Discussion Paper No.

0205

August 2002

PLANNED VERSUS IMPULSE BUYING: IMPLICATIONS TO RETAIL SEARCH STRATEGIES

by

BEN PAUL B. GUTIERREZ

Note: UPCBA discussion papers are preliminary versions circulated privately for critical comments and are not for quotation or reprinting without prior approval. They are protected by Copyright Law (P.D. No. 49)

Planned Versus Impulse Buying: Implications to Retail Search Strategies Ben Paul B. Gutierrez*
Factors influencing planned and impulse purchases are investigated using six personal care product categories. Data analyzed consisted of 982 product purchases of 502 consumers. Results of the study show that product category, intended user, purchase frequency, brand comparisons, assistance of beauty consultants, age and income appear to be dominant factors influencing planned and impulse purchases and the choice of retail search strategies. Some implications for management were identified including the need for marketers and retailers to design shopping environments that reduce the shopping difficulties of their target consumers. Key Words: Planned/impulse buying, epistemic/hedonic search strategies, retailing

1. Introduction Consumer retail search behavior is an interesting area that has challenged marketing researchers and practitioners. Unfortunately, there are very few published studies in the Philippines in this area. Even fewer Philippine retailers conduct formal marketing research studies. The dwindling disposable income of Filipinos and the entry of foreign retailers because of globalization are expected to intensify competition among retailers. Recently, the Philippine government has shown its commitment to globalization by allowing foreign retailers to operate subject to a minimum capital investment. In a highly competitive environment, only those retailers who exceed the expectations of their customers in terms of providing enjoyable shopping experience can survive and become successful. Consumers frequently encounter shopping difficulties in retail environments. Yet Titus and Everett (1995) argued that very little research has addressed questions about how consumers search through retail settings. This research will contribute to the understanding of retail search and choice behaviors as applied to personal care products. This present study utilizes the navigational search strategy selection subset of the consumers retail search process (CRSP) framework of Titus and Everett (1995). The retail search process may either employ epistemic or hedonic search strategies. Epistemic search strategies are designed for locating the desired products inside retail environments. These include strategies that rely on consumers spatial knowledge (or cognitive map) of the shopping environment (Passini 1977), or strategies that require the assistance of others (e.g. employees, other patrons) when consumers get lost (Weisman 1983). On the other hand, hedonic search strategies comprise the experiential aspect of the retail search activity (Holbrook and Hirschman 1982). Retailers attempt to enhance the overall quality of the multisensory shopping experience to satisfy the shoppers hedonic pursuit of pleasure. Bloch et al. (1989) claimed that consumers often resort to browsing activity to obtain recreational benefits. ________
*Associate Professor of Marketing, College of Business Administration, University of the Philippines Diliman, Quezon City. (Email:bpbg@up.edu.ph).

Consumers using epistemic search strategies are problem-solvers and are susceptible to utilitarian shopping trips (Babin et al. 1994). These consumers are concerned with efficiency and would want to complete their shopping trips at the soonest possible time. They would value environmental legibility more than environmental stimulation. Thus, they more likely have planned purchases and move through the store as quickly as they can. On the other hand, consumers using hedonic search strategies are likely unfamiliar with their shopping environment thus they would be proceeding more slowly and valuing environmental sensory stimulation. They would have the tendency to make impulsive purchases. The objective of the study is to determine the factors that may contribute to the choice of a retail search strategy among urban Filipino consumers, an important consideration for marketers and retailers. Another research objective is to understand shopping habits of consumers in terms of level of impulse and planned purchases. The paper is organized as follows. Section 2 reviews the literature on factors affecting the choice of search strategies and based on this review, research hypotheses are developed. Section 3 describes the survey methodology and the measures of the variables. To test the research hypotheses, collected data are analyzed using logistic regression. The results are described in Section 4. Section 5 summarizes the major findings and proposes suggestions for future research and implications for management. 2. Theoretical Foundation and Research Hypotheses Avery (1996) applied Stiglers (1961) Economics of Information Theory in a study of consumer search for frequently purchased nondurable grocery items. This model assumes that consumer search activity is influenced by cost and benefit factors. The amount of search undertaken by the customer is positively related to the purchased value of the good and the degree of price dispersion in the market. On the other hand, the amount of search is negatively related to cost of the search in terms of consumers opportunity cost of time, the cost of processing new market information and other direct costs related to search such as transportation. Better search strategies provide benefits in terms of lower prices or higher product quality. Yet, empirical evidence show that most customers do not easily recall purchase prices (Dickson and Sawyer 1990; McGoldrick and Marks 1986) and only a small percentage shop at different retail outlets to obtain the best deals (Urbany 1986; Urbany et al. 1991). One explanation of the low level of in-store price awareness is the time involved in grocery shopping. Ackerman (1989) and Thomas and Garland (1993) found that grocery shopping is a chore done on a regular basis which makes consumers resort to various shortcuts to achieve shopping efficiency. Hallsworth (1991) also suggested that convenience and access to transportation constrain search behavior. However, Stiglers theory is criticized because many noneconomic factors that affect search behavior are omitted in the model. Researchers (Sproles and Kendall 1986; Smith 1989; Kolodinsky 1990; Hafstrom et al. 1992) claimed that consumers enjoy shopping and they found that the time spent in searching is a significant contribution to this enjoyment, a support for the use of hedonic search strategies. Prus (1991) found an important type of browsing, which is mainly casual or recreational, where the consumer has no intention of making a purchase.

Slama and Tashchian (1985) reported that purchase involvement and search behavior are linked to demographic characteristics of the consumer and their household. They asserted that females are more highly involved and there is greater involvement when there are children at home. Furthermore, they suggest other factors such as marital status, age, education and income of the consumer are closely related to purchase involvement and search. High involvement and the need to plan the shopping activity would support epistemic search strategies. Babin et al. (1994) proposed a scale that identifies shoppers in terms of hedonic and utilitarian (epistemic) shopping values. They defined utilitarian shopping value based on the premise that shopping is task-related and rational and is therefore viewed more as an errand or work. Consumers high on the hedonic shopping value scale find shopping as a pleasurable experience. Here, the shopping experience is enjoyable and is viewed as an escape or adventure. This present study investigates the following hypotheses. The first three hypotheses concern the product category, end user and product purchase frequency. These are normally determined before visiting the retail outlet. The degree of involvement of the consumer and the choice of a retail search strategy depends on the importance of a product category. Six personal care categories were studied. Secondly, the intended user of the personal care product to be purchased may influence the choice of retail search strategy. If the buyer would be using the product, the search activity may be more likely involved and planned. The brand may have been decided on before visiting the retail outlet. Finally, product purchase frequency could indicate the type of retail search strategy employed. Shopping for more frequently purchased products are more likely to be planned as these purchases approach habitual purchase behavior. Product category and the brand are more likely decided on before going to the retail outlet. H1: Product category to be purchased is related to the choice of retail search strategies. H2: Shoppers buying for themselves and members of the household are more likely to employ epistemic shopping strategies. H3: Consumers buying more frequently purchased products more likely to plan their shopping and engage in epistemic search strategies. The next set of hypotheses describes what happens inside a store. This group pertains to brand comparison, store familiarity and beauty consultants. Brand comparison involves information processing and is more likely a feature of epistemic search strategies. Although comparing brands is time consuming it may be seen to be an enjoyable activity by other consumers especially when the brands are actually tried on, say for example in cosmetics. Shoppers who are regular patrons of a retail outlet are familiar with its store layout and locations of product categories. They have easier time navigating the store and tend to find the desired products more efficiently. Beauty consultants provide assistance to the shoppers of personal care products. How helpful or how polite and welcoming they are influences the choice of retail search strategy. The information these beauty consultants provide reduces time spent on information search and is relevant to epistemic retail search strategies.

H4: Shoppers who compare brands are more likely to follow epistemic retail search strategies. H5: When shopper familiarity is high epistemic search strategies are more likely to be used. H6: Assistance of friendly and helpful beauty consultants is positively related to epistemic search strategies. The third group of hypotheses concerns the demographic variables. Davies and Bell (1991) reported that an increasing number and proportion of males are doing grocery shopping. They found that males tend to spend less per shopping trip and spend less time in the store. Family income is another constraint to search behavior. Households with less income are more likely to spend less time on shopping. The effects of the demographic variables on retail search strategies are considered in the following hypotheses. H7: Younger shoppers appear to have more time to engage in the shopping activity and are likely to employ hedonic shopping strategies. H8: Lower income households tend to engage more in epistemic retail strategies, rather than hedonic retail search strategies. H9: Males are more likely to employ epistemic search strategies. H10: More educated people are predicted to use epistemic search strategies. 3. Method Store exit interviews were employed to collect data on six toiletry (used interchangeably with personal care) product categories. These categories are: 1. 2. 3. 4. Hair styling: gels, mousse, sprays Hair care: shampoo, conditioner, treatments and coloring Facial skin care: cleansers, toners, moisturizers and treatments Hand and body care: bar and liquid cleansers, moisturizers, lotion, deodorant, nail polish, body powder and feminine wash 5. Perfumes: alcohol-based fragrances and concentrates 6. Make-up: lipstick, foundation, blush-on, eye makeup, eye liner, mascara Non-probability sampling was utilized with quotas for each product category. While the sampling method limits the generalizability of the findings, it ensured that consumers actually purchased the product categories as required by a major cosmetics company that sponsored the fieldwork and provided the incentives for the respondents. It is also less costly than visiting the consumers at homes more associated with probability sampling methods. Shoppers exiting the store were intercepted, pre-qualified, and requested to participate in the

survey by five teams of marketing research students enrolled in the MBA Program of a major Philippine university during July to August 2000. There were 502 shoppers interviewed in several store locations, mostly from four major shopping malls of Rustans, Shoemart, Robinsons and Gotesco in Metropolitan Manila. Each shopper was then interviewed about two toiletry products just purchased. There were 982 toiletry products, of which 299 product categories and brands, and 181 brand names were represented. The units of analyses for the study are the 982 product purchases of the 502 shoppers. The dependent variable is operationalized by asking two simple questions answerable by yes or no: whether the product category and the brand were decided on prior to purchase of each item. While other measures such as the Barbin et al. (1994) are available, the two questions relate to an actual previous purchase in this present study. The behavioral measure rather than just the intention-to-behave (an attitudinal measure) was favored. For purchases using epistemic search strategies, consumers have already decided on both the product category and the brand before going to a retail outletcompletely planned purchase. The ability to plan their shopping shows that these consumers are able to resist impulses to buy and exercise a higher degree of self-control (Baumeister 2002). An impulsive purchase occurs when both the product category and the brand may not have been planned or when either the product category or the brand was not planned. Impulse buyers are believed to pursue hedonic retail search strategies as they tend to find browsing an enjoyable activity. Giving in to the impulse to buy is part of rewarding oneself and this act contributes to the pleasurable experience of shopping. A logistic regression model was fitted with the retail search strategy (epistemic or hedonic) as dependent variables while the independent variables are previously discussed with the hypotheses. As the response category with the higher number of occurrence, epistemic search strategy was designated as the baseline category. Thus, the fitted logit model is for Phedonic/Pepistemic. 4. Results The sample is 85 percent female, aged 26-40 years belonging to a household with four members, one of which is less than 18 years old. At least 73 percent are college graduates and at least 55 percent of the households they belong to earn, on average, at least P20,001 monthly. The sample profile may not be representative of the population, given the convenience sampling, but it approximates the urban shopping population. Most respondents were very deliberate in their choice of personal care products. They have already decided on what product category and what brand to buy before going to their shopping trips. Based on self-reports level of planned purchase is high at 82 percent and most of the shoppers have decided on a brand to buy beforehand (82.8%). These two variables are correlated (Pearson r = 0.605, p < 0.001, 2-tailed). The degree of planned versus impulse buying was further analyzed using the Antonides and van Raaij (1998) framework. Responses for product category and the brand purchased were simply crosstabulated. Table 1 shows that there are only 76.6 percent of the shoppers who have completely planned purchase, buying the intended product category and the exact brand where the epistemic retail search strategies were assumed to be employed. The remaining 23.4 percent are impulsive purchases. The proportion of completely impulse

buying (no intention of buying the category and the brand) is 11.9 percent. Occurrence of impulsive brand choice, where product category was planned but the brand was unplanned, is 5.3 percent. Impulsive product choice (18.1%) encompasses unplanned purchase in a product category regardless of the selected brand. Incidence of impulsive product choice, when the brand chosen was the usual brand, is 6.2 percent. The degree of impulsive buying could influence the navigational search strategies of consumers. Hedonic search strategy may be favored for impulsive choices.

Table 1 Degree of Impulse Shopping in Toiletries

N= 982 product purchases Yes Did you decide to buy this product before coming here? Total Count % of Total Count % of Total Count % of Total Looking for this exact brand? Yes No 752 76.6% 61 6.2% 813 82.8% 52 5.3% 117 11.9% 169 17.2% Total

804 81.9% 178 18.1% 982 100.0%

No

It would also be interesting to know the distribution of impulsive shoppers among the 502 respondents with respect to their two purchases. Out of 502 respondents, only 51 (10.2%) made completely impulsive purchases where both of their purchases were unplanned. There were 127 respondents (25.3%) where the combination planned/unplanned purchase occurred. During the logistic regression analysis there were only 851 purchases included because of missing data. The fitted regression model is found on Table 2 while the model diagnostics and classification table follow on Tables 3 and 4 respectively. Purchases of hair care and facial care product categories are planned and would more likely utilize epistemic retail search strategies. Respondents may have considered the head as the most conspicuous part of the body and the purchases to be used on the head are more likely to be involved than the other toiletry categories. This supports hypothesis 1. Purchased products intended for the buyer and his or her household appear to have followed hedonic retail search strategies the opposite of what was predicted in hypothesis 2. However, the low significant level (p<0.10) only provides weak support for hypothesis 2. About 58.6 percent of the respondents bought products for personal use while another 24.7 percent shared the product with others in the household. Still another 15.1 percent bought for other members in the household but not for themselves. Only a small percentage (1.6%) bought products for people outside of their households. While the coefficient of more frequently purchased products (purchased more than once a month) was not significant, hypothesis 3 is still supported because for purchases that have lag time of more than one month (last purchase was more than one month ago) the coefficient was positive and significant (p<0.01). Consumers buying less frequently

purchased products are more likely to engage in impulse buying and use hedonic search strategies. This result does not support the role of the degree of involvement of the buyer in a particular purchase but this finding may be category-specific. The personal care product

Table 2 Logistic Regression Model

Variable Q5ProductCategory Q5(1)HairStyling Q5(2)HairCare Q5(3)FacialCare Q5(4)SkinCare Q5(5)Perfumes Q6 UserType Q6(1)Self Q6(2)OtherHH Q6(3)Self&HH Q7 PurchaseFreq Q7(1)Morethan1monthago Q7(2)Once a month Q13(1)BrandComparison Q24 BeautyConsultant Q29 Age Q29(1)<than18years Q29(2)18-25years Q29(3)26-40years Q35 HHIncome Q35(1)<P10,000 Q35(2)P10,001-20,000 Q35(3)P20,001-35,000 Q35(4)P35,001-50,000 Constant B S.E. Wald 12.3705 .0132 8.4095 4.4098 .7210 1.4719 6.1403 3.9325 1.2523 3.3063 11.7537 6.4879 .2162 23.9821 6.7192 12.5093 5.5530 .5804 2.6975 6.1674 .1208 .0004 3.0112 2.5909 5.4263 df 5 1 1 1 1 1 3 1 1 1 2 1 1 1 1 3 1 1 1 4 1 1 1 1 1 Sig .0300 .9084 .0037 .0357 .3958 .2250 .1050 .0474 .2631 .0690 .0028 .0109 .6419 .0000 .0095 .0058 .0184 .4462 .1005 .1870 .7282 .9848 .0827 .1075 .0198 R .0496 .0000 -.0815 -.0500 .0000 .0000 .0121 .0448 .0000 .0368 .0897 .0682 .0000 .1510 -.0700 .0822 .0607 .0000 -.0269 .0000 .0000 .0000 .0324 .0248

-.0323 -.8350 -.6259 -.2266 -.3505 1.4321 .8799 1.3462 .5405 -.1062 .8297 -.0905 1.2307 -.1974 -.4305 .1000 .0047 .4815 .4563 -1.9645

.2808 .2880 .2981 .2668 .2889 .7222 .7863 .7404 .2122 .2284 .1694 .0349 .5222 .2591 .2621 .2876 .2477 .2775 .2835 .8433

Table 3 Model Diagnostics

-2 Log Likelihood Goodness of Fit Cox & Snell - R^2 Nagelkerke - R^2 870.387 838.039 .104 .154 df 8 Significance .7375

Hosmer and Lemeshow Goodness-of-fit test

Chi-Square 5.1865

Table 4 Classification Table for Type of Retail Search strategy

The Cut Value is .50 Predicted Planned Impulsive choice P I %Correct

Observed

8 Epistemic(Planned) E Hedonic(Impulsive) H 609 26 180 36 Overall 95.91% 16.67% 75.79%

category may be prone to impulsive purchases especially for less frequently bought toiletry categories such as perfumes. Consumers who compared brands used hedonic rather than epistemic retail strategies. Hypothesis 4 was supported but the coefficient had the wrong sign. It is speculated that although the brand purchase was planned, comparison may still be resorted to when there is a new product, a new packaging or a new advertisement of a competing brand. This implies that brand switchers are vulnerable to marketing promotions and other point-of-sale activities. It is to be highlighted that brand building and convincing information on the label could be factors during brand comparisons. Respondents reported price (cheapest brand=12%; buy the expensive brand=10%) and a promotion/discount (10%) to be determinant factors during brand comparisons. Surprisingly store familiarity which was supposed to be positively related to epistemic retail search strategy was not significant (hypothesis 5). Most of the shoppers (93%) were previous patrons of the store they visited. Such store familiarity makes it easier to find the toiletries needed and increase shopping efficiency. About 92 percent of the 502 shoppers found it easy to locate the products they were looking for and about 91.6 percent who claimed to easily recognize the difference between product categories. In spite of these, there were only 64 percent who experienced no difficulties. The remaining 36 percent were either lost or in a quandary in making their purchases. Problems with information appear to be most common, accounting for 75 percent of the difficulties. Except for unavailability, high price and undisclosed reasons, all the difficulties are information-related (Table 5).

Table 5 Difficulties encountered when choosing the products (Note: Several responses possible) Frequency n= 502 respondents Not enough information 27 Lack of information on product categories 18 Lack of information on brands 33 Lack of information on promotions 19 Difference between product categories confusing 34 Stockout 8 Cannot find item, changed shelf location 9 High price 7 Too many choices 2 Others, not stated 32 Total 189 Percent 14 10 17 10 18 4 5 4 1 17 100

Hypothesis 6 was supported (p < .05) indicating that the availability of friendly and helpful beauty consultants is positively related to epistemic search strategies. When the respondents were asked about the beauty consultants they have encountered, less than half of the respondents found the beauty consultants helpful (43.4%) or polite and welcoming (48.2%). Retailers can take these findings to maintain trained beauty consultants and provide adequate signs whenever some relocation of products took place. The prevalent practice of hiring temporary salespeople in most department stores is contributing to some of these problems. Among the demographic variables, age and income appeared to be factors influencing the choice of a search strategy. As predicted by hypothesis 7, younger people are more likely to use hedonic retail search strategies (p<0.05) while females with children appear to follow epistemic search strategies (p<0.10). However, the opposite of hypothesis 8 was supported. Consumers with higher income appeared to favor impulsive shopping and hedonic retail search strategies, rather than epistemic search strategies. The effects of gender and education were not significant. Hypotheses 9 and 10 were not supported. 5. Conclusion This study highlights the factors influencing the choice of retail search strategies. Before the visit to a retail outlet, the product category, the intended user and the purchase frequency appear to be significant factors. Epistemic search strategies appear to be favored when buying hair and facial care product categories. Hedonic search strategies may be followed when buying products for the self and members of the household and when buying infrequently-purchased products. Since most respondents bought toiletries for personal use communication campaigns that encourage self-indulgence or pampering self would be effective. Products that enter the daily personal rituals of consumers could become successful in the market. Inside the store as consumers process information during brand comparisons and as they deal with beauty consultants they appear to use epistemic search strategies. This implies that marketers and retailers need to design shopping environments that reduce the expected shopping difficulties of their target consumers. It is important to understand that shoppers have varying types and amounts of product and store knowledge, which may affect their navigational search strategies. The findings reveal that shopping efficiency is paramount. Most purchases (76.6%) were completely planned. More than 90 percent of the consumers easily recognized product categories and found their brands. Store familiarity is also high because 93 percent of the respondents were previous patrons. Shopping difficulties were utilitarian as they were dominated by information-related problems. The difficulties seem to indicate that the communication efforts of local marketers and retailers are ineffective and that the shopping environment may be unfriendly. Having friendly and informative beauty consultants is one of the suggested improvements of the consumers. Other factors may need to be investigated such as time available for shopping and environmental stimuli. The presence of environmental stimulation variables such as scent and sound may moderate the choice of search strategy. The knowledge of product categories associated or purchased together is valuable in store layout designs and shelving strategies

10

Researchers also suggest that consumers engage in multipurpose shopping trips and they shift back and forth between epistemic and hedonic search strategies throughout their shopping experience (Titus and Everett 1995). There is an opportunity for local retailers to improve their performance in terms of merchandise assortment, making shopping more enjoyable, having helpful consultants and providing more information about their products.

References Ackerman, NM (1989). Money resources, time demands, and situational factors as predictors of shopping time. Journal of Consumer Studies and Home Economics, 13. Antonides G and van Raaij WF (1998). Consumer behaviour: A European perspective, Wiley. Chichester, England. Avery RJ (1996). Determinants of search for nondurable goods: An empirical assessment of the economics of information theory. The Journal of Consumer Affairs, 30(2), 390-404. Babin BJ, Darden WR and Griffin M (1994). Work and/or fun: Measuring hedonic and utilitarian shopping value. Journal of Consumer Research, 20 (March), 644-656. Bloch P Ridgway NM and Sherrell DL (1989). Extending the concept of shopping: An investigation of the browsing activity. Journal of the Academy of Marketing Science, 17, 13-21. Baumeister RF (2002). Yielding to temptation: Self-control failure, impulsive purchasing, and consumer behavior. Journal of Consumer Research, 28(4), 670-676. Davies G and Bell J (1991). The Grocery Shopper Is He Different? International Journal of Retail and Distribution Management, 19(1), 25-28. Dickson PR and Sawyer A (1990). The price knowledge and search of supermarket shoppers. Journal of Marketing, 54(July), 42-53. Hafstrong JL Chae JS and Chung YS (1992). Consumer decision-making styles: Comparison between United States and Korean young customers. The Journal of Consumer Affairs, 26(Summer), 146-158. Hallsworth AG (1991). Who shops where? And why? International Journal of Retail and Distribution Management, 19(3), 19-26. Holbrook MB and Hirschman EC (1982). The experiential aspects of consumption: Consumer fantasies, feelings, and fun. Journal of Consumer Research, 9(September), 132-140. Kolodinsky J (1990). Time as a direct source of utility: The case of price information search for groceries. The Journal of Consumer Affairs, 2(1), 89-109.

11

McGoldrick JB and Marks H (1990). How many grocery prices do shoppers really know? Retail and Distribution Management, 14(1), 24-27. Passini R (1977). Wayfinding: A study of spatial problem solving with implications for physical design. Dissertation. Pennsylvania State University. Prus R (1991). Just browsing, thanks: focused and diffused shopping practices in Terry Childers et al. (Eds.), Proceedings of the AMA Winter Educators Conference, American Marketing Association, Chicago, 296-302. Slama ME and Taschian A (1985). Selected socioeconomic and demographic characteristics associated with purchasing involvement. Journal of Marketing 49(1), 72-82. Smith MF (1989). An empirical investigation of changing and sustaining consumer shopping enjoyment. Dissertation. University of Texas, Arlington. Sproles GB and Kendall EL (1986). A methodology for profiling consumers decisionmaking styles. The Journal of Consumer Affairs, 20(Winter), 267-279. Stigler GJ (1961). The economics of information. The Journal of Political Economy, 69, 213225. Titus PA and Everett PB (1995). The consumer retail search process: A conceptual model and research agenda. Journal of the Academy of Marketing Science, 23(2), 106-119. Thomas A and Garland R (1993). Supermarket lists: Their effect on consumer expenditure. International Journal of Retail and Distribution, 21(2), 8-14. Urbany JP (1986). An experimental examination of the economics of information. Journal of Consumer Research, 13, 257-271. Urbany JP Dickson R and Key RJ (1991). Actual and perceived vigilance in the retail grocery market. Marketing Letters, 2(1), 15-25. Weiman G (1983). Improving wayfinding and architectural legibility in housing for the elderly in V. Regnier and J. Pynoos (Eds), Housing the Aged: Design Directives and Policy Consideration, Elsevier, New York, 441-464.