The Effect of Price Ladder Position on Consumer
Purchasing Behaviour
Master Thesis
Chris Dekker
University of Groningen
Faculty of Economics and Business
MSc Marketing Management
c.dekker.2@student.rug.nl
S2973685
First supervisor: Prof. Dr. L. M. Sloot
Second supervisor: Dr. M. Keizer
2
Abstract
This paper aims to research consumer reactions to price changes. More specifically, this paper tries to find the effects of changing position on the price ladder, and researches whether there is flexibility in moving in between competitor prices. Additionally, differences between more hedonic and less hedonic products were tested in this context. A survey was distributed to and completed by 349 Dutch respondents. In this survey, respondents rated the likelihood they would purchase a certain option for every brand in their choice set. This was done for the product groups beer and laundry detergent. Respondents only saw one out of five possible scenarios per product group, and prices differed in every scenario. The data was analysed through the use of a multiple linear regression model. The main finding of this paper is that moving up in price, away from one competitor, negatively impacts
purchase intention, even if other, higher priced competitors are not surpassed in price in the process. This impact becomes lower as the price relative to the average competitor price becomes higher. Furthermore, being the most expensive product further weakens this impact. The paper found no significant differences between more hedonic and less hedonic products in this context. The findings of this paper offer valuable information to managers, as it explains the degree of flexibility managers have when changing product prices relative to competitor product prices.
KEYWORDS: PRICE LADDER – PRICE POSITION – PRICING FLEXIBILITY – HEDONIC
3
Table of Contents
Introduction ... 4
Research questions and management questions ... 6
Literature review ... 7
Price elasticity of demand and sensitivity ... 7
Price positions and price distance... 8
Hedonic and utilitarian products ... 9
Purchase intention ... 9
Conceptual model and hypotheses ... 10
Methodology ... 14 Pre-test ... 14 Experiment ... 15 Sampling ... 18 Variable measurement ... 20 Results ... 24 Initial analyses ... 24 Assumptions ... 25
Interpretation of linear regression and testing of hypotheses ... 27
Conclusion ... 33
Discussion ... 33
Managerial implications ... 34
Limitations and further research ... 35
Acknowledgements ... 37
References ... 38
Appendix 1 – pre-test questions ... 41
Appendix 2 – survey choice sets ... 42
Appendix 3 – survey questions... 46
Appendix 4 – Spearman’s Correlation table ... 48
Appendix 5 – linear relationship ... 49
Appendix 6 – normal distribution of residuals ... 50
Appendix 7 – homoscedasticity... 51
Appendix 8 – moderation effect most expensive ... 52
4
Introduction
Psychological pricing is a very relevant aspect of modern marketing. The rise of omni-channel retailing comes with an increasing price transparency: consumers can quickly compare prices of different retailers by simply searching online (Lazaris, & Vrechopoulos, 2014). Pricing decisions have always been of importance in marketing as part of the four P’s (Monroe, 2003; Borden, 1964) – product, price, place and promotion. This importance may have increased more so in recent decades due to increased price transparency. Consumers have a clear view of all prices in a certain market, so price sensitivity becomes higher (Granada, Gupta, & Kauffman, 2008).
Several pricing decisions have been researched. A few examples of retailers considerations in pricing decisions are psychological price endings (Gendall, 1998; Gendall, Holdershaw, & Garland, 1997), promotional pricing and quality signalling (Quigley, & Notarantonio, 2014). An area that has been researched less extensively is the area of price positions. In this sense, a price position is the position a certain product or store has when compared to other stores or products: the cheapest product would be in the highest position possible, and the most expensive product would be in the lowest position possible. This concept is more specifically discussed in the literature section of this paper.
Consumers nowadays “base their timing and quantity decisions increasingly on price (promotions)” (Bijmolt, van Heerde, & Pieters, 2005; van Heerde, Gupta, & Wittink, 2003). It is often assumed that a change in price leads to a change in demand, and that this relationship is somewhat linear (e.g. in Henderson, 1956): an increase in price of X leads to a decrease in demand of Y, an increase in price of 2X leads to a decrease in demand of 2Y, etc. This might be the case if a unique product is sold by one retailer, without any competition. In a market with more players, however, it is possible that the price position changes the way a price change is received by customers. Furthermore, price distance to the next competitor may also influence the effect of a change in price.
Example
5
Brand Price Annual demand expectations
Six-pack A €6,00 50000
Six-pack B €5,50 55000
Six-pack C €5,40 56000
Six-pack D €5,10 59000
Six-pack E €5,00 60000
Table 1 – Brand overview example
In this simplified scenario, E is ranked as number one in cheapest beer brands, D is ranked as number two, etc. Consequently, Six-pack A is ranked fifth and last, as this brand is the most expensive. If brand associations and quality are not big influencing factors, it can be expected that E is sold most and A is sold least. This can lead to the demands in table 1, which show a linear relationship to the price of every brand.
It is possible that the relationship between price and demand is not linear, but dependent on price position and price distance to competitors. Specifically, it is possible that, in this situation, there is a bonus of being the cheapest and either a punishment or bonus for being the most expensive. An example of this: six-pack A (the most expensive brand) sells 10% less than they would in a linear relationship, because they are considered as “the expensive brand”, and six-pack E (the cheapest brand) sells 10% more than they would in a linear relationship, because they are considered as “the cheapest brand”. In this case, the relationship would appear as in figure 1.
Figure 1 – Annual demand per price level six-packs of beer
6 brand. The price elasticity may be relatively lower in between price positions, whereas elasticity may increase when price change causes a position change. In this example, only a few influencing factors are mentioned. In a realistic setting, more factors would need to be taken into account, like specific product characteristics.
Relevance
The intention of this research is to add to the existing literature on price elasticities and sensitivities by including price position as an important market characteristic. Bijmolt et al. (2005) name the
importance of understanding how price sensitivities vary with differing market characteristics for developing a successful marketing strategy. However, the possible relevance of price positions is not considered in their generalizations. This research tries to fill this gap by analysing different
experimental shopping scenarios that include price positions and purchase intention. The study will be managerially relevant, as a small price change can easily be made, but may have more extreme consequences in some pricing scenarios. Furthermore, the effect of changing a price towards a competitor price without changing actual price position will be researched. Additionally, the difference between more hedonic and less hedonic products will be researched in this situation, in order to test whether conclusions still hold when another type of product is used.
This paper consists of a literature review where hypotheses will be made clear, a description of research methods and data collection, a results and discussion section, and a conclusion. The product groups that are used in the research will typically be sold in grocery stores, as these are accessible by almost anyone participating in the research.
Research questions and management questions
It is of importance to answer the following research and management questions in order to fill the aforementioned gap in literature:
RQ1: How do consumers react to price position changes?
RQ2: How does price distance to the next competitor affect consumer reactions?
RQ3: How do product characteristics affect consumer reactions in different price position scenarios? MQ1: To what extent do retailers have flexibility in making products more expensive, if they do not
overtake competitor product prices?
7
Literature review
In order to fully understand the role of price positions in this paper, it is necessary to evaluate the existing literature on price sensitivity. This may lead to valuable insights on factors that can influence this research. Areas that will be reviewed are (1) generalizations on price elasticity and sensitivity, (2) price positions and price distance, (3) purchase intention and (4) hedonic and utilitarian products. Lastly, a conceptual model with hypotheses will be formed.
Price elasticity of demand and sensitivity
Price elasticity of demand shows how sensitive consumers are to price changes. The formula to obtain this number is as follows:
𝑃𝑟𝑖𝑐𝑒 𝑒𝑙𝑎𝑠𝑡𝑖𝑐𝑖𝑡𝑦 𝑜𝑓 𝑑𝑒𝑚𝑎𝑛𝑑 =%∆ 𝑖𝑛 𝑄𝑑 %∆ 𝑖𝑛 𝑃
In this formula, the top part of the fraction is the percentage change in quantity demanded of a certain product, while the bottom part is the corresponding percentual change in price (Intelligent Economist, 2018). If consumers are rational and reasonably informed, price elasticity is usually negative (Tellis, 1988). The number that results from this calculation shows the price sensitivity of consumers towards certain products: consumers have a higher sensitivity to price changes if the price elasticity is more negative, and a lower sensitivity to price changes if the price elasticity is less negative (Tellis, 1988). In other words, a higher negative price elasticity leads to a bigger decrease in demand if the price is increased.
When reviewing the literature on price elasticity of demand, it stands out that there are many different studies on price elasticity of demand of specific products, like gasoline (Brons, Nijkamp, Pels, & Rietveld, 2008) and electricity (Lijesen, 2007). Because of the nature of this paper, a more general perspective is needed, as the factors taken into account should be applicable to all markets. This perspective can mainly be found in two meta-analyses: one older study by Tellis (1988) and a relatively recent study by Bijmolt et al. (2005).
8 Bijmolt et al. (2005) found that price elasticity is higher at SKU-level than at brand level. This means that consumers switch between SKU’s faster than they switch between brands when price is increased. In situations where consumers have to decide between brands, they would be less sensitive to price changes than in situations where they have to decide between SKU’s.
Price positions and price distance
Interestingly, both meta-analyses by Bijmolt et al. (2005) and Tellis (1988) do not mention anything about price positions specifically, which indicates this view on elasticity of demand is relatively scarcely researched. Therefore, this section briefly evaluates the existing literature on price positions, price distance and bonus effects. Shankar and Bolton (2004) found that competitor price levels are an important determinant for retail pricing. Therefore, in pricing optimization, it is useful to know what the effect is of passing a competitor on the price ladder.
Price position is described by Noone, Canina and Enz (2012) as “relative price— higher or lower than, or on par with, the competition”. Bolton and Shankar (2003) use price position as a dimension in pricing, although this is done at a retailer level, with weighted averages per brand. This paper
researches pricing at a product level, rather than at a retail level. The terminology, however, still is the same.
Noone et al. (2012) performed their aforementioned study on revenue management and price position in the hotel sector. They found that revenue performance is higher for hotels that are more expensive than competitor hotels. Additionally, they found that hotels should try to maintain this price level relative to competitors. The study by Noone et al. (2012) shows that a price position effect exists in the hotel branch. This makes it possible that similar effects can also be found in different markets. On the other hand, a positive effect on revenue does not necessarily mean there is an effect on purchase intention and demand as well.
9
Relative price thinking
Azar (2011) mentions the importance of relative price thinking: consumers making decisions based on relative price differences rather than absolute price differences. If consumers indeed think relatively about price differences, then price changes would make less of a difference for higher prices, as relative differences become lower. A €1 change in price makes less of a difference from €500 (0.2%) than from €100 (1%), just because the relative change is lower. Therefore, it would make sense that consumers are less price sensitive when prices get higher.
Hedonic and utilitarian products
The difference between hedonic and utilitarian products has been an influencing factor on many occasions when looking at purchase intent (e.g. in Kivetz, 2000; Palazon, & Delgado-Ballester, 2013). Hedonic products “relate to the multisensory, fantasy and emotive aspects of product usage
experience” (Hirschman, & Holbrook, 1982), while utilitarian products are the primarily functional products that we need (“utility”). Although categorization is ultimately dependent on context, examples of typically hedonic products are designer clothes and luxury watches, while examples of typically utilitarian products are microwaves and toilet paper (Dhar, & Wertenbroch, 2000).
Consumers make different decisions depending on the hedonic (or utilitarian) nature of the product. Therefore, it is of importance to integrate both more hedonic and less hedonic products into the research, in order to measure whether the hedonic difference has an effect. The expectation is that the purchase of a hedonic product gives the consumer more guilt (Palazon, & Delgado-Ballester, 2013), making them more sensitive to price and price position changes. On the other hand, if a purchase is of hedonic nature, price might be less important because of the obtained pleasure that comes with the purchase. Consumers can consider the higher price as an affordable luxury (Mundel, Huddleston, & Vodermeier, 2017), therefore picking the more expensive option.
Purchase intention
10
Conceptual model and hypotheses
The discussed literature and research questions lead to the following conceptual model and hypotheses.
Figure 2 – Conceptual model
Based on this model, the following hypotheses can be made:
H1: A higher price distance to the average product price leads to a lower purchase intention.
This is the main effect of this paper. As consumers usually have a negative reaction to increases in price, the expectation is that a higher price distance leads to a lower purchase intention. Usually, a negative price elasticity exists (Tellis, 1988). This has been proven by many different researchers in many different areas (e.g. Hanssens, Parsons & Schultz, 2004, p333). There is no reason to believe that this situation will be different in this paper. It is, however, more important how certain aspects actually affect this relationship. These aspects are shown in hypotheses 2 to 4b.
H2: A higher relative price distance to the first lower priced alternative positively moderates the
(negative) effect of price distance to the average product price on purchase intention.
11
Figure 3 – Expected moderation effect relative price distance to first lower priced alternative
H3: If a product is the most expensive product, this negatively moderates the (negative) effect of price
distance to the average product price on purchase intention.
The expectation is that being the most expensive gives a “punishment effect” to the relationship that is described in H1 – consumers are expected to react even more negatively, if a product becomes the most expensive option in the product group. Noone et al. (2012) describe that being the most
expensive can be an advantage in the hotel sector. The situation in the grocery store sector, however, is expected to be different, as the branch has different characteristics that are important. An example of this is that prices do not necessarily signal quality in a grocery store, as low priced product can still be of high quality (Steenkamp & Sloot, 2019). Therefore, being the most expensive is not necessarily an advantage in this sector. Furthermore, as people generally react negatively to price increases (Tellis, 1988), it would make sense that this effect becomes even more negative if a certain product does this more than its competitors. In figure 4, the expected moderation effect is portrayed in a graph.
12 According to literature, there are two possible effects that hedonic level can have on the relationship in H1. This is why the following hypotheses are contradictory hypotheses:
H4a: A higher perceived hedonic level of the product positively moderates the (negative) effect of
price distance to the average product price on purchase intention.
H4b: A higher perceived hedonic level of the product negatively moderates the (negative) effect of
price distance to the average product price on purchase intention.
A common factor in the research on hedonic products is that the hedonic nature of a product has an effect on purchase intent (e.g. in Kivetz, 2000). On the one hand, it is expected that the purchase of a more hedonic product gives the consumer more guilt (Palazon, & Delgado-Ballester, 2013). Therefore, it would make sense if a price increase would have a negative effect on purchase intention, if a product is more hedonic. On the other hand, consumers can see the hedonic product as an affordable luxury, and decide to buy it based on this thought (Mundel, Huddleston, & Vodermeier, 2017). If this is the case, it would make more sense if the hedonic nature of a product weakens the effect that is described in H1. Due to contradictory expectations in existing literature, hypotheses have been created in both a positive and a negative direction. In figure 5 and 6, both expected moderation effects are portrayed in a graph. Figure 5 visualizes hypothesis H4a, while figure 6 visualizes hypothesis H4b.
13
14
Methodology
In this chapter, the experiment and analysation methods will be explained This also includes operationalisation of the variables in the conceptual model, explanations about the pre-test and a description of the sample. Furthermore, a complete overview of the relevant variables will be supplied.
Pre-test
According to Malhotra (2019), a pre-test can help eliminate problems before the actual experiment is launched. It can also help to identify the characteristics of product groups, to decide which product groups are suitable for this experiment. In order to test which product groups can be considered as more hedonic and which products groups can be considered as less hedonic, a pre-test was created and distributed among ten respondents: four of them were men and six were women. The mean age is 29.2, although the median age is 22. This difference can be attributed to convenience sampling (Malhotra, 2019), as parents and students were asked to complete the survey. In this pre-test, people were asked the questions that can be found in appendix 1. They were asked to answer questions about the product groups toothpaste, beer, laundry detergent, chocolate, sun cream and crisps. The products group were chosen because of the nature of the experiment. In order to give consumers enough choice, it is useful to have at least four known brands. One of those brands needs to be relatively popular, as this will be the experimental brand that is important for data gathering. This will be the brand which changes price, while the rest of the prices stay the same.
Scale hedonic values
15
Figure 7 – Hedonic scores per product group
In figure 7, a clear division between the hedonic group and the non-hedonic group is visible. Beer, chocolate and crisps belong to the more hedonic group, while toothpaste, sun cream and laundry detergent belong to the less hedonic group. In order to properly measure difference between more hedonic and less hedonic products groups, the groups beer and laundry detergent have been chosen for the actual experiment. This decision was made, because they are the highest and lowest scoring products on hedonic attitude.
Additionally, there needs to be enough space for products to change price in different scenarios. Therefore, the product groups need to have a notable price range between the lowest and highest product price of the group. This is the case for both beer and laundry detergent: at the moment of writing, normal prices for crates of beer range from around €13 to €17 while normal prices for laundry detergent liquid colour range from around €3 to €8 (Albert Heijn, 2019; Superscanner, 2019).
To get closer to realistic results, the experiments will be done with brands that are known to consumers – this consequently means that quality is known. As mentioned before, this makes consumers more price sensitive (Tellis, 1988). Furthermore, as the choice sets consist of choices of brand, switching because of a change in price happens less frequently (Bijmolt et al., 2005).
Experiment
In order to test the previously mentioned hypotheses, an experiment was created on Qualtrics. A laboratory (online) experiment was chosen because of multiple reasons. Firstly, an online experiment offers a high degree of environmental control – factors can be controlled for, that cannot be controlled for in a field experiment (Malhotra, 2019). Music choice in a store, for example, could have an effect on purchasing behaviour in a field experiment, while this can be filtered out in a laboratory
4,36 4,3 4,18 2,7 2,7 2,42 1 2 3 4 5 6 7
Beer Chocolate Crisps Toothpaste Suncream Laundry detergent H ed o n ic s core (1 -7) Product group
16 experiment. Furthermore, the researcher in this paper had limited time and resources, which makes a laboratory experiment more efficient. Lastly, internal validity is often high in laboratory experiments: they tend to have similar results when repeated with similar subjects (Malhotra, 2019). The decision to create a laboratory experiment also comes with a disadvantage: external validity is likely to be low (Malhotra, 2019), which means that results may be hard to generalize to the Dutch society.
Experiment description
In the experiment, respondents have to decide which product they want to buy out of four options. All options are in the same product group, but with different prices. There are five different situations per product group, and in each situation, the price of product X (in this experiment: Heineken and Ariel) is changed, while the rest of the prices in the product group stay the same. This is an advantage of the earlier mentioned environmental control (Malhotra, 2019). In this way, the hypotheses can be tested, as:
• Absolute price distance to the average competitor product price varies with every price change of X;
• Relative price distance to the first lower priced alternative varies with every price change of X;
• There is a price change scenario where product X is the most expensive option;
• Two product groups will be tested – one that is perceived as hedonic and one that is perceived as not hedonic. The scale by Voss, Spangenberg and Grohmann (2003) is used again to test the hedonic scores of the product groups again. This is done on a scale that ranges from 1 to 7.
17
Figure 8 – Examples of choice sets
Purchase intention
The dependent variable purchase intention will be measured by asking respondents how likely they are to purchase the products, on a 7-point Likert scale that ranges from “definitely not buy it” to
“definitely buy it”. This specific item was taken from a research by Spears and Singh (2004) and has a very high standardized factor loading for the variable purchase intention, while the item has both a very high coefficient alpha and a high composite alpha. Additionally, respondents have to pick one of the four options in both scenarios they get.
Manipulation checks
The experiment has several additional questions to test reliability. In appendix 3, questions are shown which test whether respondents already have a brand preference for beer or laundry detergent. Furthermore, a few manipulation checks are done to test whether respondents have actually seen the aspects that are important – specifically brand and price. This is necessary, as it makes sure that the data that is analysed is actually accurate. Respondents are asked whether they remember which brands were shown and which brands were not. Additionally, they are asked whether they could remember the specific prices of the experimental products.
Final pre-test
The experiment was pre-tested by six respondents (3 men, 3 women) to check whether everything was clear. This is necessary to eliminate potential problems, before the actual experiment is distributed (Malhotra, 2019).The main problem that came forward was that it was hard for respondents to remember specific prices in the manipulation check. Instead, rather than a specific price, respondents were asked whether they could remember the most expensive brands. This made it easier for
18
Sampling
The experiment that was created on Qualtrics was distributed through social media – Whatsapp and Facebook specifically. This method of convenience sampling was most suitable for distribution, because of a limited amount of time and resources (Malhotra, 2019). Data was collected from the 31st
of May until the 4th of June, 2019. This led to 458 responses in total, out of which 349 were fully
completed. All of the data was interpreted and analysed in SPSS. The survey was only distributed in Dutch, which means that all respondents are either Dutch, or Dutch speaking.
Data cleaning
From the sample of 349 respondents, the data was cleaned further. Respondents who could not give the right multiple-choice answers to the question “Which of these brands were not in the assortment and which were?” were filtered out (ca. 45). This was done to make sure that respondents actually considered the assortment, while not just quickly clicking through the survey. Furthermore, some respondents gave some illogical answers, like preferring a brand of beer and later on stating they would never buy this kind of beer. Those answers (ca. 18) were also removed from the sample. Additionally, there were some questions that asked respondents whether they could remember which brand was the most expensive. Even though this type of question was pre-tested and successfully answered by most in the pre-test, the amount of wrong answers to this question was too high in the actual survey. Therefore, this question was not used to further clean the dataset. An aspect to consider is that this might influence the variable that measures which brand is the most expensive. With the help of a box plot, a few outliers were removed (ca. 5). Lastly, respondents who stated they do not drink or buy beer were removed (ca. 10). The group that remained consisted of 271 respondents.
Age
19
Figure 9 – Division of age groups
Education level
In figure 10, it can be observed that most respondents in the sample group have done some form of higher education (MBO, HBO or University), with HBO being the most frequently picked option. All of the three groups are well represented, although the collected data is not very suitable for comparison with CBS data, as CBS uses different education categories. The variable education level will be used as a control variable in the analyses.
20
Gender
Out of everyone in the sample, 226 (83,4%) are female and 42 (15,5%) are male. A minority of three people chose not to specify their gender. The division of gender in figure 11 is one that is not
representative of Dutch society, as men are relatively underrepresented in the sample group. In 2018, the population consisted of 49.6% men (CBS, 2019). This means that the results from the experiment will not be fully representative of Dutch society, as the sample is skewed towards women. In order to control for this, gender will be used as a control variable in the analyses. In this way, the difference can be accounted for.
Figure 11 – Gender
Variable measurement
Before variable creation, the different scenarios and questions per product group were combined into one variable, with a new variable that identified the scenario in which the answer was measured. This eliminated a lot of blank values, while putting all available data under one variable, rather than a variable for each scenario. For instance, after the combination of variables, only the variables Purchase Intention Heineken and Purchase Intention Ariel existed, rather than PI_Heineken_S1,
PI_Heineken_S2, and other variables that belong to a specific scenario.
Hedonic Scores
21
Construct Number of items Cronbach’s Alpha Mean group score
Perceived hedonic level Beer
5 0.870 3.18
Perceived hedonic level Laundry Detergent
5 0.785 2.80
Table 2 – Reliability Analysis and mean values
As a Cronbach’s Alpha of at least 0.6 suffices (Malhotra, 2019), all items can be combined as one score and this can be done for both product groups. The mean scores on perceived hedonic level are different compared to the scores from the pre-test. In the pre-test, laundry detergent had a mean value of 2.42, while beer had a mean value of 4.36. In the actual experiment, the mean scores are much closer together. A possible explanation for this is that respondents may have been affected by the choices they got to see before they answered the questions on hedonic attitude, while they were not affected by this in the pre-test. Furthermore, it may be possible that the amount of respondents in the pre-test was too low to draw conclusions from. Lastly, consumers may have wanted to take less time answering the questions on hedonic attitude, because they were part of a big survey. The values from the actual experiment are the values that need to be worked with. Because of this, it is possible that it is harder to test for the effects of differences in hedonic attitude.
Pricing variables
The exact prices per scenario can be found in appendix 2. Eight variables (four per brand) were used in SPSS, all measured from the prices of Heineken and Ariel:
• Absolute price distance to average competitor price • Relative price distance to average competitor price • Absolute price distance to first lower priced alternative • Relative price distance to first lower priced alternative
22
Data restructuring
Initial analyses seemed to have relatively low predictive power and significance. The variables “relative price distance to average competitor price” and “absolute price distance to first lower priced alternative” were dropped after initial testing. This was done, because there is no real difference in effect between absolute and relative numbers, especially after standardisation. The variables that are in the conceptual model were kept. To create more reliable analyses, the data was restructured to get more cases. As respondents answered questions for two scenarios, there were separate variables for both product groups. The variables were transposed to create a dataset with double the amount of cases. A few variables that were measured individually, rather than with product groups, were chosen as fixed variables: these were stretched in SPSS rather than transposed. Table 3 (on next page) shows a complete overview of both the transposed and the stretched variables. The variables in the table are all final variables in the dataset. In the new dataset, standardised variables were created for all
23
Final Variable Transposed
or stretched?
Original Variables Additional comments
Gender Stretched - As “don’t want to specify” was a possible response, this variable was coded as: 1 = “don’t want to
specify or male” 2 = “female” Age Stretched - Measured in age categories: 16-25, 26-35 up until
66 years and older.
Studied_At_HBO_Uni Stretched - To be able to measure an actual difference, a dichotomous variable was computed and coded as: 0 = “did not study at HBO/University” and 1 = “did
study at HBO/University”
Product_Group ID variable - Variable to identify whether a case belonged to beer or laundry detergent before data was
restructured. Purchase_Intention Transposed Purchase intention Heineken
& Purchase intention Ariel
This variable has a scale from 1 (I would definitely not buy) to 7 (I would definitely buy). Abs_avg_PD Transposed Absolute price distance to
average competitor product price Heineken & Absolute price distance to average competitor product price
Ariel
Absolute price distance is measured in Eurocents.
Rel_Dist_Lower Transposed Relative price distance to first lower priced alternative Heineken & Relative price distance to first lower priced
alternative Ariel
Relative price distance is measured as a percentage, with the lower priced option being 100%. The increase in percentage to the higher priced option is
the relative price distance.
Most_Expensive Transposed Most Expensive Heineken & Most Expensive Ariel
Dichotomous variable, with 0 = “Heineken/Ariel is not the most expensive option” and 1 = “Heineken/Ariel is the most expensive option”. Hedonic_Score Transposed Hedonic Score beer &
Hedonic Score laundry detergent
This variable has a scale from 1 to 7.
Price_Check Transposed Price check Heineken & Price check Ariel
Dummy variable, with 0 = “Respondent could not remember the most expensive option” and 1 = “Respondent could remember the most expensive
option”. Quantity_Preference Transposed Estimated real-life purchase
frequency (0-10) Heineken & Estimated real-life purchase
frequency (0-10) Ariel
Variable that measures how often the respondents would buy Heineken or Ariel if they went on ten shopping trips. This variable has a scale from 0 to
10.
24
Results
In this chapter, the clean dataset will be analysed and the hypotheses will be tested. Multiple linear regression will be used to analyse the variables, although several assumptions need to be met before this can be done. These are a linear relationship, a normal distribution of residuals, no
multicollinearity, and lastly, homoscedasticity. Moderation effects will also be tested, before they are interpreted and tested against the hypotheses.
The ID variable product group was left out of the multiple regression model. This was done because of three reasons:
1. It is not the intention of this paper to test the difference in purchase intention between two different product groups. The researcher is not interested in whether respondents have a higher purchase intention for beer or for laundry detergent.
2. The hedonic score variable already accounts for a difference in product characteristics. 3. Two different product groups were tested in the surveys, because this made sure there would
be a noticeable difference in hedonic score.
Initial analyses
Although the initial plan was to use the average price distance to competitor price (after this, ADCP) as main independent variable, several problems arose while analysing this relationship. Firstly, multicollinearity was a big problem in the initial conceptual model. Almost all variables that had to do something with pricing turned out with a high VIF (above 40). This problem was hard to solve, and it makes sense: if a product is more expensive when compared to competitors, this automatically makes the chance of the product being the most expensive product higher. A Spearman’s correlation was used to analyse this, and a Spearman’s rho of 0.813 (p < 0.001) resulted from this (appendix 4). This high correlation is the likely cause of the high VIF-scores (Malhotra, 2019). Secondly, the model seemed to lack predictive power and significance in general. When analysing moderators, the main independent variable ADCP was often insignificant, while other variables seemed much more important.
25 answered. The research questions ask about price distance and price position changes, and with RDLP as independent variable and ADCP as moderator variable, those questions can still be answered. The hypotheses just need to be altered slightly to make more sense. This is possible, because the literature is linked to the whole concept of pricing theory, and not specifically to the variables ADCP and RDLP. The new conceptual model and hypotheses are as follows (figure 12):
Figure 12 – Revised conceptual model
H1: A higher relative price distance to the first lower priced alternative leads to a lower purchase
intention.
H2: A higher price distance to the average product price positively moderates the (negative) effect of a
higher relative price distance to the first lower priced alternative on purchase intention.
H3: If a product is the most expensive product, this negatively moderates the (negative) effect of a
higher relative price distance to the first lower priced alternative on purchase intention.
H4a: A higher perceived hedonic level of the product positively moderates the (negative) effect of a
higher relative price distance to the first lower priced alternative on purchase intention.
H4b: A higher perceived hedonic level of the product negatively moderates the (negative) effect of a
higher relative price distance to the first lower priced alternative on purchase intention.
Assumptions
Linear Relationship
26 means that a lot of the points on the scatter plot will have significant overlap, making the plot
uninterpretable. This problem was solved by using the jitter function in SPSS. By doing this, the points on the scatter plot are spread out at the places where they would normally overlap. The result of using a scatter plot with the jitter function can be seen in appendix 5. A linear relationship can be interpreted. Purchase intention goes down slightly as RDLP goes up in the scatter plot. This can especially be noticed in the low density of plot points in the top right corner. The amount of points that show both a high purchase intention and a high RDLP are relatively low, while the bottom right corner (high RDLP, low purchase intention) has a higher density. The assumption of a linear relationship is met.
Normal distribution of residuals
In order to test for a normal distribution of residuals, a normal P-P Plot of Regression Standardized Residual was created. The result of this can be seen in appendix 6. Observed cumulative probability and expected cumulative probability should be as close together as possible. Interpretation of the P-P plot shows that this is the case, even though there are a few small deviations. A normal distribution of residuals can be observed. The assumption of a normal distribution of residuals is met.
Homoscedasticity
Homoscedasticity can be tested by plotting a scatter plot with regression standardized residual on the y-axis and regression standardized predicted value on the x-axis. The data (see appendix 7) seems to be clustered together in lines. This makes sense, because there is a limited amount of different values per variable. The distribution of data does look like it is random, with a lot of positive and negative data points on both the x-axis and y-axis. There is no clear pattern: the seven lines that are visible are a direct result of the seven values of purchase intention. The assumption of homoscedasticity is met.
No multicollinearity
Multicollinearity can be detected by checking the VIF-scores that are produced when the linear regression is made. Appendix 9 shows the complete linear regression model with all relevant
27
Interpretation of linear regression and testing of hypotheses
The final multiple regression model (see appendix 9) consists of eight variables:
Variable Description
Independent • Relative price distance to first lower priced alternative (RDLP) Moderator • Interaction: RDLP*Price distance to average competitor price (ADCP)
• Interaction: RDLP*Hedonic score • Interaction: RDLP*Most Expensive
Control • Age
• Gender
• Studied at HBO/University
• Estimated real-life purchase frequency (1-10)
Dependent • Purchase intention
Table 4 – Relevant variables in final regression model
The following analyses have been done to test whether moderation effects exist in this model.
Moderator variable: Price distance to average competitor price
Aiken, West and Reno (2018) propose a graphical way to portray moderation effects. To analyse whether a moderation effect exists between RDLP and ADCP, a new variable was created that split ADCP in three groups:
• Low: 179 cases with the lowest ADCP value • Moderate: 180 cases with the middle ADCP values • High: 179 cases with the highest ADCP value
28
Figure 13 – Scatter plot moderation effect ADCP
The scatterplot shows a difference in the relationship between RDLP and purchase intention, as ADCP takes on a different value. When ADCP takes on a low value, there is a relatively strong negative relationship (y = 3.98-0.91x), and if ADCP takes on a moderate value, there is a mildly negative relationship (y = 3.49-0.04x). On the other hand, if ADCP takes on a high value, there is a relatively strong positive relationship (y = 5.07+2.07x). The fit lines per subgroup follow clear different equation lines. There is a possible moderation effect, and the effect will be more specifically tested in the hypotheses part.
Moderator variable: Hedonic score
The earlier used method that is derived from Aiken, West and Reno (2018) can be used for this variable as well. Another new variable was created for hedonic score, with the same distribution of 179-180-179. The scatter plot that resulted from this can be seen in figure 14. A clear division between low, moderate and high hedonic level can be interpreted again, although the effect seems to be less strong than the effect for ADCP. The three effects per value of hedonic score are:
1. Low hedonic score: negative relationship (y = 3.47-0.27x) 2. Moderate hedonic score: negative relationship (y = 3.75-0.04x) 3. High hedonic score: negative relationship (y = 4.03-0.06x)
29
Figure 14 – Scatter plot moderation effect hedonic score
Dichotomous moderator variable: Most Expensive
As the scatter plot approach is harder to interpret with a dichotomous categorical variable, a multiple regression model with all relevant variables was created in SPSS. A scatter plot approach is harder to interpret, as the change in relationship between independent variable and dependent variable is much more abrupt with only two groups. A scatter plot was still made, just with the support of an additional model. All variables were put in block 1 (for model 1), except for the interaction RDLP*Most
30
Figure 15 – Scatter plot moderation effect “most expensive”
Hypothesis testing
The hypotheses that need to be tested are:
H1: A higher price distance to the first lower priced alternative leads to a lower purchase intention. H2: A higher price distance to the average competitor product positively moderates the (negative) effect of a higher price distance to the first lower priced alternative on purchase intention.
H3: If a product is the most expensive product, this negatively moderates the (negative) effect of a higher price distance to the first lower priced alternative on purchase intention.
H4a: A higher perceived hedonic level of the product positively moderates the (negative) effect of a higher price distance to the first lower priced alternative on purchase intention.
H4b: A higher perceived hedonic level of the product negatively moderates the (negative) effect of a higher price distance to the first lower priced alternative on purchase intention.
31 For all hypotheses that are only tested in one direction, significance can be divided by two. This is the case for hypotheses 1, 2 and 3. Significance cannot be divided by two for hypothesis 4a and 4b. Adjusted significances for the hypothesis variables can be seen in table 5.
Tested variable p / 2
RDLP 0.019
Moderator ADCP 0.016
Moderator most expensive product 0.085
Table 5 – adjusted significances
The control variables age (p= .222), gender (p= .890) and “studied at HBO/University” (p= .888) do not have a statistically significant effect on purchase intention. The control variable estimated real-life purchase frequency does have a significant impact (p < .001) and is the biggest predictor of purchase intention (B = .537). The inclusion of this variable makes the whole model more predictive and statistically significant.
H1: A higher price distance to the first lower priced alternative leads to a lower purchase intention.
RDLP has a significant negative effect on purchase intention (B= -.106, t=-2.083, p= .019). This means that a higher price distance to the first lower priced alternative leads to a lower purchase intention. H1 is accepted.
H2: A higher price distance to the average competitor product positively moderates the (negative) effect of a higher price distance to the first lower priced alternative on purchase intention.
Moderator variable ADCP weakens the negative effect significantly in H1 (B= .101 t= 2.168, p= 0.016). This effect can especially be seen in figure 13. The fit line for points scoring low on ADCP shows a strong negative effect. The line for the moderate category is still negative, but a lot milder. Lastly, the fit line for points scoring high on ADCP even shows a positive effect. This proves that a moderation effect is happening in this relationship. H2 is accepted.
H3: If a product is the most expensive product, this negatively moderates the (negative) effect of a higher price distance to the first lower priced alternative on purchase intention.
The interaction between RDLP and the dichotomous variable “most expensive” significantly weakens the negative effect in H1 (B= .071, t= 1.375, p= 0.085). As can be seen in figure 15, if a product is the most expensive option, purchase intention becomes higher as RPLD goes up. This contradicts
32
H4a: A higher perceived hedonic level of the product positively moderates the (negative) effect of a higher price distance to the first lower priced alternative on purchase intention.
H4b: A higher perceived hedonic level of the product negatively moderates the (negative) effect of a higher price distance to the first lower priced alternative on purchase intention.
33
Conclusion
Discussion
This paper went into detail on price ladder effects for two different product groups. The intention of this research was to add to the existing literature on price elasticities and sensitivities by including price ladder effects as an important market characteristic. This has been done by exploring the existing literature on price elasticities and sensitivity, and by conducting an experiment on this topic with several scenarios per product group. In the experiment, respondents were shown one out of five choice sets per product group at random. The product groups that were used are beer and laundry detergent. Every choice set consisted of four relatively well known brands in the specific category. Out of those four brands, one brand had a different price in every choice set. This resulted in different price positions and price distances for every scenario, and the choices made by respondents led to valuable data.
By analysing the data that resulted from conducting the experiment, this paper tried to find out (1) how consumers react to price position changes, (2) how price distance to the next competitor affects
consumer reactions, and (3) how product characteristics affect consumer reactions in different price position scenarios. Additionally, this paper tries to advice managers on pricing flexibility in specific pricing scenarios. In order to answer these questions, five hypotheses were tested. Out of these hypotheses, two were accepted and three were rejected.
Hypothesis testing
Firstly, relative price distance to the first lower priced alternative (RDLP) has a significantly negative effect on purchase intention. This means that price sensitivity does not change a lot when moving away from the first lower priced alternative: an increase in price has a negative effect regardless of price position change. For retailers, this would mean that there is no space to move price upwards without negative effects, even if they do not become more expensive than another brand while doing so. An explanation for this may be that consumers group products that are around the same price as equally expensive. If product A moves towards product B in price, they may both be considered as expensive.
34 relatively expensive already. This is in line with relative price thinking (Azar, 2011). Figure 13 shows a clear moderation effect, as the relationship between RDLP and purchase intention clearly changes as ADCP takes on different values. For retailers, this means that there is more freedom in pricing
decisions when products are relatively expensive already. In this case, moving up the price ladder has a less significant effect.
Thirdly, as discussed earlier, Noone et al. (2012) found that hotels that were more expensive than others often have higher revenue performance. In the context of this paper, it was expected that being the most expensive would have an opposite effect, and that being the most expensive would make the impact of RDLP on purchase intention even more negative. The analyses in this paper indeed show that being the most expensive indeed moderates the relationship between RDLP and purchase intention, as shown in appendix 9 and figure 13. However, the variable most expensive weakens this effect, rather than strengthening it. A “punishment effect” for being the most expensive was expected, but this hypothesis was rejected. The reason for this may be in line with the moderation effect of ADCP that was discussed in the previous paragraph: consumers may become more indifferent to price changes when it is already relatively expensive, even if a product already is the most expensive option.
Lastly, the hedonic attitude of a consumer towards a product group does not significantly moderate the relationship between RDLP and purchase intention. The scatter plot in figure 14 shows that an
interaction effect may be happening, but the complete multiple regression model shows that the interaction is not in fact statistically significant. Aiken, West and Reno (2018) state that statistical insignificance does not have to mean a moderation effect is not present, In the context of this study, however, all other hypotheses have been tested for statistical significance in a multiple regression model. Statistical significance of hypotheses on hedonic attitude are therefore tested through multiple regression as well. Because of this, hypotheses in both directions are rejected. A possible explanation for the statistical insignificance may be found in a difference in hedonic attitude towards product groups between the pre-test and the actual experiment. In the pre-test, beer had a mean score of 4.36 and laundry detergent had a mean score of 2.42, a difference of 1.94 on a scale of 1 to 7. In the actual experiment, however, mean scores for beer and laundry detergent converged: beer had a mean score of 3.18 and laundry detergent had a mean score of 2.80, a difference of 0.38. The product groups may have been too close together too properly test differences between the product groups.
Managerial implications
35 no competitors are surpassed in price in the process. This means that managers should not think lightly about changing prices at all times, not just when a price change also changes price position. However, there is more flexibility to increase product prices if a product is already relatively more expensive. If a product is the most expensive option, there is even more flexibility to change price. Managers are advised to set a clear pricing strategy for their product. As mentioned before, it is commonly known that a lower price often leads to a higher purchase intention. On the other hand, if a higher price is set, the impact of a price increase on purchase intention becomes lower. Managers should take this into account on both the bottom and the top of the price ladder, and make price changing decisions accordingly.
Limitations and further research
The effects that were found in this paper might not be entirely generalizable to the entire population of the Netherlands. Although the amount of cases was sufficient, a relatively big percentage of the respondents was female. In future research, an experiment that is more representative of Dutch society should be conducted. Additionally, it is interesting to know whether the effects that were found are also applicable to other countries and other product groups.
As discussed in the literature section, omitting quality can cause a positive bias (Tellis, 1988). As quality is a factor when choosing out of four different brands of beer and laundry detergent, consumers are likely to have been more price sensitive in the scenarios than they would have been if quality was not a factor. Furthermore, price elasticity is likely to be higher, as a choice is made out of brands, rather than individual SKU’s (Bijmolt et al., 2005). These effects have not been measured in the model, but might have been a factor as well.
Another limiting factor in this research is the lack of difference in hedonic score between the two product groups in the actual experiment. To properly test the effect of hedonic score, future research should try to find product groups that have more diverged scores.
A disadvantage of the experimental method is that purchasing behaviour may be different in real-life situations. Similar scenarios should be used in real shopping situations, in order to test for real-life behaviour.
36 Lastly, the manipulation check that was supposed to check whether respondents actually processed the given prices was not very successful. Even though the question was pre-tested, the question that asked whether respondents remembered which brand was the most expensive option in their scenario was too hard, and too many respondents did not give the right answers. In future research, an easier question should be used that actually checks whether they consciously looked at the prices.
37
Acknowledgements
38
References
• Aiken, L. S., West, S. G., & Reno, R. R. 2018. Multiple regression: testing and interpreting
interactions. Newbury Park: SAGE Publications.
• Albert Heijn. https://www.ah.nl/producten/bier-sterke-drank-aperitieven/bier/pilsener, May 2019.
• Azar, O. H. (2011). Relative thinking in consumer choice between differentiated goods and services and its implications for business strategy. Judgment and Decision Making, 6(2), 176.
• Bijmolt, T. H., Heerde, H. J. V., & Pieters, R. G. 2005. New Empirical Generalizations on the Determinants of Price Elasticity. Journal of Marketing Research, 42(2): 141–156.
• Bolton, R. N., & Shankar, V. 2003. An empirically derived taxonomy of retailer pricing and promotion strategies. Journal of Retailing, 79(4): 213–224.
• Borden, N. H. 1964. The concept of the marketing mix. Journal of advertising research, 4(2): 2-7.
• Brons, M., Nijkamp, P., Pels, E., & Rietveld, P. 2008. A meta-analysis of the price elasticity of gasoline demand. A SUR approach. Energy Economics, 30(5): 2105–2122.
• Centraal Bureau voor de Statistiek. 2019. Consulted on 04-06-2019.
• Dhar, R., & Wertenbroch, K. 2000. Consumer Choice between Hedonic and Utilitarian Goods. Journal of Marketing Research, 37(1): 60–71.
• Gendall, P. 1998. Estimating the effect of odd pricing. Journal of Product & Brand
Management, 7(5): 421–432.
• Gendall, P., Holdershaw, J., & Garland, R. 1997. The effect of odd pricing on demand. European Journal of Marketing, 31(11/12): 799–813.
• Granados, N., Gupta, A., & Kauffman, R. J. 2008. Designing online selling mechanisms: Transparency levels and prices. Decision Support Systems, 45(4): 729–745.
• Hanssens, D. M., Parsons, L. J., & Schultz, R. L. 2004. Market response models:
econometric and time series analysis. Boston: Kluwer Academic Publishers.
• Henderson, H. 1956. Supply and demand. London: Nisbet & Co.
• Heerde, H. J. V., Gupta, S., & Wittink, D. R. 2003. Is 75% of the Sales Promotion Bump Due to Brand Switching? No, Only 33% Is. Journal of Marketing Research, 40(4): 481–491. • Hirschman, E. C., & Holbrook, M. B. 1982. Hedonic Consumption: Emerging Concepts,
Methods and Propositions. Journal of Marketing, 46(3): 92.
• Kivetz, R. 2000. Hedonic and utilitarian motivations in consumer choice. Advances in
39 • Lazaris, C., Vrechopoulos, A. 2014. From Multichannel to "Omnichannel" Retailing: Review
of the Literature and Calls for Research. 2nd International Conference on Contemporary
Marketing Issues, Athens.
• Lijesen, M. G. 2007. The real-time price elasticity of electricity. Energy Economics, 29(2): 249–258.
• Malhotra, N. K. 2019. Marketing research: an applied orientation. New York, NY: Pearson. • Monroe, K. B. 2003. Pricing: making profitable decisions. Boston, MA: McGraw-Hill/Irwin. • Mundel, J., Huddleston, P., & Vodermeier, M. 2017. An exploratory study of consumers’
perceptions: What are affordable luxuries? Journal of Retailing and Consumer Services, 35: 68–75.
• Noone, B. M., Canina, L., & Enz, C. A. 2012. Strategic Price Positioning for Revenue Management: The Effects of Relative Price Position and Fluctuation on Performance.
Journal of Revenue and Pricing Management, 12(3): 207–220.
• Palazon, M., & Delgado-Ballester, E. 2013. Hedonic or utilitarian premiums: does it matter? European Journal of Marketing, 47(8): 1256–1275.
• Park, J., Lennon, S. J., & Stoel, L. 2005. On-line product presentation: Effects on mood, perceived risk, and purchase intention. Psychology and Marketing, 22(9): 695–719. • Price Elasticity of Demand. 2018, September 26. Intelligent Economist.
https://www.intelligenteconomist.com/price-elasticity-of-demand/. Consulted on 17-03-2019. • Quigley, C. J., & Notarantonio, E. M. 2014. An Exploratory Investigation of Perceptions of
Odd and Even Pricing. Proceedings of the 1992 Academy of Marketing Science (AMS)
Annual Conference Developments in Marketing Science: Proceedings of the Academy of Marketing Science, 306–309.
• Rippé, C. B., Weisfeld-Spolter, S., Yurova, Y., Dubinsky, A. J., & Hale, D. 2017. Under the sway of a mobile device during an in-store shopping experience. Psychology & Marketing, 34(7): 733–752.
• Shankar, V., & Bolton, R. N. 2004. An Empirical Analysis of Determinants of Retailer Pricing Strategy. Marketing Science, 23(1): 28–49.
• Spears, N., & Singh, S. N. 2004. Measuring attitude toward the brand and purchase intentions. Journal of current issues & research in advertising, 26(2), 53-66.
• Steenkamp, J.-B., & Sloot, L. 2019. Retail disruptors: The spectacular rise and impact of
the hard discounters. London: Kogan Page Limited.
• Superscanner. https://superkeuze.nl/c/33445/pilsener, May 2019.
40 • The Multiple Linear Regression Analysis in SPSS. n.d. Statistics Solutions.
https://www.statisticssolutions.com/the-multiple-linear-regression-analysis-in-spss/, June 11, 2019.
41
Appendix 1 – pre-test questions
42
Appendix 2 – survey choice sets
Beer S1
Beer S2
43
Beer S4
Beer S5
44
Laundry detergent S2
45
Laundry detergent S4
46
48
49
Appendix 5 – linear relationship
Without fit line
50
51
52
53
54
Powerpoint slides thesis defence
Dia 1
55 Dia 2 TABLE OF CONTENTS • Short summary • Motivation • Literature review
• Conceptual model and hypotheses
• Research methodology
• Sample descriptives
• Main outcomes
• Results/hypothesis testing
56 Dia 3
SHORT SUMMARY
• Researching price ladder position and consumer reactions
• Experiment was created
• Grocery stores: beer and laundry detergent
• Multiple linear regression model
• Main findings:
• No extra flexibility
57 Dia 4
PERSONAL MOTIVATION
• Initially started researching odd and round pricing
• Convinced to change topic:
• Felt like topic was scarcely researched - potential for new findings.
• New topic is more interesting to me.
• Challenges an existing theory.
• Lots of options for the creation of an experiment.
58 Dia 5
LITERATURE REVIEW
• Price positions not considered in Bijmolt et al. (2005) and Tellis (1988)
• Managerial relevance – consumer reactions
• Azar (2011): consumers sometimes make decisions based on relative difference rather than absolute difference.
• Does this mean consumers are less price sensitive when prices are higher?
• Noone et al. (2012) – price position effects exist in the hotel branch
• Do they exist in different markets?
• In the hotel branch, being in the highest (most expensive) price position has positive effects.
59 Dia 6
LITERATURE REVIEW
• Positive effects only exist when there is a fit between quality and price (Noone et al., 2012).
• Hotel branch: higher price often signals higher quality.
• Supermarket branch: quality and price do not necessarily go hand in hand.
• Hard discounters offer high quality for a relatively low price (Steenkamp & Sloot, 2019).
• Negative effect for being the most expensive is expected – why shop the most expensive options if there are cheaper options of equal quality?
• Hedonic nature
• Palazon & Delgado-Ballester (2013): purchase of a product of hedonic nature gives consumers more guilt.
60 Dia 7
CONCEPTUAL MODEL AND HYPOTHESES
Revised version • H1: A higher relative price distance to the first lower priced alternative leads
61 Dia 8
CONCEPTUAL MODEL AND HYPOTHESES
• H2: A higher price distance to the average product price positively moderates the (negative) effect of
a higher relative price distance to the first lower priced alternative on purchase intention.
• H3: If a product is the most expensive product, this negatively moderates the (negative) effect of a
62 Dia 9
CONCEPTUAL MODEL AND HYPOTHESES
• H4a: A higher perceived hedonic level of the product positively moderates the (negative) effect of a
higher relative price distance to the first lower priced alternative on purchase intention.
• H4b: A higher perceived hedonic level of the product negatively moderates the (negative) effect of a
63 Dia 10
RESEARCH METHODOLOGY
Pre-test Experiment Scales and questions
4,36 4,3 4,18 2,7 2,7 2,42 1 2 3 4 5 6 7 H ed oni c sc or e (1 -7) Product group
Hedonic score per product group
• Hedonic nature • Voss, Spangenberg and
Grohmann (2003) • Notable price range • Brands that are known
• Hedonic nature
• Purchase intention (Spears and Singh, 2004):
• Definitely buy it- Definitely not buy it
• Age/Gender/Education • Estimated real-life purchase
frequency • Manipulation checks
• Brand preference • Did respondents actually
pay attention to brands and prices?
• Qualtrics: online experiment • 2*5 scenarios with different prices
• Respondents only see two • Only Heineken and Ariel change price • Randomized order
64 Dia 11 SAMPLE DESCRIPTIVES Ra w da ta 458 responses were collected through the use of Qualtrics Fin ish ed res pon se s 349 responses were fully completed Data clea nin g 78 responses were removed in the data cleaning process Fin al da ta se t 271 responses remained in the final dataset
Data cleaning process
• Ca. 45 respondents were not able to remember the correct brands in a multiple-choice question. • Answers that do not make sense were removed (ca. 18) –
e.g. stating that they like a brand, but then later on indicating that they would never buy the brand. • 5 outliers were removed.
65 Dia 12
SAMPLE DESCRIPTIVES
66 Dia 13
MAIN OUTCOMES
• Relative price distance to the first lower priced alternative (RDLP) has a significantly negative effect on purchase intention.
• Price distance to the average competitor product price weakens the negative effect of RDLP on purchase intention.
• Being the most expensive product also moderates the relationship between RDLP and purchase intention. The relationship between RDLP and purchase intention is weakened.
67 Dia 14
RESULTS AND HYPOTHESIS TESTING
• Assumptions: Linear relationship, normal distribution of residuals,
homescedasticity, no multicollinearity – all conditions were met
• H1: A higher relative price distance to the first lower priced alternative leads
68 Dia 15
RESULTS AND HYPOTHESIS TESTING
• H2: A higher price distance to the average product price positively moderates
the (negative) effect of a higher relative price distance to the first lower priced alternative on purchase intention. – accepted (B= .101, t = 2.168, p= 0.016)
• H3: If a product is the most expensive product, this negatively moderates the
69 Dia 16
RESULTS AND HYPOTHESIS TESTING
• H4a: A higher perceived hedonic level of the product positively moderates the
(negative) effect of a higher relative price distance to the first lower priced alternative on purchase intention. – rejected (p > 0.100)
• H4b: A higher perceived hedonic level of the product negatively moderates the
70 Dia 17
