Since the impact of product-line advertisement on reference price is central to this research, two experiments were enforced to test the hypothesized relationships.
The first experiment focused the effect of brand advertisement while the purpose of the second experiment is to examine the influence of informational advertisement on reference price toward buyers with different target purchase.
Both of two experiments manipulated the goods that subjects who need to buy and the materials of advertisement, including high-end, low-end and one certain horizontal product, meanwhile, measured the change of reference price. Since the response of reference price toward subjects is primarily based on their perception of product line and advertisement, the establishment of vertical/horizontal product line and the design of advertisement are particularly critical.
Products
Three criteria were used for selecting the product to show in the advertisement.
First, durable is the first category for product selecting in this research. Since Tridib and Insrajit (2005) suggested further study can do more research in durable category to investigate the anchoring effects of a default option, owing to low extent research of reference price in durables in the past. Second, the product had to be the one that most subjects were familiar with and may had purchased or used it, so that they could have enough knowledge to respond to the advertisement. Last, the product must have a set of attributes. On the basis of three criteria, digital camera is selected for experiments.
The design or product line and advertisement
Two pretests were performed ahead of product line design. Usually digital camera is described with a number of features. Considering the intensity of advertisement stimuli, the mainly concerned attributes of digital camera for most consumers were picked in the first pretest. Based on the results of the first pretest, the content of product line was designed. Each of vertical and horizontal product line was composed of two products. Vertical product line contained one high-end product and one low-end product distinguished by their function difference. Moreover, horizontal product line included two products with the same price and part of identical functions but different emphasized feature. The description of product line was listed in Table 1 and Table2.
Table 1
The Design of G series Product Line (Vertical)
Attribute High-end Low-end
(G-X100) (G-A1)
Specification of Price 12,800 5,680
Easy-to-evaluate Attribute Megapixels 12 million 6 million
Optical Zoom 5 X 3 X
Digital Zoom 7 X 4 X
ISO Equivalencies 1600 400 Hard-to-evaluate Attribute Anti-Shake User Interface Emphasis
Table 2
The Design of P-Shot Product Line (Horizontal)
Attribute Product 1 Product 2 (P-Shot I) (P-Shot II)
Specification of Price 9,800 9,800
Easy-to-evaluate Attribute Megapixels 10 million 10 million Optical Zoom 4.6 X 4.6 X
Digital Zoom 6 X 6 X
[Specific Feature] Focal Length ISO Hard-to-evaluate Attribute User Interface Anti-Shake Emphasis
The product-line advertisement was for sure designed according to the characteristic of each product mention antecedently. It was divided into transformational and informational advertisement on the basis of previous advertising research (Puto and Wells 1984). The transformational advertisement in this research was represented by brand advertisement, which intends to evoke identification toward consumers with an amount of description of brand image and aspiration.
Comparatively, the informational advertisement contained rich objective illustration of product specification. Both of brand advertisement and informational advertisement were designed for high-end, low-end and two products toward horizontal product line respectively, and used as the advertisement stimuli in experiments. The magnitude of advertisement stimuli had been tested in the second pretest.
Measures
Both of two formal experiments involved in the measurement of reference price.
While prior literature had recognized the formulation of reference price includes both internal and external conceptualization, various definitions of reference price could be found (Oded 2002). Broadly, the concepts of reference price could be defined with retailer/advertised price, past experience and marginal utility/perception base.
Tow criteria helped us select the measurement of reference price. First, since the change of reference price influenced by advertisement is the most important thing to observe in this research, it was more reasonable to measure the reference price on the basic of marginal utility rather than advertised price or previous experience. Second, the question in experiments is necessary to help subjects reflect their overall change of reference price. Therefore, “Reservation Price”, the maximum that individual would pay, was used to ask the initial reference price and the altered reference price formed after product-line advertisement stimuli toward subjects. The effect of other advertised prices or past experience was controlled or ignored temporarily.
Table 3
Definition of Reference Price
Definition
Retailer/Advertised Price Base
The range of prices normally charged by retailer when the product is
not on special (Emory 1970).
Average price of similar products (Emory 1970; Gabor 1997).
Highest market price (Monoroe 1990); The average retail price (Klein and Oglethorpe 1987); Average market price (Sampson 1964)
Advertised seller’s price; Other store’s price (Biswas and Blair 1991)
Past Experience Base Last price paid (Gabor 1977).
Price most frequently charged (Olander 1970); Price of most frequently purchased brand (Gabor and Granger 1961).
Historical price: the price that consumer usually pay (Klein and Oglethorpe 1987).
Marginal
Utility/Perception Base
Reservation price: the maximum the individual would pay (Thaler 1985).
Fair price: the price reflecting consumers’ beliefs on how much a product should cost (Thaler 1985).
Good buy price: the price consumers would considered as a good buy (Klein and Oglethorpe 1987).
Just price: the price or below it that consumers are prepared to pay (Winer 1988).
Expected future price: The expected price to be charged on the next purchasing occasion (Jacobson and Obermiller 1990).
Subjects
The 574 subjects participating in the experiments were recruited on the internet or from college of business randomly. Subjects were approximately composed of hale male and hale female and roughly range from 16 to 35 years old. There was 94.8 percent of subjects owned experience of reading digital camera ads and 70.4 percent had ever bought at least one digital camera, fitting our request for the structure of subjects. Each volunteered subject was randomly assigned in different experiments and given several specific simulated purchase situations to respond.
3.1 Experiment I (Brand Advertisement)
The purpose of the first experiment was operated to investigate whether a transformational advertisement (brand advertisement) will increase the reference price significantly no matter within the vertical or horizontal product line. Besides the overall impact of brand advertisement on reference price was discussed, this
experiment further compared the impact of brand advertisement with high/low-end product toward low/high-end product buyers respectively.
Procedures
Subjects were divided into four groups and given a stimulated purchase situation first. They were appointed to buy one product within vertical or horizontal product line. While they got their individual task, subjects were requested to read the information of vertical or horizontal product line to know the advertised price and basic specification of the digital cameras. After reading the information of product line, subjects required to write down the maximum price they would like to pay. This was done to ensure their original reference price before advertisement stimulus. For example, a subject appointed to buy a low-end product was asked to read the description of vertical product line first, including the advertised price and functions of high-end and low-end product. The subject’s original reference price was expected to be formed during the process of reading product line description.
The advertisement stimulus was manipulated sequentially. The advertisement was assigned for each subject according to the product that the subject was appointed to buy. The high-end buyer was exposed to the brand advertisement with low-end product, whereas the low-end buyer was exposed to the brand advertisement with high-end product. The same criterion was used in the experiment of horizontal product line.
Table4
The Design of Experiment 1
Vertical Product Line Horizontal Product Line 1-1 1-2 1-3 1-4 Step of Experiment
Appointed High-end Low-end Product 1 Product 2 Purchase G-X100 G-A1 P-Shot I P-Shot II Subject of Brand Low-end High-end Product 2 Product 1 Advertisement Stimuli G-A1 G-X100 P-Shot II P-Shot I
Results
The purpose of experiment 1 focused on the impact of brand advertisement on reference price. To evaluate the change of reference price, a repeated measure of pair/independent t-test was used to analyze. In the analysis, the difference of the mean of original reference price (R1) and altered reference price (R2) is measured by pair t-test directly. However, considering the impact shown on R2 may be offset by
increase and decrease simultaneously due to the inconsistency of subsets’ responds, the dependent t-test may be further used to compare reference price remaining group with reference price increasing or decreasing group to distinguish the effect on reference.
The results of experiment 1 for vertical product line are summarized in Table5.
In the pair T-test, the brand advertisement does not show significant impact on reference (t = -1.001, p=0.317), however, with detailed analysis we find the outcomes of reference price increasing and decreasing may both significant exist (t = 4.211, p=0.000; t=-2.689, p=0.007). Besides, the range of reference price increasing is apparently wider than that of reference price decreasing.
However, the expected result does not appear in horizontal product line section.
The impact of brand advertisement on reference price is significant (t = 5.073, p<0.000), but it results in the decrease of reference price rather than increase, not supporting H1b.
Table 5
Summarized Result of Pair / Independent T-test (vertical product line) Statistic
Mean t-value p-value
Pair ( R1-R2 ) -266.06 -1.001 0.317
R2 ( 1 , 3 ) 5440.14 4.211 0.000***
R2 ( 2 , 3 ) -1078.24 -2.689 0.007**
ABS ( 1 , 2 ) 5416.61 2.402 0.01*
p.s. Choice “1”,”2”,”3” represent reference price increasing, decreasing and remaining.
p.s. ABS means absolute value of (R1-R2)
With hypothesis constructed by prior literature, the quality perception of product was expected to mediate the impact of brand advertisement on reference price. Hence, subjects were divided into two groups to investigate: brand advertisement with low-end product and high-end product. The outcome of experiment represented brand advertisement with low-end product cause the reference price significantly decrease (t
= 4.937, p<0.000) , whereas the brand advertisement with high-end product cause the reference price increase (t = 5.073, p<0.000). Besides, the impact of the advertisement with low-end product is obviously stronger than that with high-end product, explicitly supporting H2.
Table 6
Summarized Result of Pair / Independent T-test (vertical product line) Statistic
p.s. Choice “1”,”2”,”3” represent reference price increasing, decreasing and remaining.
3.2 Experiment 2 (Informational Advertisement)
The second experiment was manipulated to investigate whether informational advertisement of different types of product, high-end, low-end or certain horizontal product, will bring varied impact on buyers with their individual purchase mission. To test the H3, subjects were arranged to have their targeted purchase and hypothesized advertisement. The change of reference price, after a high-end buyer saw the informational advertisement of low-end product, a low-end buyer sees the informational advertisement of high-end product, and a certain product buyer sees the informational advertisement of another product within the same horizontal product line, were expected to be observed. In addition, the product attributes emphasized in the informational advertisement were further divided into easy-to-evaluate and hard-to-evaluate type to understand the mediating effect of concreteness of product attributes.
Procedures
In experiment 2, subjects were divided into eight groups. Similar with experiment 1, subjects of each group were arranged a stimulated situation and appointed purchase. According to the stimulated purchase task, subjects were asked to read the description of vertical or horizontal product line with advertised price and basic specifications. And then, the first question asked them to write down the maximum price they would like to pay as their initial reference price. The informational advertisements in this experiments consisted of a head-line, a (simulated) brand name, a picture of product and a body of text and photo describing the features of product. After exposing to the assigned advertisement, subjects were requested to fill in their second reference price no matter it changed or not. In the end
of experiment, two reference prices for each participant were collected to measure the impact of informational advertisement.
Table7
The Design of Experiment 2
Group Appointed
Purchase
Subject of Informational Advertisement Stimulus 2-1 High-end Low-end with lower megapixels 2-2 High-end Low-end with user interface 2-3 Low-end High-end with higher megapixels Vertical
Product Line
2-4 Low-end High-end with anti-shake function 2-5 Product I Product II with ISO value
2-6 Product I Product II with anti-shake function 2-7 Product II Product I with focal length function Horizontal
Product Line
2-7 Product II Product I with user interface function
Results
Experiment 2 concentrated on the impact of informational advertisement for examinating H3. There are two sections discussed in H3. The first section investigates the impact of informational advertisement on reference price within the vertical product line; meanwhile, examines whether informational advertisements with low-end and high-end product bring different effects on high-end and low-end product buyers.
The results indicated that the informational advertisement with low-end product significantly increase reference price toward the high-end buyers(t= 2.992, p=0.003), supporting H3a. On the other hand, two significant effects were found in the informational advertisement with high-end product. In the beginning pair t-test for H3b, the outcome did not show any evident influence of informational advertisement with high-end product. However, considering the increase and decrease of reference price may offset to each other, some specific analyses were enforced. Consequently, the test results showed that there are two significant impacts of informational advertisement with high-end product (t=6.697, p=0.000; t=-2.845, p=0.005). In other words, the informational advertisement with high-end product may let low-low product buyer increase or decrease his reference prices. These two effects on buyers are considerably equal.
Table 8
Summarized Result of Pair / Independent T-test (vertical product line) Statistic
Mean t-value p-value
with low-end product
Pair ( R1-R2 ) 59.16 0.303 0.762
R2 ( 1 , 3 ) 3265.13 2.992 0.003**
R2 ( 2 , 3 ) -186.46 0.822 0.412
With high-end product
Pair ( R1-R2 ) -4.03 -0.024 0.981
R2 ( 1 , 3 ) 2015.61 6.697 0.000***
R2 ( 2 , 3 ) -816.25 -2.845 0.005**
p.s. Choice “1”,”2”,”3” represent reference price increasing, decreasing and remaining.
The second section of H3 tests the impact of informational advertisement on reference price within the horizontal product line. This experiment indicated the informational advertisement with certain product of horizontal product line may let the buyer, who would like to buy another product within the same product line, decrease his initial reference price (t=-1.951, p=0.05), supporting H3c. But the impact of ads within horizontal product line is much smaller than the impact occurring in vertical product line.
In the end of the experiment, product attributes were separated into two types, easy-to-evaluate and hard-to-evaluate, to test H4. But the results did not suggest there is significant difference existing between informational advertisement with easy-to-evaluate and hard-to-evaluate attributes. H4 did not be supported.