Of VARIANCE, FOR FINANCIAL RISK:

TABLE 5

ANALYSIS OF VARIANCE FOR PERFORMANCE RISK

TABLE 6

ANALYSIS OF VARIANCE FOR PHYSICAL RISK

Social Risk

It is apparent that situational factors have entered at a highly significant level (P <.01) as a means of explaining variations in responses to social risk. In this case, 13.05 of the total variation was explained by situational factors. The social risk involved in considering the purchase of an electric knife for home usage is significantly less than considering the same product as a wedding gift, It appears that the social risk involved in buying a product for oneself would be less than buying the same product as a gift for someone else. With respect to levels of social risk, there appeared to be no statistically significant store effect.

The interaction effects due to store-situation combinations significantly (P <.02) explained 1.6% of the variation in social risk. This result would indicate that the experimental results attributed to the situational factor are slightly influenced by the store factor. In this specific case, it appears that the social risk involved in buying an electric knife for home usage is lower for Gibson's Discount store than for Montgomery Ward. However, the social risk associated with buying an electric knife from Gibson's Discount store as a wedding present is significantly greater than from Montgomery Ward.

Financial Risk

Situational factors significantly (P < .01) explained 13.92 of the variation in financial risk with financial risk being greater for the home usage situation than for the wedding gift situation. Here, the direction of mean differences is reversed from that obtained for social risk. This would imply that the respondents were more concerned about the price of a product which they intended to use themselves than when giving it as a wedding present.

In this particular case, by buying a more expensive model for a wedding present than they would have considered for themselves, could possibly be interpreted as a risk reduction method for handling perceived social risk,

There were no statistically significant store or interaction effects upon financial risk.

Performance and Physical Risk

With performance and physical risk, the only significant effect was related to store differences.

A significant (P <.02) store effect on performance risk was indicated and explained 3.03% of the total variation for that dependent variable. Since the performance risk for an electric knife was less for Montgomery Ward than for Gibson's Discount, it would appear that these respondents felt that a better quality of electric knife could be found at a Montgomery Ward store than at a Gibson's Discount store.

No other main effects or interaction effects were found which significantly influenced the levels of perceived risk for either of these risk dimensions.

DISCUSSION OF RESULTS

Four hypotheses concerning the independent effect of buying situations upon four dimensions of perceived risk were tested; two were statistically significant. When significant results were found, the amount of total variation explained by situation ranged from 13.05% to 13.92%. Store factors had a minor influence on performance risk. The magnitude of the variance explained supports Lutz and Kakkar's (1975) contention that one should not expect too much for situation in isolation. The results do, however, suggest that buying situations are related to "selected" intervening variables which may influence brand performance or purchase.

Theoretically, it appears that buying situations, along with individual differences and product characteristics, may play a part in determining perceived risk. This would imply that it would he appropriate to identify specific buying situations when experimental efforts are being made to measure perceived risk. This may eliminate a significant portion of unexplainable variations which has previously existed between individual respondents.

In a practical sense, it may be useful for advertisers to identify buying situations which tend to elicit high levels of perceived risk and either attempt to lessen these effects by illustrating the appropriateness of the product in the high risk situation, or simply choose to avoid advertisement formats which depict such high risk situations.

REFERENCES

J. Arndt, "Word-of-Mouth Advertising and Informal Communication," In Donald F. Cox (ed.) Risk Taking and Information Handling in Consumer Behavior (Boston: Division of Research, Harvard Business School, 1967).

Raymond A. Bauer, "Consumer Behavior as Risk Taking," Dynamic Marketing for a Changing World, R. S. Hancock (ed.) (Chicago: American Marketing Association, 1960), 389-398.

Russell W. Belk, "An Exploratory Assessment of Situational Effects in Buyer Behavior," Journal of Marketing Research, XI(May, 1974), 159-163.

Russell W. Belk, "The Objective Situation as a Determinant of Consumer Behavior," Advances in Consumer Research, Vol. II, M. J. Schlinger (ed) (Chicago: 1975).

Donald F. Cox, "Perceived Risk and Consumer Decision Making," Journal of Marketing Research, I(November, 1964), 32-39.

Donald F. Cox, (ed.). Risk Taking and Information Handling in Consumer Behavior (Boston: Harvard University Press, 1967).

S. M. Cunningham, "Perceived Risk as a Factor in Product-Oriented Word-of-Mouth Behavior: A First Step," In Peter Bennett (ed.) Proceedings American Marketing Association, 1965, 229-238.

W. L. Hays, Statistics for Psychologists (Chicago: Holt, Rinehart, and Winston, 1964).

Robert D. Hisrich, Ronald J. Dornoff, and Jerome B. Kernan, "Perceived Risk in Store Selection," Journal of Marketing Research, IX (November, 1972), 435-439.

Richard J. Lutz and Pradeep Kakkar, "The Psychological Situation as a Determinant of Consumer Behavior," Advances in Consumer Research, Vol. II, M. Schlinger (ed.) (Chicago: 1975).

Kenneth E. Miller, "A Situational Multi-Attribute Model," Advances--Consumer Research, Vol. II, M. Schlinger (ed.)(Chicago: 1975).

Michael Perry and B. Curtis Hamm. "Canonical Analysis of Relations Between Socioeconomic Risk and Personal Influence in Purchase Decisions," Journal of Marketing Research, Vol. VI, (August, 1969), 351-354.

Ted Roselius, "Consumer Rankings of Risk Reduction Methods," Journal of Marketing, XXXV (January, 1971), 56-61.

Ivan Ross, Perceived Risk and Consumer Behavior: A Critical Review, Working Paper Number 21. (Minneapolis: University of Minnesota, 1974).

R. G. Sandell, "Effect of Attitudinal and Situational Factors on Reported Choice Behavior," Journal of Marketing Research, V, 1968, 405-408.

H. E. Spence, J. F. Engel, and R. D. Blackwell, "Perceived Risk in Mail-Order and Retail Store Buying," Journal of Marketing Research, VII (August, 1970), 364-369.

Bernard J. Winer, Statistical Principles in Experimental Design. (New York: McGraw-Hill, 1962), 298-318.

W. G. Zikmund and J. E. Scott, "A Multivariate Analysis of Perceived Risk, Self-Confidence and Information Sources," Advances in Consumer Research, Volume I. Proceedings of the Association for Consumer Research. S. Ward and P. Wright (ed.) (Chicago: 1973), 406-416.

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