• No results found

And Then There Was Light! And Then There Was Light!

N/A
N/A
Protected

Academic year: 2021

Share "And Then There Was Light! And Then There Was Light!"

Copied!
59
0
0

Bezig met laden.... (Bekijk nu de volledige tekst)

Hele tekst

(1)
(2)

And Then There Was Light!

A study in consumer decision making in residential lighting

University of Groningen Faculty of Economics and Business

Msc Business Administration, Marketing Management August, 2012

(3)

Management Summary

The European Union has recently implemented regulations banning incandescent light bulbs. Little is known about consumer decision making in residential lighting. Using the Theory of Planned Behavior, a framework is developed analyzing variables influencing purchase behavior. Two variables are added to the Theory of Planned Behavior, location and socioeconomic situation. An exploratory research was used to find product attributes as well as indications for behavior. Then, a quantitative research, using an online questionnaire finds some variables influencing decision making. These include locational aspects, residential situation and the following product attributes: ramp-up time, light quantity and atmospheric attributes.

(4)

Index

1. Introduction 6 1.1 Introduction 6 1.2 Problem statement 8 1.3 Structure 8 2. Literature Review 10

2.1 Theory of Planned Behavior 10

2.2 Attitudes Toward behavior 11

2.2.1 Lamp Types and Characteristics 11

2.2.2 Environmental Attitude 13 2.2.3 Innovative Attitude 13 2.3 Subjective Norms 13 2.4 Behavioral Control 14 2.5 Location 14 2.6 Socioeconomic Situation 15 2.7 Conceptual Model 16 3. Research Methodology 17 4. Exploratory Research 18 4.1 Research Methodology 18 4.2 Industry Experts 18 4.3 Focus Group 18 4.4 Survey 20 5. Conclusive Research 21

5.1 Sample and Procedure 21

5.2 Operationalisation of Constructs 21

5.3 Estimation and Assessment 22

5.4 Measurements and Scaling 22

5.5 Results 23 5.6 Sample Descriptives 23 5.6.1 Demographics 23 5.6.2 Product Attributes 24 5.6.3 Subjective Norms 25 5.7 Factor Analysis 26 5.7.1 Interpretation of Factors 28 5.8 Scale Construction 29 5.8.1 Environmental Attitude 29 5.8.2 Innovative Attitude 29 5.9 Research Questions 30

(5)

5.9.3 Socioeconomic Variables and the Theory of Planned Behavior 36

5.9.4 Lamp Location and Purchase Behavior 37

5.9.5 Socioeconomic Variables and Purchase Behavior 38 6. Conclusion, Limitations and Recommendations

6.1 Conclusions 41

6.2 Limitations 42

6.3 Managerial Recommendations 42

6.4 Academic Recommendations and further Research 43

7. References 45

8. Appendices 49

(6)

6

1

Introduction

Climate change and Environmentalism

In recent years, energy-efficiency has become a popular topic in society. Due to climate-studies (brought famously under attention by “an Inconvenient Truth”) and geo-political circumstances (oil crisis in the 80s, political instabilities in oil-regions) governments around the world are trying to persuade consumers to become increasingly energy-efficient. It seems as though (western) society is become increasingly aware of environmental issues. Companies have in the last decade also made an increased effort to act (or persuade society that they act) environmentally responsible. Car

manufacturers, luxury consumer goods producers, and others try to use energy efficiency as a product benefit.

Saving potential

The amount of electricity used for residential lighting is not known precisely. However, some estimations can and have been made (Wajer and Kemna (1991), Bertoldi and Atanasiu (2006), Mills (2002), Taskforce Verlichting(2008)). These estimations vary from between 498 and 524 kWh per household per year. For the Netherlands, this totals about 4 Twh per year.

According to the “Taskforce Verlichting” (2008), it would be possible to reduce the amount of electricity used for residential lighting by 50% in 2013 (2007 is the index year). This would result in electricity savings of about 2 Twh per year, making one medium-sized power plant obsolete. Legislation

In December 2008, the EU has set a target of “reducing energy consumption by 20% through increased energy efficiency” for 2020. In order to achieve this ambitious goal, several initiatives have been undertaken.

(7)

7 Picture 1: EU Ban on Bulbs (EU Regulations, 2008)

Electricity Prices

Apart from the EU regulations, the electricity prices have been fluctuating. In the period from 1996-2006, the price of electricity almost doubled (Statline, CBS, 2012). After 1996-2006, prices dropped a little. However, with prices fluctuating 10% or more each year, there is a lot of uncertainty concerning the future price development of electricity. Many analysts expect a large increase in prices the coming decade.

A brief history of luminary devices

In modern society, most people take the ability to properly see their surroundings for granted. However, before the mid-eighteenth century, people lived in near-complete darkness except in the presence of sunlight and moonlight (Fouquet and Pearson, 2006). Low quality candles, oil lamps and torches provided some relieve of darkness but these where relatively expensive, hazardous and provided poor levels of illumination. Since then, the use, effectiveness and efficiency of luminary devices have drastically increased. This increase has not been steadily however.

Incandescent lamps

As Gendre (2003) shows for the case of incandescent light bulbs, incremental innovations resulted in an increase in efficiency of 2 lumens of light per Watt in 1879 (Thomas Edison's first economically viable light bulb) to 12 lumens of light per Watt in 1912. It took about 50 years for a next step was made, cranking efficacy levels up to 26 lm/W. Most incandescent light bulbs sold today have an efficacy of between 13 and 30 lm/W, translating in an efficiency of about 2-4%. The remainder of the bulbs' output is mostly heat (95%). To state Gendre (2003): “Their pathetic efficacies make

incandescent lamps more suitable for heating purpose than lighting.”

Since then (100 years exactly!), the luminous efficacy has stayed more or less the same.

Other types of lamps

(8)

8 Up till recently however, lamps other than incandescent lamps have not been embraced by consumers for household purposes, for different reasons.

Conclusion

Due to the developments described above, especially the EU-instigated Ban on Bulbs policy, the use of incandescent light bulbs will marginalize in the EU.

Although there have been various attempts to persuade consumers to adopt energy-efficient lighting in the past and legislation has been made and implemented by the EU, there is little known about how consumers decide which type of lamp they buy. There has been little research in this subject, most research concentrating on the adoption of CFL's. Therefore, it is unclear what results this Ban on Bulbs policy will have on consumer decision making.

1.2 Problem statement

There is little information concerning consumer decision making processes in household lighting. This information is essential in understanding the adoption of (new) technologies so effective marketing strategies and/or government policies can be developed.

Research goal

Describing variables that influence consumers when deciding upon the purchase of household lighting. Main research question

Which variables influence consumers when deciding upon the purchase of household lighting? Which attributes do consumers find important?

1.3 Structure

First, all concepts which are dealing with our problem statement are discussed. In case possible, literature is used. This will result in a conceptual model. Then, an exploratory research will be

(9)
(10)

10

2

Literature Review

2.1 The Theory of Planned Behavior

The purpose of this study is describing consumer behavior and identifying variables which influence decision making processes. Behavioral attitudes are therefore of great importance. A suitable framework for explaining consumer behavior is the Theory of Planned Behavior (TPB). This study will use this framework and expand it. The TPB has been proven itself in numerous studies covering a wide array of topics (Smith et al, 2007).

Developed by Icek Ajzen (1991), the TPB is a further development of the theory of reasoned action which was proposed by Martin Fischbein together with Icek Ajzen in 1975. According to this theory, Behavioral intention is a function of attitude toward a behavior and subjective norms toward that behavior. Behavioral intention is a predictor for actual behavior.

As a revision and extension of this model, Azjen added a variable. Perceived behavioral control, defined as an individual's perceived ease or difficulty of performing the particular behavior (Ajzen, 1991), was added “to account for times when people have the intention of carrying out a behavior, but the actual behavior is thwarted because they lack confidence or control over behavior" (Miller, 2005, p. 127).

(11)

11 Limitations of the TPB

The theory of planned behavior can be a strong framework explaining consumer behavior. However, it has its limitations. First, behavioral intentions may not be needed as a mediator between attitude, subjective norm, behavioral control and behavior (Bentler & Speckart, 1979).

Secondly, the theory assumes rationality and systematic decision making. Unconscious motives and habitual behavior are therefore not taken into account.

Third, the theory uses internal factors to explain behavior (and/or behavioral intentions). External factors however, are not explicitly taken into account. There may however be other factors that influence behavior. Empirical studies have shown that only 40% of the variance of behavior could be explained by the TPB (Ajzen 1991; Werner 2004). This decreases the TPB's power for behavioral predictions. Also, from a marketing perspective, since external/situational factors are not taken into account, the TPB lacks a strong base for segmentation.

In an attempt to compensate for these limitations, the TPB will be adjusted by an extension of the TPB using external factors and the omission of behavioral intentions. In the remainder of this chapter the components of the TPB will be discussed first, after which these extensions will be discussed. At the end of this chapter a conceptual model will be presented.

2.2 Attitude Toward Behavior

Attitude(s) toward behavior refers to the degree to which a person holds an attitude towards a particular behavior. This attitude is a function of a belief about consequences of behavior and the evaluation of these consequences (Azjen, 1991). In the case of household lighting, an attitude is thus formed toward a particular type of lamp by a consumer's perception of a behavioral consequence and the evaluation of this consequence.

These behavioral consequences are expected to be mostly related to the lamp (type) characteristics, or product attributes. For instance, if a consumer believes buying and using an incandescent light bulb (purchase behavior) has the consequence of an atmospheric light dispersion (product attribute), an evaluation is made. If this consumer desires such an atmospheric light dispersion (evaluation of the behavioral consequence), a positive attitude is formed and this will likely result in the consumer buying an incandescent light bulb (depending on other attitudes).

There is a wide array of attitudes that could impact the buying behavior of consumers in household lighting. Since little to none research is publicly available at the moment, this study will try to complement this “blind spot”, testing a wide array of possible variables that might affect purchase behavior.

First, an overview of existing lamp types and characteristics will be provided, therefore identifying relevant product attributes. As just discussed, these are expected to influence behavioral attitudes. Thereafter, Environmental attitude and Innovative attitude will be discussed.

2.2.1 Lamp types and characteristics

There is a wide and increasing range of lighting products available to consumers. Without extensively discussing technical details, the most common types and their characteristics will be discussed in this chapter. These characteristics are the attributes that are assumed to determine consumers' attitudes toward behavior.

(12)

12

Incandescent light bulbs

Compact Fluorescent Lamp (CFL)

Light Emitting Diode lamp (LED)

Halogen Lamp Lamp characteristics

A lamp has a number of characteristics. These can be split up into quantitative light characteristics, qualitative light characteristics, aesthetic characteristics and costs.

Quantitative Light Characteristics

Illumination

The distribution of light on a horizontal surface. The purpose of all lighting is to produce illumination.

Lumen

A measurement of light emitted by a lamp. As reference, a 100-watt incandescent lamp emits about 1600 lumens.

Footcandle

A measurement of the intensity of illumination. A footcandle is the illumination produced by one lumen distributed over a 1-square-foot area. For most home and office work, 30–50 footcandles of illumination is sufficient. For detailed work, 200 footcandles of illumination or more allows more accuracy and less eyestrain. For simply finding one's way around at night, 5–20 footcandles may be sufficient.

Qualitative Light Characteristics

Color temperature

The color of the light source. By convention, yellow-red colors are considered warm, and blue-green colors are considered cool. Color temperature is measured in Kelvin (K) temperature. Higher Kelvin temperatures (3600–5500 K) are what we consider cool and lower color temperatures (2700–3000 K) are considered warm.

Color rendition

How colors appear when illuminated by a light source. Most objects are not a single color, but a combination of many colors. Light sources that are deficient in certain colors may change the apparent color of an object. The Color Rendition Index (CRI) is a 1–100 scale that measures a light source's ability to render colors the same way sunlight does. The top value of the CRI scale (100) is based on illumination by a 100-watt incandescent light bulb.

Glare

The excessive brightness from a direct light source that makes it difficult to see what one wishes to see. Glare is primarily the result of relative placement of light sources and the objects being viewed.

Ramp-up time

Ramp-up time describes the duration in which a lamp reaches its maximum illumination. CFL's, especially early generations, take a notoriously long time to “ramp-up”.

Light spread

The directional behavior of the light beam emitted by a lamp.

Dimming Possibility

(13)

13 Aesthetic Characteristics

These entail the shape, size and general appearance of the lamp itself. Costs

Retail Price

The price for which a lamp is purchased by the consumer.

Usage Costs

The result of the use of the lamp in terms of electricity. These usage costs can be split up further by monetary consequences (electric bill) and environmental consequences (co2 emission).

The lamp characteristics just discussed could all be of importance in forming attitudes toward behavior and therefore influencing it. In Chapter 4, these characteristics will be subject of an exploratory research. The purpose is to assess which of these will prove fruitful for usage in a conclusive research (Chapter 5)

2.2.2 Environmental Attitude

In past research, environmental attitude has been linked to a number of environmentally friendly behaviors. For example, environmental attitude has been linked to avoiding non-disposable

merchandise, looking for recyclable products, and buying biodegradable goods (Laroche et al., 2001). This is in correspondence with the cognitive consistency theory (Festinger, 1957), which entails the idea that a person who is concerned about (ecological) problems, will be motivated to take actions that will minimize them.

Several studies have shown that “environmentally sensitive” consumers are very likely to change their behavior and purchase products that they perceive to be “environmentally friendly” (e.g. Kilbourne &Pickett, 2008; Mostafa, 2007; Schlegelmilch et al., 1996).

Since the use of electricity has environmental consequences, it is expected that environmental attitude has influence on buying behavior of consumers in household lighting.

2.2.3 Innovative Attitude

Consumer innovativeness (or innovative attitude) has been extensively linked to the adoption and diffusion of new products in much research (e.g. Rogers, 1995, Steenkamp & Gielens, 2003). The overall assertion is that innovative consumers are among the early adopters of new products (Rogers, 1995). Much effort has therefore been made in identifying traits and characteristics that may predict consumers’ innovative adoption behavior.

Since we are dealing with several types of lamps, it is possible that innovative attitude affects purchase behavior in the case of household lighting. LED as lighting bulbs are a relatively new technology, thus its adoption is possibly affected by consumers' innovative attitude.

2.3 Subjective Norms

(14)

14 2.4 Perceived Behavioral control

In the Theory of Planned Behavior, perceived behavioral control refers to an individual's perceived ease or difficulty of performing a particular behavior (Ajzen, 1988). In household lighting, retail prices as well as usage costs might impact this perceived behavioral control.

Retail prices

Retail prices vary widely between the available lamps. Incandescent bulbs are available from 0,20 eurocents while a single LED lamp can cost more than 70,- euros. These retail prices might influence perceived behavioral control.

Usage Costs

Every electric appliance uses electricity and lamps are of course no exception. Electricity use has monetary consequences (electric bill) and environmental consequences (co2 emission). Both of these usage cost might influence perceived behavioral control.

The previously discussed sections are all dealing with the traditional components of the TPB. Next, two additional components, which are expected to benefit the applicability of the TPB in explaining consumer behavior are discussed.

2.5 Location

There has been no research regarding the influence of a lamps’ location on the consumer decision making process. However, location is expected to be of great importance. The location of a lamp is related to its function, visibility and usage intensity.

Function

The function of a lamp is interrelated with its location. For example, if the function of a lamp is to illuminate a kitchen surface, its location will be near the kitchen (mostly above). If the function of a lamp is to illuminate an attic, providing save access and overview, it will be placed in this attic. There can be made a distinction between two primary functions:

Functional lighting: Providing illumination to perform tasks

Atmospheric Lighting: Providing illumination to create a desired atmosphere

A third function sometimes mentioned is accent lighting. This is referred to the situation when the function of a lamp is providing illumination to accent a detail, for example a painting. Since this function is a combination of functional and atmospheric lighting, this third function shall not be taken into account in this research.

Visibility

The location of a lamp will determine its visibility. Visibility is here referred to as the degree of which a lamp as well as its illumination is visible. The fitting and/or armature of a lamp also affects its visibility, and can in this respect be viewed as integral part of location.

Examples of highly visible lamps are a dining table luminaire or hallway chandelier. A garden lamp or attic lamp are less visible in everyday life.

Usage Intensity

(15)

15 amount of time it is switched on. Some lamps are used almost all day, while others are (almost) never used.

Location, as a function of these three variables, is expected to have an influence on the consumer decision making process. Desired lamp attributes will vary depending on location. For instance, a consumer is expected to desire other characteristics from a desk lamp than his living room lamp. Since no literature has been published on this subject, the topic of location will be subject to an exploratory research, discussed in chapter 4.

2.6 Socioeconomic Situation

There may well be other variables affecting purchase behavior not previously discussed.

Although the topic of consumer behavior with respect to residential lighting has not been extensively researched, some research has been undertaken, mostly dealing with the (non)uptake and diffusion of cfl’s. Early examples of these are a research by Scott (1997) and some research by Kumar et al.(2003). Scott (1997), who’s research took place in Ireland, finds few relations between uptake of CFL’s and socio-demographic characteristics. Kumar et al.2003) find a positive relation between income and education and the uptake of cfl’s in India.

Building upon this research, Mills & Schleich (2010) have tried to extend the research further. Using a double hurdle model, they make a distinction between adoption barriers/market entry and adoption intensity, relating this to socio-demographic characteristics. Market entry is referred to here as to the purchase of (at least) a single CFL. Adoption intensity measures the number of CFL’s purchased. They find some interesting results. These include a positive correlation between the following characteristics and market entry for CFL’s : floor space, detached houses, household size, young children, retirees, retail price, household awareness of the energy situation.

A negative correlation is proven for the low(est) income groups and senior management position or skilled professional and market entry. A perhaps interesting result was that they found no relation between the high income groups and market entry.

For the adoption intensity, positive relations are found with: residence size, detached homes, pre 1951 build houses, retirees, household size, lower income groups and age (positive but declining). A

negative relation was found with the relative number of sockets and the higher incomes. Children and regional power price differences had no influence on the adoption intensity.

Mills and Schleich (2010) conclude that barriers to market entry of cfl’s are low for all households except the lowest income classes.

(16)

16

2.7 Conceptual Model

Based on this literature review, the following conceptual model is developed.

The following research questions follow from the conceptual model:

RQ1: To what extend does the Theory of Planned Behavior describe consumer purchase behavior? RQ2: To what extend does lamp location influence the relative importance of the variables described

in the Theory of Planned Behavior?

RQ3: To what extend do socioeconomic variables influence the relative importance of the variables described in the Theory of Planned Behavior?

RQ4: To what extend does lamp location influence consumer purchase behavior?

(17)

17

3

Research Methodology

In the previous chapter, a literature review has been used in order to develop a suitable framework which can describe consumer behavior in residential lighting. Since not much research has been published in this topic, an exploratory research has been used to further strengthen the conceptual model. After refining this conceptual model it is empirically tested, using a conclusive research design.

(18)

18

4

Exploratory Research

4.1 Research Methodology

The explorative part of this research consists of three parts. First, industry experts were asked to shed their light upon attributes which they thought consumers deemed important. The outcome of these interviews where the input of a focus group meeting. Finally, a questionnaire was distributed within a small sample in order to assess in which way lamp locations can best be operationalized (allowing the most possible effects of locations’ constructs on consumer behavior, as discussed in chapter 2). Due to the small scale of the exploratory part and clarity, the methods as well as the outcomes of the

exploratory research will be discussed briefly in the subsequent sections. 4.2 Industry experts

Since little (recent) literature on the topic of residential lighting exists, industry experts where consulted. It seemed wise to select those people with a lot of exposure to consumers, making them aware of the sentiments and behavioral aspects of these consumers. Therefore, three companies who mainly sold lighting appliances to consumers where selected.

Industry experts where selected based on direct experience with consumers in their search for residential lighting. Convenience sampling led to interviews with representatives/employees of two residential lighting specific shops in Groningen. These where “Lighthouse”

(www.lighthouse-groningen.nl) and “De Compaen” (www.decompaen.nl). These interviews where semi-structured (Appendix 1) in order to cover relevant information.

The purpose of these interviews was twofold: getting a grasp of the typical consumer and the way in which they view household lighting and to establish the product attributes which consumers used in their decision making process as well as the probable importance of those attributes. A list of possible attributes was provided as guidance (Appendix 1). Also, the level of knowledge of these consumers was asked to be estimated.

Results

The main result was a list of product attributes these industry experts expected to be of importance for consumers. Both companies were under the impression most consumers did not make much distinction between attributes. Rather, a more “holistic view” was suggested. When asked which of the attributes suggested would mainly be part of this holistic view, some were discarded. These are illumination, footcandle, glare and color rendering. The other suggested attributes where expected to be of greater importance. Also, these industry experts suggested that customers typically have

relatively little knowledge on the subject and are easy to influence. 4.3 Focus Group

In order to get a grasp of the way in which consumers view household lighting and some related topics which could be of value for this research, a focus group was set up.

(19)

19 Results

Knowledge on lighting

There was a consensus that the participants knew little about the technical aspects of lamps. Some quotes: “I am not used to think about lamps. They are just there and if they break, I’ll buy the same one.”

“I know the terms like Watt and stuff but not exactly what they mean…”

Difference between lamp types

Respondents agreed upon large differences between the several lamp types. Some strong negative sentiments came across when discussing CFL’s.

“I dislike CFL’s. The warm-up time is very impractical and they don’t shine as bright as an incandescent light bulb.”

“CFL’s are expensive and I think they break just as often as regular lamps! Besides, I had one in the bathroom and it lasted like a minute before I could properly see!”

On LED’s:

“I haven’t got any but I hear they are very cold. Also, I wouldn’t pay 25,- euro for a lamp!”

Lamp attributes

When discussing lamp attributes the view of the industry experts was confirmed at first. The focus group participants were not used to think in terms of (technical) attributes. Rather, a more holistic approach was taken. Most frequently, purchase behavior was the direct result of replacement with the same lamp as the previously owned.

When stressed however, all of the attributes suggested by industry experts were more or less confirmed. Actually, when discussing the previously discarded attributes (illumination, footcandle, glare and color rendering), color rendering was suggested to be of importance sometimes as well. Therefore, the following product attributes will be used for the conclusive/qualitative research (Chapter 5) :  Color Temperature Light Spread Color Rendering Light Quantity Ramp-up time Dimming Possibility Esthetics Retail Price

Electricity use, cost consequences

(20)

20 4.4 Survey (Appendix 2)

Since one of the variable suspected to influence consumers' purchase criteria is the lamp location, a small quantitative research was undertaken.

Using email, 25 questionnaires were distributed. This led to 10 respondents.

In this research, consumers were asked to rate several locations on Visibility, Usage intensity and Function, since it is expected that these variables are key to the influence location has on consumer behavior. Assuming an intrinsic impact of each of the variables just described, the goal of this research was to find locations where variables would fluctuate the most on each respective (variable)spectrum. Results

It was concluded that the most suitable lamp locations (covering each end of the spectra mentioned above the most) are a Desk lamp, Plafonniere/Ceiling lamp and a Living room lamp.

Conclusion

(21)

21

5

Quantitative Research

In the previous chapter, a (preliminary) exploratory research has been described. Processing the outcomes of that research has led to a revised conceptual model. This chapter will be dealing with testing this conceptual model empirically and answering the research question resulting from the model. For this end, a questionnaire was developed.

5.1 Sample and Procedure

Since the population subject to this study consists of all Dutch consumers of household lamps, no further selection criteria are used for the sampling purposes. A convenience sampling method is chosen, spreading the questionnaire using a snowball method in the author’s surroundings. 5.2 Operationalization of constructs

Before testing, all constructs used in the conceptual model need to be operationalized. A balance has been sought between completeness, validity and questionnaire (time)constraints.

Location

Resulting from the exploratory research, Location is operationalized through testing the importance of all product attributes on three locations. A Desk lamp, Plafonniere/Ceiling lamp and a Living room lamp where used.

Attitude toward Behavior

Respondents are asked to state the importance of the following product attributes:

Color Temperature Light Spread Color Rendering Light Quantity Ramp-up time Dimming Possibility Esthetics

These stated importance’s are assumed to be a reflection of their attitude toward that attribute. Environmental Attitude

For measuring Environmental Attitude, the New Environmental Paradigm (NEP) was used. Containing 9 questions, this scale has proven its value for decades (Dunlap et al., 2008).

Innovative Attitude

Respondents were asked how much they agreed with 6 statements to determine their innovativeness. These statements are derived from the Domain Specific Innovativeness (DSI) scale, developed by Goldsmith and Hofacker(1991), which is an established tool to measure Domain Specific

(22)

22 Subjective Norms

Subjective norms are operationalized by respondents stating to which degree they are influenced by (significant) others. Colleagues, Partners, Children and Friends where considered to be the possible significant others. Therefore, all of these are tested in this respect.

Perceived Behavioral Control

Respondents are asked to state the importance of the following product attributes:

 Retail Price

 Electricity use, cost consequences

 Electricity use, environmental consequences

These stated importance’s are assumed to be a reflection of Perceived Behavioral control. Purchase Behavior

Purchase Behavior is measured by respondents stating which lamp they have bought last for each location. Incandescent, CFL, LED or Halogen were the possible choices in this respect.

Socioeconomic Situation

A broad definition of socioeconomic situation is used. Apart from some standard variables (age, gender, income, marital status) respondents were asked some specific information about their residential situation and lighting situation. The following variables were disclosed by respondents:

- Gender - Income - Education - Marital Status - Residence Type - Residence Owner

- Last Reconstruction or Last Move - Lamp-purchase Decision making - Responsibility Electricity Bill - Age

- Household Size

- Number of lamps in use 5.3 Estimation and assessment

After data collection, SPSS 16 will be used to process and analyze data. A variety of analytic techniques and procedures will be utilized in order to properly answer the relevant research questions. These include Factor Analysis for possibly identifying latent variables , One Way ANOVA and Chi-square for testing variance, and, as Post Hoc statistical test, Fishers (unprotected) Least Significant Difference tests (LSD). The standard confidence level of 0,05 is used for all testing purposes.

5.4 Measurements and Scaling

(23)

23 5.5 Results

In the following sections, the results of this research will be discussed. Procedural comments will be given in the first section, after which the sample will be discussed. Then a factor analysis is done, after which there will be a section dealing with reliability of constructs. Thereafter, chapter 5.9 and further will be dealing with testing the conceptual model.

Procedure

Since the quantitative part of this research entails researching a vast number of variables and relations, some choices have to be made for the purpose of overview. To this end, it is chosen to discuss each research question, as discussed in the conceptual model, separate. Each chapter will remind the reader which relation is investigated, as to provide a clear picture of overview.

Also, scales are constructed from variables as much as possible. To this end, a factor analysis is done for the relevant product attributes and the variables environmental attitude and innovativeness are constructed, but not after it is established that this is appropriate by a reliability analysis.

5.6 Sample Descriptives

The sample consists of 98 respondents. Due to the length of the questionnaire, some respondents have not completed the questionnaire or skipped questions. 77 Respondents provided all answers. Since the snowball method is used, it is unclear what the response percentage is.

First, an overview of the demographics of the sample is provided. Thereafter, some other important statistics, dealing with the TPB, are displayed.

5.6.1 Demographics

The average age of the respondents is 42 years, with the youngest being 21 and the oldest 66 years old.

Frequency Percentage

Gender Male 56 57,7%

Female 41 42,3%

Marital Status Single 38 39,2%

Married 38 39,2% Cohabit 17 17,5% Household Size 1 2 3 4 5 6 34 35 8 17 2 1 35,1% 36,1% 8,2% 17,5% 2,1% 1%

Income Less than Modal

Modal

(24)

24

MBO 4 4.1%

HBO 45 46.4%

WO 42 43.3%

Table 1, Demographic statistics.

Housing

Housing type Family home, detached 31

Semi-detached 10 Row-house 18 Apartment/flat 21 Single room 10 Other 7 Owner Yes 64 No 33

Last time moved or reconstruction

<1 year 17

1-5 Years 44

5-10 Years 16

>10 Years 20

Table 2, Residential statistics.

Representativeness of sample

When comparing the demographic statistics of the sample to the Dutch population in 2012 (as described by the CBS), the following discrepancies should be acknowledged:

 Males are slightly overrepresented

 The level of education is higher than the Dutch population. 5.6.2 Product Attributes

Since part of the research question (as discussed in chapter 1) is dealing with which product attributes are relevant to consumers, the descriptives dealing with these attributes are rather important.

Respondents were asked to state the importance of these attributes on a 7 point Likert-Scale.

Attitudes Toward Behavior Mean Desk Lamp

Mean Living Room Lamp

Mean Ceiling

Lamp Mean Total

(25)

25 Perceived Behavioral Control

Retail Price 5,32 5,00 5,20 5,16

Electricity use, cost consequences 4,13 4,23 4,22 4,18

Electricity use, environmental consequences 4,22 4,26 4,31 4,27

Table 3, Mean scores Product Attributes per Location.

The mean scores range from 3,81 to 6,00. On a 7 point scale, this is not a dramatic difference. All attributes used in the questionnaire seem to be important or very important to consumers, with light quantity being the most important and start-up time being the least important.

A OneWay Anova analysis will be used to analyze if scores vary significantly based upon their location. This will be discussed in chapter 5.9, where all main research questions resulting from the conceptual model are discussed.

5.6.3 Subjective Norms

When asked which opinions respondents take into account when deciding which lamp to buy, partners were most influential. However, mean scores are pretty low all together (all being measured at a 7 point Likert Scale), meaning that no great importance is said to be given on the opinions of others altogether, although partners do have some influence.

Mean Std. Deviation

Opinion Partner 3.96 2.73

Opinion Friends 1.97 1.76

Opinion Children 1.67 2.03

Opinion colleagues 1.35 1.22

(26)

26 5.7 Factor Analysis

Since not much is known about lamp attributes influencing purchase behavior, the variables derived from the preliminary research are subjected to a factor analysis. The purpose of this analysis is twofold; discovering latent (or at least interrelated) variables and reducing the number of variables (which is helpful for further analysis).

The variables included in this analysis:

Color Temperature Light Spread Color Rendering Light Quantity Ramp-up time Dimming Possibility Esthetics Retail Price

Electricity use, cost consequences

Electricity use, environmental consequences

According to Malhotra (2007), the test is appropriate when the Kaiser-Meyer-Olkin (KMO) measure of sampling adequacy is higher than 0,5 and when the Bartlett’s test of Sphericity is lower than α = 0,05.

KMO and Bartlett's Test

Kaiser-Meyer-Olkin Measure of Sampling Adequacy.

,636

Bartlett's Test of Sphericity Approx. Chi-Square 705,468

Df 45

Sig. ,000

Table 5, KMO and Bartlett’s Test

In table 5, it is shown that: KMO=0,636 and BTS=0,000 <0,05. Therefore, a factor analysis is appropriate.

After establishing the appropriateness of the factor analysis, the amount of factors has to be established. Three conditions need to be met in order to use factors (Malhotra, 2007):

 Eigenvalues > 1

 Factor Variance explained > 5%

(27)

27

Total Variance Explained

Comp onent Initial Eigenvalues Extraction Sums of Squared Loadings

Total % of Variance Cumulative %

1 3,050 30,501 30,501 2 1,481 14,814 45,315 3 1,291 12,908 58,223 4 ,983 9,826 68,049 5 ,845 8,455 76,504 6 ,723 7,235 83,739 7 ,548 5,481 89,220 8 ,477 4,768 93,988 9 ,420 4,202 98,190 10 ,181 1,810 100,000

Table 6 Factor Analysis: Eigenvalues and Variance

Table 6 shows that the cumulative variance explained is <60% for components with eigenvalues>1, therefore the necessary conditions have not been met. After discarding Light Quantity and Ramp-up Time (which have maximum loads<0,5) another test is done. Again passing the KMO and Bartlett's test ( KMO=0,621 and BTS=0,000 <0,05) it is apparent all conditions are met (table 7).Three factors explain 68,393% of variance (>60% needed), all having an eigenvalue above 1 and independently explain more than 15% each (> 5% needed). Therefore, this factor analysis seems suitable.

Total Variance Explained

Comp onent Initial Eigenvalues Extraction Sums of Squared Loadings Rotation Sums of Squared Loadings

Total % of Variance Cumulative %

1 2,732 34,156 34,156 2 1,469 18,367 52,523 3 1,270 15,870 68,393 4 ,744 9,295 77,688 5 ,600 7,497 85,184 6 ,560 7,006 92,190 7 ,433 5,418 97,608 8 ,191 2,392 100,000

(28)

28 The rotated components matrix shows the following coinciding components:

1 2 3

Color Temperature (Kelvin) ,781 ,038 ,026

Light Spread/Dispersion ,688 ,287 ,153

Color Rendering (CRI) ,729 ,281 -,175

Esthetics ,362 -,080 ,734

Retail Price -,125 ,195 ,849

Electricity use, monetary consequences ,120 ,928 ,106

Electricity use, Environmental consequences ,144 ,895 ,008

Dimming Possibility ,617 -,085 ,301

Table 8: Rotated component table (Varimax)

Next step is checking the factors for reliability. For factor 1, 4 variables are checked for reliability using Cronbach's Alpha. Since Cronbach's Alpha is more then 0,6 this factor has a high reliability and

consistency and it is usable for further analysis. The same goes for factor 2. Cronbach's

Alpha N of Items

Factor 1 .695 4

Factor 2 .869 2

Factor 3 .518 2

Table 9 Reliability Analysis

However, factor 3 has an Cronbach's Alpha below 0.6. Therefore, according to Malhotra (2007), it cannot be used for further analysis. Although other authors claim that a Cronbach's Alpha of >0.5 is sufficient (e.g. Bowling (1997)) The 0.6 boundary is used in this research. With this in mind, the underlying variables of factor 3 will be tested individually.

5.7.1 Interpretation of Factors

Since factor 1 is dealing with attributes that all encompass certain atmospheric properties, the underlying variables will be constructed into a new scale titled Atmospheric Criteria. Factor 2 is revealing attributes that are dealing with implications of use. Therefore, a new scale is constructed using these attributes, to be titled Usage Implications. Note that “Atmospheric Criteria” are part of the attitudes toward behavior, while “Usage Implication” are part of Perceived Behavioral Control,

enforcing the distinction between these attributes made by the Theory of Planned Behavior.

Factor 1 “Atmospheric Criteria” (Cronbach's Alpha 0,695)

Color Temperature

Light Dispersion

Color Rendering

(29)

29

Factor 2 “Usage Implications” (Cronbach's Alpha 0,869)

Electricity use, Monetary Implications

Electricity use, Environmental Implications

In the upcoming sections, analyses will thus make use of the following lamp attributes/variables:

Atmospheric Criteria (Factor 1)

Usage Implications (Factor 2)

Ramp-up Time

Retail Price

Light Quantity

Esthetics

5.8 Scale Construction

This chapter will be dealing with the operationalization of several constructs. These will be recoded, tested for reliability and then, if possible, converted into a single variable. These constructs are:

Environmental Attitude

Innovativeness 5.8.1 Environmental attitude

Respondents were asked to rate 9 statements on agree-ability using a 7-point Likert Scale. These statements are derived from the NEP-scale (New Environmental Paradigm), which is a widely accepted scale for measuring environmental attitude. Of these 9 statement, 4 where formulated in a positive way, 5 in a negative way. After recoding, a reliability test shows a high Cronbach’s alpha (table 10), ensuring a high enough reliability and consistency to create with these nine items the new variable Environmental Attitude.

5.8.2 Innovative Attitude

Respondents were asked how much they agreed with 6 statements to determine their innovativeness. These statements are derived from the Domain Specific Innovativeness (DSI) scale, developed by Goldsmith and Hofacker(1991), which is an established tool to measure Domain Specific

Innovativeness.

Also, respondents were asked if they had already bought a LED lamp. One of the statement was negatively formulated, while 5 where formulated in a positive way. After recoding, a reliability test shows a Cronbach's alpha of 0,532. One of the statements was discarded, which resulted in a

Cronbach’s alpha of 0,633. Surprisingly, excluding the question whether respondents already bought a LED lamp resulted in an even higher reliability: 0,713. These internal five items where compressed into the variable Innovativeness.

Construct Number of questions Cronbach's Alpha Mean Score Std Deviation

Environmental attitude 9 ,789 5,12 ,87

Innovativeness 5 ,713 2,26 ,74

(30)

30 5.9 Research Questions

The following section will be dealing with all research questions as described in the conceptual model.

RQ1: To what extend does the Theory of Planned Behavior describe consumer purchase behavior? RQ2: To what extend does lamp location influence the relative importance of the variables described

in the Theory of Planned Behavior?

RQ3: To what extend do socioeconomic variables influence the relative importance of the variables described in the Theory of Planned Behavior?

RQ4: To what extend does lamp location influence consumer purchase behavior?

(31)

31 5.9.1 The Theory of Planned Behavior vs Purchase Behavior (RQ1)

This chapter will answer the following research question:

To what extend does the Theory of Planned Behavior describe consumer purchase behavior? Since the TPB consists of three components (attitude toward behavior, subjective norms and perceived behavioral control), each will be dealt with separately.

Attitude toward behavior

To test if there is variance in the type of lamp purchased (purchase behavior) and the attitudes toward behavior expressed by respondents, a One-Way ANOVA test is done.

Independent variables are the purchase criteria.

Dependent variables: Type of lamp purchased (Incandescent, Halogen, CFL, LED)

Homogeneity of Variance

Because one of the assumptions of analyses of variance is that variance within each group is equal, this is tested before the analysis. Two options are available in this respect: The Levene test of homogeneity and the rule of thumb that the difference in value of standard deviations should not exceed 2 (Huizingh, 2006). This research opts for the rule of thumb.

The standard deviations within the groups are between 0.6 and 2.1, thus not exceeding the critical value of the rule of thumb. Therefore, we can assume variance within each group is equal (enough).

Results

The result of this analysis show a significant variance for “Light Quantity” [F(4,258) = 2,583, p = 0,038] as well as “Ramp-up Time” [F(4,255) = 3,711, p = 0,006]. Esthetics show a non-significant

trend[F(4,257) = 2,206, p = 0,069], as well as Innovative Attitude [F(4,85) = 2,355, p = 0,061]. Atmospheric Criteria and Environmental attitude show no significance.

Post Hoc Test (LSD)

Using Fishers (unprotected)Least Significant Difference tests (LSD), some (other) significant results surfaced.

(32)

32 Ramp-up Time was more important to respondents who had purchased a halogen lamp than any other type of lamp. Surprisingly, ramp-up time was less important for respondents who ended up buying a LED lamp (which technical characteristics allow for a negligible ramp-up time).

Another somewhat surprising result is that esthetics where less important for respondents who had bought an incandescent lamp (which is often considered to be most esthetically pleasing) than other types of lamps.

(33)

33 not know which type of lamp they have bought. Surprisingly, the most innovative respondents chose a CFL as opposed to a LED type lamp (which technique is far more recent). Perhaps this is the result of rebates or introductory offers made by LED (re)sellers as a market introduction tactic, therefore successfully persuading non-innovative consumers to enter the LED light bulb market.

Respondents who scored higher on Environmental attitude tend to choose LED lamps. Respondents with a low score respectively tend to be unaware of the last purchased lamp type. This confirms the positive relations of environmental attitude with energy-efficient behavior.

Finally, as expected, respondents ending up buying a CFL deemed atmospheric criteria less important than respondents who purchased a halogen lamp.

Subjective Norms

(34)

34 Perceived Behavioral Control

There is no significant relation between PBC and the type of lamp purchased. Involvement

When looking at the types of lamp purchased and the average scores on the variables included in the TPB, a pattern emerges. Both for respondents who didn’t know which type of lamp they had bought as well as for respondents who bought incandescent light bulbs, the averages scores of almost all

constructs are lower. This could be caused by variables not taken into account by the TPB. Another reason could be that the decision making process was less elaborate for these respondents. This could be the result of low involvement in the decision process, for whatever reason. Since perceived

behavioral control (operationalized by costs) shows no significant relation, this suggested low involvement does not lead to a significant rise in price sensitivity (associated with low involvement). That being said, especially for respondent unaware of the type of lamp they had purchased, the suggestion of low involvement seems likely.

5.9.2 Location vs Purchase Behavior (RQ2)

This chapter will answer the following research question:

RQ2: To what extend does lamp location influence the relative importance of the variables described in the Theory of Planned Behavior?

This research attempts to discover variables influencing the variables within the TPB of consumers when they decide to buy a lamp. From the exploratory research, it is hypothesized that a relation exists between the location of a lamp and the consumer’s attitudes toward behavior. For instance, atmospheric criteria are expected to be more important for a living room lamp than a desk lamp. Therefore, an Anova test is done to check for differences within the disclosed attitudes and behavioral control variables and their respective weight distributions among the three tested locations.

(35)

35 In this Anova analysis the dependent variables are:

Atmospheric criteria Light Quantity Ramp-up Time Esthetics Purchase Price Usage Implications

The independent variables are the locations (expressed by Desklamp, Plafonniere and Living room lamp).

Homogeneity of Variance

In this case, the standard deviations within the groups range between 1 and 2, thus not exceeding the critical value of the rule of thumb. Therefore, we can assume variance within each group is equal (enough).

Results

(36)

36 5.9.3 Socioeconomic variables vs the Theory of Planned Behavior (RQ3)

In this section, the following research question will be discussed:

RQ3: To what extend do socioeconomic variables influence the relative importance of the variables described in the Theory of Planned Behavior?

Since there are 12 socioeconomic variables and 16 variables within the TPB, 192 analyses should be performed to answer this research question completely. Chapter 5.9.1 showed that of the 16 variables describing the TPB only 2 show a real significant relation with behavior (light quantity and ramp-up time). Therefore, the other 14 are discarded for these analyses. Analyses were performed for the following variables, testing for influence on Light Quantity and Ramp-up Time:

- Gender - Income - Education - Marital Status - Residence Type - Residence Owner

- Last Reconstruction or Last Move - Lamp-purchase Decision making - Responsibility Electricity Bill - Age

- Household Size

- Number of lamps in use

(37)

37 Another significant relation [F(3, 94 = 4.928, p = 0,003] was found between the last move or

reconstruction of respondents and their attitude towards ramp-up time. Respondents who hadn't moved or reconstructed recently deemed ramp-up time more important. Perhaps they have

experienced or are experiencing the consequences of early generation cfl's, which are known to have long ramp-up times.

None of the socioeconomic variables found to influence cfl adoption or market entry by Mills & Schleich (2010) (as discussed in chapter 2) where found to be significant.

5.9.4 Location vs. type of lamp purchased (RQ4)

In this section, the following research question will be answered:

RQ4: To what extend does lamp location influence consumer purchase behavior?

(38)

38 As displayed in the following chart, the most popular lamp differs per location. This affirms the

suspected importance of lamp’s location on the purchase decision. Led is least popular in a ceiling armature, where CFL’s are the most popular choice. Incandescent lamps are evenly distributed. Relatively many people don’t know which type of lamp they have bought for their ceiling lamp. Somewhat surprising is the most popular lamp used at the desk: a Halogen type lamp.

5.9.5 Socioeconomic Situation vs. Purchase Behavior (RQ5)

To what extend do socioeconomic variables influence consumer purchase behavior?

Since socioeconomic status in this research is measured by variables on different scales, different tests are used, depending on the measurement scale of the respective variables. These variables are:

- Gender - Income - Education - Marital Status - Residence Type - Residence Owner

- Last Reconstruction or Last Move - Lamp-purchase Decision making - Responsibility Electricity Bill - Age

- Household Size

(39)

39 Chi-square tests are used for the first 9 variables, while the last three are subjected to a variance test (Oneway Anova).

Of the 12 variables, Last move or Reconstruction was found to be significant at (p= 0.036). Also, some relations were discovered in post hoc testing. These results will now be discussed.

Last Reconstruction or Last Move

Chi-Square Tests Value df Asymp. Sig. (2-sided) Pearson Chi-Square 22.155a 12 .036 Likelihood Ratio 25.418 12 .013 Linear-by-Linear Association .238 1 .626 N of Valid Cases 94

Most striking is the choice of lamps for new home owners or after reconstructions. Halogen type lamps seem to be the most popular choice, whereas incandescent bulbs are virtually nonexistent. This could be a result of the ban on bulbs. However, CFL’s are also relatively unpopular. This may be the result, as suggested before, of past experience with long ramp-up times, a characteristic common in early generation and cheap CFL’s. Where we might expect an uptake of LED lighting in relatively new residential surroundings, this is not the case. Perhaps the market for LED bulbs is not mature enough, causing uncertainty about attributes combined with high prices.

Gender

(40)

40 they seem to be ahead in the uptake of LED lighting. Also, CFL’s are rather unpopular with women. Retesting for differences in gender versus the TPB (RQ3) showed no significance. Therefore, these different preferences could not be explained by the TPB.

Number of lamps in Use

(41)

41

6

Conclusion, limitations and Recommendations

This chapter will deal with the conclusions of this research. After discussing the conclusions, possible limitations of this research are discussed. Academic and Managerial recommendations are provided in the subsequent section.

6.1 Conclusions

The main research question of this study was:

Which variables influence consumers when deciding upon the purchase of household lighting? Which attributes do consumers find important?

Product attributes

Qualitative and Quantitative research have led to the conclusion that the following attributes are used by consumers when deciding upon the purchase of household lighting:

Atmospheric Criteria (color temperature, light dispersion, color rendering, dimming possibility)

Esthetics

Ramp-up Time

Retail Price

Light Quantity

Usage Implications (cost of electricity, environmental implications)

The importance of these attributes vary, albeit slightly. They are all relatively important. The location of a lamp does have some influence: esthetics and atmospheric criteria are more important for a living room lamp than for a desk lamp or a ceiling lamp.

Research Questions

Research Questions Strength of the relation

Most important conclusions

RQ1 Medium Ramp-up Time and Light Quantity have the most

effect on purchase behavior

RQ2 Medium Atmospheric attributes and esthetics are most important for living room lamps

RQ3 Minimal Ramp-up time is more important for home owners who don't often move or reconstruct

RQ4 High CFL's and Halogen lamps are inversely popular as ceiling lamps and desk lamps respectively

RQ5 Minimal Halogen lamps are most popular with new home

(42)

42 Extension of the Theory of Planned Behavior

This study has extended the TPB, adding location and socioeconomic situation as factors influencing purchase behavior through the TPB. RQ2 and RQ3 were the research questions related to these factors. The purpose of these extensions was twofold: allowing a better explanation of behavior for the TPB and providing a better basis for segmentation.

These added variables do provide some additional variance of behavior explained, although the effects aren’t overwhelming. This can be contributed to the fact that the TPB itself does not seem to explain consumer behavior in residential lighting completely (the subsequent section will explore possible reasons for the TPB’s shortcoming further).

The same holds for segmentational benefits. Since many components of the TPB had no significant effect on purchase behavior, socioeconomic variables in relation to these components (RQ3) were of little value.

It is expected that if the TPB was more successful in explaining consumer behavior, the same would hold for the added variables.

Null results

Off all hypothesized relations, few are statistically significant. Three options remain. The first is that measurements where not accurate in some way. This possibility will be discussed in the subsequent sub-chapter “limitations”. Another option is that there are variables that influence the purchase decision which are not taken into account by this research. For example, the armature could dictate the lamp most suitable to fit this armature, either for compatibility reasons or for habitual reasons. It seems as though past behavior can also have effect on purchase behavior. The third reason would be that consumers just do not care much about lamp types. Involvement could just be too low for the suggested variables to affect the purchase decision. This hypothesis is supported by the focus group, as well as the following outcomes:

 Consumers have little knowledge on household lighting

 Criteria are all said to be relatively important (no clear distinguishing hints at low involvement)

 There are little significant relations between the product attributes and the types of lamp purchased, while lamp types differ largely on these attributes.

 Relatively many respondents did not know which type of lamp they had purchased The overall impression talking to people in-depth about the subject, as well as talking to industry experts and the focus group supports the suggestion of low involvement.

6.2 Limitations Sample limitations

Due to the relatively small sample the chance of type II errors is rather large. Since the confidence level (α) is set at 0,05, some hypotheses might not have been rejected while with a larger sample or a lower confidence level, they would. However, these situations are mentioned in the results and most of the time it wasn't even close. The effect of these type II errors is therefore expected to be small. The sample collection method resulted in a sample of the Dutch population with a higher than average education. This could have affected the sample and might jeopardize the validity of the results. The same could be said for a slight over-representation of males in the sample.

(43)

43 constructs were selected properly. The locations where the result of a preliminary test in a very small sample, for which statistical significance could be detected.

Although the questionnaire used questions to test the constructs of environmental attitude and innovativeness which have been validated and widely used, both scales (Domain Specific

Innovativeness and the New Environmental Paradigm) are not without discussion. Also, the translation of these scales into Dutch could have created linguistic subtleties hampering validity.

Finally, due to the length of the questionnaire and the just suggested involvement issue, it is not ruled out that respondents got fed up with the questionnaire and “freewheeled home”.

6.3 Managerial Recommendations

This research provides important insight for (marketing)managers and policy makers in several ways. Although there has been extensive campaigning for cfl's in the past, these campaigns do not appear to have a large effect on the purchase decision. Consumers, on average, largely over-estimate the cost of electricity of lighting. This overestimation does however not lead to lower powered lamps. Therefore, campaigning in this respect does not seem fruitful. This is supported by research in the past. The Ban on Bulbs is perhaps going to prove a wise decision. The proposed ban on Halogen (starting 2016) is recommended, since it seems as halogen is becoming more popular. Other countries could follow the EU in this respect.

It is recommended to managers of lamp producers that they view their customers as having low involvement in the purchase decision. Marketing strategies should be adjusted to this low involvement.

Ensuring armatures are fitted with the managers' desired lamp could for instance prove fruitful. Lobbying or cooperating with armature manufacturers is therefore recommended. Since the EU is banning Halogen lamps starting 2016, it is expected that a lot of armatures will be in need of replacement, providing a great opportunity in the years to come in this respect.

Furthermore, it is recommended that marketing management take locational aspects into account. Specifically, for a living room lamp it is wise to put more effort in the lamp design and the qualitative lighting properties.

6.4 Academic Recommendations and further research

The purpose for this study was to provide for a literature gap in the understanding of the consumer decision making process in household lighting. A start has been made.

One of the contributions of this research was that a factor analysis discovered two sets of attributes, which can be translated into the theory of planned behavior. Further research could perhaps find other sets of attributes.

Using the theory of planned behavior to explain consumer behavior in low involvement products can be of value, although the theory as “as is” might be more useful when expanding the research subject to high involvement products. It is expected that especially subjective norms and perceived behavioral control have a larger influence when the stakes are higher.

The TPB can benefit from additional (external) variables. In this study, location and socioeconomic situation added value to the TPB, increasing its power to explain purchase behavior. For researchers using the TPB, it is highly recommended to add such external variables. Adding socioeconomic variables, past behavior and habitual behavior might be of use in all similar researches.

(44)
(45)

45

7 References

Azevedo, I., Morgan, M., and Morgan, F. (2009). The transition to solid-state lighting. Proceedings of the IEEE, 97(3):481–510.

Ajzen, I. (1985). From intentions to actions: A theory of planned behavior. In J. Kuhl & J. Beckmann (Eds.), Action control: From cognition to behavior. Berlin, Heidelber, New York: Springer-Verlag. Ajzen, I. (1988). Attitudes, Personality, and Behavior, Open University Press, Milton-Keynes Ajzen, I. (1991). The theory of planned behavior. Org. Behav. Hum. Decis. Process. 50, 179–211. Armitage, J., Connor, M. (1998). Extending the Theory of Planned Behavior: A Review and Avenues for Further Research. Journal of Applied Social Psychology 28. 1429-1464.

Bartlett, S. (1993). Shedding light on residential consumers. Energy, 18(2):171–183.

Bentler, P.M., Speckart, G., Models of attitude-behavior relations. Psychology Review, 86 (1979), 452-464

Bertoldi, P. and Atanasiu, B. (2006). Residential lighting consumption and saving potential in the enlarged EU. Proceedings of EEDAL’06 Conference, London, UK, 21-23.06.2006. European Commission, Institute for Energy.

Calwell, C. and Granda, C. (1999). Halogen torchiere market transformation: A look at

progress to date and future strategies. Technical report, Prepared for the Natural Resources Defense Council (NRDC).

Calwell, C. and Mills, E. (1997). Halogen torchieres: A look at market transformation in progress. IAEEL Right Light 4 Conference. Proceedings., pages 41–47.

CEC (1998). Commission Directive 98/11/EC of 27 January 1998 implementing Council Directive 92/75/EEC with regard to energy labelling of household lamps. Official Journal of the European Union, L 71:1–8.

CEC (2005). European compact fluorescent lamps quality charter. Technical report, European Commission.

CEC (2009a). Commission adopts two regulations to progressively remove from the market non-efficient light bulbs. Memo, 18 March 2009.

CEC (2009b). Commission Regulation (EC) No 244/2009 of 18 March 2009 implementing Directive 2005/32/EC of the European Parliament and of the Council with regard

(46)

46 European Union, L076:3–16.

CEC (2009c). Commission Regulation (EC) No 245/2009 of 18 March 2009 implementing Directive 2005/32/EC of the European Parliament and of the Council with regard

to ecodesign requirements for fluorescent lamps without integrated ballast, for high intensity discharge lamps, and for ballasts and luminaires able to operate such lamps, and repealing Directive 2000/55/EC of the European Parliament and of the Council. Official Journal of the European Union, L076:17–44.

CEC (2009d). FAQ: phasing out conventional incandescent bulbs. Memo, 1 September 2009.

Crosbie, T., Stokes, M., and Guy, S. (2008). Illuminating household energy demand and the policies for its reduction. Energy and Environment, 19(7):979–993

Dunlap, R. E., Van Liere, K.D., Mertig, A.G., Jones, R.E. (2000). Measuring endorsement of the new environmental paradigm: A revised NEP scale. Journal of Social Issues, 56, 425-442.

Dunlap, R. E., Frederick H. Buttel, Peter Dickens, and August Gijswijt (eds.) 2002. Sociological Theory and the Environment: Classical Foundations, Contemporary Insights.

Dunlap, Riley E.(2008), 'The NEP Scale: From Marginality to Worldwide Use.' Journal of Environmental Education 40 (1),3-18

Festinger, L. (1957) A theory of cognitive dissonance, Stanford, CA: Stanford University Press

Fouquet, R. & Pearson, P.J.G. (2006). "Seven Centuries of Energy Services: The Price and Use of Light in the United Kingdom (1300-2000), The Energy Journal, vol. 0(Number 1), pages 139-178.

Gatignon, H. and T. S. Robertson (1991). Innovative decision processes. Handbook of Consumer Behavior. T. S. Robertson and H. H.

Gendre, M.F.: Two Centuries of Electric Light Source Innovations. (2003)

Goldsmith, R. E., & Hofacker, C. F. (1991). Measuring consumer innovativeness. Journal of the Academy of Marketing Science, 19(3), 209-221.

Huizingh (2006), SPSS 14.0 voor Windows en Data Entry. Den Haag: Sdu Uitgevers b.v.

Kilbourne, W. & Pickett, G. (2008). How materialism affects environmental beliefs, concern, and environmentally responsible behavior. Journal of Business Research, 61, 885-893

Referenties

GERELATEERDE DOCUMENTEN

Part I: Numerical Method 13 2 SimpleX2: Radiative transfer on an unstructured, dynamic grid 15 2.1

Despite the di fficulties in simulating the evolution of the cosmological density field, the biggest challenge for numerical simulations of cosmic reionisation is presented by

License: Licence agreement concerning inclusion of doctoral thesis in the Institutional Repository of the University of Leiden. Downloaded

Photons that were sent out by a source travel in one radiative transfer time step from grid point to grid point, where an interaction with the medium takes place.. (2.26) is

We can conclude from these tests that a dynamic triangulation algorithm can significantly increase the performance of SimpleX in radiation hydrodynamics if vertices need to be added

Due to the high angular momentum of this galaxy almost all sources are confined in the high density disc, so the additional ionising flux that the low mass sources contribute

De eerste lichtbronnen in het heelal die aldus ontstonden hebben een grote invloed gehad op de vorming van latere structuren, zoals sterren en sterrenstelsels.. De details van

Door het inter-academiaal college over numerieke sterrenkunde werd mijn interesse voor theoretisch werk gewekt, waardoor de keuze om mijn afstudeeronderzoek in de theoriegroep