Understanding behavioral responses to product innovation : how do consumers’ personality traits influence the relationship between analytic thinking and the likelihood of innovation adoption?

(1)

Understanding behavioral responses to product innovation

How do consumers’ personality traits influence the relationship between analytic thinking and

the likelihood of innovation adoption?

Written by Geneviève Lekner

Master Thesis

Geneviève Lekner - 11398442

MSc in Business Administration - Entrepreneurship and Innovation Submission date: 14 June 2018

(2)

Statement of originality

This document is written by Geneviève Lekner who declares to take full responsibility for the contents of this document. I declare that the text and the work presented in this document is original and that no sources other than those mentioned in the text and its references have been used in creating it. The Faculty of Economics and Business is responsible solely for the

(3)

Abstract

It is well established that in today’s rapidly changing environment, innovation drives economic growth, leads to job creation, heightened productivity, and economy expansion. Yet, approximately 90% of today’s product innovations do not survive on the market. It is a critical insight for stakeholders involved in innovations to understand why consumers adopt certain product innovations and resist others. It allows them to be more efficient in their development efforts and helps them identify and improve competitiveness and profitability.

This paper examines whether certain behavioral or personality traits give reason for the barriers that result in innovation resistance. Precisely, this study aims to deliver first empirical evidence on how consumers’ personality traits influence the relationship between analytic thinking and the likelihood of adopting a product innovation. The empirical background of this study includes innovations within the food (production) industry that are meant as a substitute for highly appreciated products such as meat or dairy.

The results demonstrate that analytic thinking is not as strong of a predictor for innovation adoption as initially expected. Besides that, none of the personality traits moderated the relationship between analytic thinking and innovation adoption. Hence, these results show room for further exploration. Overall, this paper aims to contribute to the limited research regarding possible determinants for innovation adoption of innovative food alternatives.

(5)

Introduction

Today, the economy seems to be characterized by continuous innovation resulting in countless new product launches every day (Cornescu and Adam, 2013). According to several studies by Crawford (2008) approximately 90 percent of these new products do not survive on the market (Crawford, 2008). This is conflicting with the results of thorough market research conducted by many organizations (Cornescu and Adam, 2013). Marketers conclude that consumers who “talk the talk” in surveys do not always “walk the walk” with regards to innovation adoption (Arts, Frambach and Bijmolt, 2011). Put differently, consumers intentions to adopt a new product are usually poor predictors of adoption behavior (Frambach, Prabhu and Verhallen, 2003). Nevertheless, the understanding of why consumers adopt certain innovations and resist others is a critical insight for stakeholders involved in innovations (Arts, Frambach and Bijmolt, 2011). Particularly because failure means a lag in future revenue streams and a potential loss of reputation (Heidenreich, Kraemer and Handrich, 2016). Organizations need to understand consumer product resistance and the reasons behind it in order to become more efficient in their development efforts and to identify and improve competitiveness and profitability (Dunphy and Herbig, 1995). Overall, it can be concluded that developing new products which are attractive to consumers is a huge challenge (Hoek et al., 2011). Yet, it is even more complicated as soon as these new products are meant as a substitute for highly appreciated products (Hoek et al., 2011). For that reason, the empirical background of this study concerns the launch of several innovative food products that are healthy and/or environmental friendly alternatives to highly appreciated products such as meat, dairy and sugar. The reason why it was decided to investigate meat alternatives is that consuming animal originating products is nowadays under debate as it is proven to be an environmentally unfriendly food choice: “animal agriculture is responsible for 16 percent of the greenhouse gas emissions which is more than the combined exhaust from all transportation” (WRI, 2008). Additionally, livestock is the number one water polluter worldwide (Wilson, 2014) and in the US responsible for almost 90 percent of the water consumption (Facts and Sources, 2014). Above all, the world population is expected to grow to around 9.7 billion in 2050 (Elferink and Schierhorn, 2016). These figures put a huge pressure on the food production system as the world has to feed more people with the same amount of raw materials. Put differently, the world is in need of alternative protein sources as soon as possible (‘Vegetable proteins’, 2017). Consequently, a lot of organizations and policy makers are concerned with sustainable consumption and production in order to stimulate consumers to make a shift to a more

(6)

sustainable food product (Elferink and Schierhorn, 2016). Similar reasons underlie the choice to study sugar-free alternatives to sweets. Even though there are still many myths about sugar and its effects, one thing is for sure; the average American consumes more than four times what the World Health Organization (WHO) recommends as safe (Lee, 2017). In addition to that, consumers are generally becoming more aware of the risks of too many sweets resulting in many food producers inventing alternatives to these (formerly)

appreciated products (Lee, 2017).

Even though meat substitutes and sugar-free products are thought to be commercially viable, the usual answer to the question whether consumers really want to buy more sustainable products was a profound “sort of”. Put differently, surveys show that consumers aspire to buy responsibly but this purchase intent does not always translate to real sales (Winston, 2015). As businesses often struggle to make sustainable products more mainstream (Winston, 2015), this study aims to concretize the phenomenon of innovation adoption and discover possible determinants related to this consumer reaction.

(7)

Literature review

For decades, innovation literature has been reporting enormous high failure rates with regards to product innovations (Andrew and Sirkin, 2003). According to Heidenreich e.a. (2016) most product innovations get rejected because of consumers’ innovation resistance (Heidenreich, Kraemer and Handrich, 2016). This is in line with Claudy e.a. (2015) who argues that because of this high failure rate, researchers should focus on resistance factors that keep consumers from adopting a new product instead of further analyzing attitudes of innovation adoption (Claudy, Garcia and O’Driscoll, 2015). Studying this phenomenon is extremely relevant for academic research as well as management practice (Talke and Heidenreich, 2014). Yet, so far, research into the phenomenon of consumer resistance to innovation is particularly scarce (Kleijnen, Lee and Wetzels, 2009). On the one side, this gap is surprising as most firms are familiar with market introduction failures since innovations demand change in consumer behavior and resistance to change is a common consumer response (Ellen, Bearden and Sharma, 1991; Kleijnen, Lee and Wetzels, 2009). According to Schneider and Hall (2011), American families repeatedly buy the same 150 items, which is equal to almost 85% of their household needs; it is difficult to get a new innovation on the radar (Schneider and Hall, 2011). Nevertheless, on the other side, innovation literature is largely subject to a pro-change bias; innovation is usually portrayed as being essentially “good” (Cornescu and Adam, 2013). The extant literature follows the assumption that consumers are principally open to change and therefore willing to adopt innovations (Ellen, Bearden and Sharma, 1991; Heidenreich, Kraemer and Handrich, 2016). Researchers have generally a pro-innovation attitude and since the positive consequences of innovation outweigh the negative ones, consumers’ innovation resistance required less research attention (Cornescu and Adam, 2013). In other words, resistance to product innovation is a new concept and so far rather difficult to define (Cornescu and Adam, 2013).

In this study the following definition for product innovation resistance is used: “resistance to innovation represents an unfavorable evaluation of a new product (Nabih, Bloem and Poiesz, 1997) because of the potential changes made in the consumers’ day-to-day existence which disrupts their established routines and habits” (Ram and Sheth, 1989). Put differently, as soon as consumers’ perception of certain product-specific factors do not meet their expectations, innovational barriers may arise which might cause innovation resistance (Laukkanen, Sinkkonen and Laukkanen, 2008). Moreover, innovation adoption is seen as the result of overcoming persistent product resistance (Cornescu and Adam, 2013). Still, consumers’ resistance to innovation is a complex process to investigate due to its subjective nature (Cornescu and Adam,

(8)

2013). Nevertheless, assessing consumers’ resistance to an innovation is viral to an organization in order to create an image close to the real barriers of consumers to the adoption of a new product (Cornescu and Adam, 2013). Analyzing these barriers aids organizations to design better (commercialization) strategies including counteractions which face the potential criticism and the importance factors of resistance (Ellen, Bearden and Sharma, 1991; Cornescu and Adam, 2013). Even though consumers’ resistance to innovation is considerate one of the main reasons behind the high innovations failure rate (Ram and Sheth, 1989), still a large number of studies did not include certain behavioral or personality traits that give reasons for the barriers that results in consumers innovation resistance (Claudy, Garcia and O’Driscoll, 2015). Empirical evidence on the influence of consumers capacity to evaluate a new products value on the likelihood of product innovation, still lacks (Kleijnen, Lee and Wetzels, 2009; Heidenreich, Kraemer and Handrich, 2016). In addition, according to Claudy e.a. (2015) “analyzing direct comparisons and generalizations about the mediating or moderation influence of contextual variables or consumer traits provide fruitful opportunities for future research” (Claudy, Garcia and O’Driscoll, 2015). Hence, the overall aim of this study is to explore whether certain behavioral or personality traits give reason for the barriers that results in consumer innovation resistance (Claudy, Garcia and O’Driscoll, 2015). To be precise, this paper aims to contribute overcoming the pro-change bias by delivering first empirical evidence on how consumers’

personality traits influence the relationship between analytic thinking and the likelihood of

adopting a product innovation that is meant as a substitute for a highly appreciated product. Consumers increasing concern about the environment and their personal health forms the empirical background of this study. The corresponding preference to buy so-called “environmentally friendly products” forces businesses to know their customers’ environmental preferences (Fisher, Bashyal and Bachman, 2012).

Analytic thinking

and corresponding hypothesis

Cornescu and Adam (2013) argued that many studies explored the factors which contribute to “positive” new product decisions (Cornescu and Adam, 2013). Yet, it is also important to understand the barriers consumers experience while incorporating a new product into their consumption pattern (Kleijnen, Lee and Wetzels, 2009). For that reason, this study aims to incorporate both the expected barriers as well as the benefits with regards to a product innovation. This “reasoning” may be a sign of analytic thinking which is defined as the consumers’ capacity to identify the new products utility within consumers’ day-to-day existence

(9)

(Ram and Sheth, 1989). According to Sá e.a. (1999), consumers with a high cognitive ability are more likely to project a relationship, evaluate information and then make a decision based on these findings (Sá, West and Stanovich, 1999). Hence, this sign of analytic thinking relates to intelligence (Sá, West and Stanovich, 1999) and will be measured by testing consumers’ ability to answer several IQ related questions. This shows whether the different levels of analytic thinking influence the likelihood of innovation adoption. Since analytics are generally better able to indicate the connection between a cause and effect (Pennycook, Fugelsang and Koehler, 2015) it is expected that they are more capable to understand what they should purchase in order to safeguard their health and the environment. According to Stanovich e.a. (2011), thinking analytically is seen as the reasoning behind overriding our intuitions and gut feelings; and often our only hope for a better, more rational future (Quintais, 2006; Stanovich, West and Toplak, 2011). For this reason, hypothesis 1 is formulated as follows:

In addition to that, the potential benefits and barriers a consumer expects to experience while adopting a new product were measured to identify, apart from intelligence, the reasoning behind consumers behavior towards a product innovation (Kleijnen, Lee and Wetzels, 2009). The new product innovation barriers are analyzed by determining the functional (what does it do?) as well as the psychological barriers (what does it do for me?) that consumers experience towards a product innovation (Ram and Sheth, 1989; Joachim, Spieth and Heidenreich, 2017). Joachim e.a. (2017) conducted a quantitative large-scale study in order to evaluate the relative importance of each of these innovation barriers. The functional barriers give account to four different areas: (1) product value (e.g. do consumers believe the innovation is superior to competing products) (2) product compatibility (e.g. do consumers perceive that the innovation is incompatible with past or existing products?) (3) product complexity (e.g. are consumers worried that the new product is cumbersome and that they do not know for sure how to prepare it properly?) and (4) product realization (e.g. do consumers believe they have to wait a certain time until the benefits of using

the product evolve?) (Joachim, Spieth and Heidenreich, 2017). Besides that, Joachim e.a., (2017) concluded that the psychological barriers give account to six

different areas: (1) economic risks (e.g. do the consumers believe the product offers a good price-performance ratio?) (2) personal risks (e.g. do consumers think that using the new product H1: “Analytics” are more likely to adopt a product innovation which is meant as a healthier

(10)

might result in negative consequences?) (3) norm barriers (e.g. do consumers believe that the innovation match their values and norms?) (4) social risk barriers (e.g. do consumers believe there is a chance that their friends might respond negatively if they purchase the new product?) (5) information barriers (e.g. do consumers believe that they have all the necessary information about the new product in order to evaluate it before purchase) and/or (6) usage barriers (e.g. does using the product require new behavior/s for example in terms of different preparation skills?) (Joachim, Spieth and Heidenreich, 2017).

The benefits consumers expect to experience while adopting a healthier and/or more environmental friendly alternative to a highly appreciated product are categorized in the most commonly mentioned reasons for taking on green behavior by Claudy e.a., (2015): (1) ecological welfare (e.g. reduction of C02 emissions), (2) animal welfare (e.g. less raising and killing animals for food consumption) and (3) health related benefits (e.g. the higher interest in health, weight control and the natural contents of foods) (Claudy, Garcia and O’Driscoll, 2015). The relative importance of ecological welfare, animal welfare and personal health also control for the effects of being a generally more conscious consumer on the likelihood of innovation adoption.

Personality traits

and corresponding hypothesis

Besides the relationship between consumers’ level of analytic thinking and the likelihood of innovation adoption (H1), this study explores the possible influence of personality traits on this relationship (figure 1). As analytics are better able to indicate the connection between cause and effect (Pennycook, Fugelsang and Koehler, 2015), they are more likely to understand what to purchase in order to safeguard their health and the environment. Yet, analytics’ personality might influence this relationship to innovation adoption (Goldsmith, 1984). Based on the conceptual model of Claudy et al., (2015) consumers’ personality traits are accommodated in three separate consumer-specific factors which have a moderating function in this study: (1) taste for novelty (Im, Mason and Houston, 2007), (2) taste for variety (Mcalister and Pessemier, 1982) and (3) general resistance to change (Oreg, 2003).

Taste for novelty- the personality trait “taste for novelty” most likely relates to new product

adoption (Im, Mason and Houston, 2007). Yet, empirical evidence shows uneven results consisting of positive, negative, very weak or no relationship at all (Im, Mason and Houston, 2007). According to Im e.a. (2007) these weak or inconsistent results may arise because the conceptual models have not taken valuable intervening variables into account (Im, Mason and Houston, 2007). In this study we use the definition of Kirton (1976) and Im e.a., (2007) and

(11)

define taste for novelty as: “a generalized unobservable trait that reflects a person’s inherently innovative personality, predisposition, and cognitive style”. Put differently, this personality craves to be surrounded by the latest innovations. An example might be a vegetarian consumer purchasing a certain meat substitute until a new alternative is launched which has possibly better sensory characteristics or contains fewer additives. Hence, as soon as an analytic shows to have a certain taste for novelty, the effect between analytic thinking and innovation adoption is expected to be stronger. Especially because of the fact that even though analytics are generally better able to understand what they should purchase in order to safeguard their health and the environment (Pennycook, Fugelsang and Koehler, 2015), they can be quite skeptical when it comes down to trying new things (Norton, 2011). Though, it is expected that a consumers’ inherently innovative personality will counteract this form of being skeptical. For that reason, the hypothesis 2 is formulated as follows:

Taste for variety- Taste for variety differs from taste for novelty as it includes switching among

product variants regardless whether these products are new or existing innovations (Mcalister and Pessemier, 1982). An example might be a vegetarian consumer regularly switching among existing and new meat alternatives pure for the sake of experiencing variety. Taste for variety is expected to modify the relationship between analytic thinking and innovation adoption because it leads individuals to engage in varied behaviors (Mcalister and Pessemier, 1982). For that reason, hypothesis 3 is formulated as follows:

General resistance to change- Todays modern societies value consumers who are willing and

able to positively respond to change (Oreg, 2003). Nevertheless, organizations that initiate changes and launch innovative new products are often hindered by consumers who resist the changes (Oreg, 2003). This resistance to change is often caused by consumers who believe the changes are not consonant with their personal interests (Oreg, 2003). Nevertheless, some H2: Consumers’ “taste for novelty” will positively influence the relationship between analytic

thinking and the likelihood of adopting a product innovation.

H3: Consumers’ “taste for variety” will positively influence the relationship between analytic

(12)

individuals even resist changes that are consistent with their personal interests (Oreg, 2003). As a result, this “general resistance to change” exerts a weaker effect on innovation adoption. For that reason, hypothesis 4 is formulated as follows:

Please refer to figure 1 for the conceptual model based upon above mentioned hypothesis.

Figure 1: The conceptual framework

H4: Consumers’ “general resistance to change” will negatively influence the relationship

(13)

Method

Design

This quantitative study will be set up through a cross-sectional survey design in order to best generalize the findings at a specific point in time. It aims to identify the relationship between analytic thinking and the likelihood of innovation adoption and analyzes three types of personality traits to see whether these moderated the relationship. The survey will be generated through Qualtrics in which participants will be asked to participate online. The design consists of two phases: one phase to identify the relationship between analytic thinking and the likelihood of innovation adoption. The second phase aims to identify the influence of three separate personality traits as moderators on the relationship between analytic thinking and innovation adoption.

Participants and Procedure

In this study, there is no need to target a specific population as it strives to reach a generalizable image with regards to the adoption of a new product and its corresponding variables. Approximately half of the participants (+/- 100 respondents) were reached through Facebook and LinkedIn and the other half (+/- 100 respondents) through Amazon Mechanical Turk (Mturk). Mturk is a marketplace for work that requires human intelligence (Mturk, 2018). Mturk enables requesters to programmatically access this marketplace and a diverse, on-demand workforce. A pre-test, conducted among 20 participants, helped to make sure there were no problems with regards to the survey that might result in biased answers. It clarified whether the survey design was realistic and made sense to the respondents.

Measures

Analytic thinking (IQ) (α= .298). The sign of analytic thinking will be measured by testing

consumers’ ability to answer several IQ related questions. These logical and mathematical questions are based on the most scientifically valid IQ test available online today (‘IQTest’, 2017). An example item is: “this sequence of four words “triangle, glove, clock, bicycle” corresponds to this sequence of numbers “3, 5, 12, 2”. Respondents were asked to answer by either “true” or “false”. Please refer to table 1 for the full list of items.

Lastly, the most common measure of scale reliability, Cronbach’s Alpha (α), is below the lower limit of .70 for scales and .30 for separate items (table 7) (Field, 2013). This might be due to poor inter-relatedness between the items or heterogeneous constructs measuring IQ (see

(14)

limitations chapter).

Table 1: IQ test

Q1 _{This sequence of four words “triangle, glove, clock, bicycle” corresponds to this sequence of} numbers “3, 5, 12, 2”

Q2 A round wall clock that has been rotated until it is hanging upside down will have a minute hand that points to your right when it is two forty-five

Q3 This sentence has thirty-five letters

Q4 The sum of all the odd numbers from zero to 16 is an even number Q5 _{The letters of the word, “sponged” appear in reverse alphabetical order}

Q6 _{The number 64 is the next logical number in the following sequence of numbers: “2,6,14,30…”}

Innovation Barriers (α= .721). Both barrier types are measured based on existing scales within

the AIR typology using a 5 point Likert scale (1= strongly disagree and 5= strongly agree) (Claudy, Garcia and O’Driscoll, 2015). This resulted in four separate functional barriers types and six psychological barrier types (Kulviwat et al., 2007). An example item of functional barriers is “It should not take a long time before the benefits of using the product occur” (Talke and Heidenreich, 2014). An example item for the psychological barriers is “The new product should offer a good price-performance ratio”. Please refer to table 2 and 3 for the full list of items. Lastly, the Cronbach’s alpha (α) of both the scale (>.70) and separate items (>.30) fall within the acceptable values. This means that the measure is consistently reflecting the construct that it is measuring (table 7) (Field, 2013).

Table 2: Functional barriers

Value barriers “This product should offer advantages not offered by

competing products” (Claudy, Garcia and O’Driscoll, 2015)

Compatibility barriers “It should be possible to use the new product in

combination with products I already use” (Talke and Heidenreich, 2014)

Complexity barriers “The new product should be easy to use/prepare” (Kleijnen, Lee and

Wetzels, 2009)

Realization barriers “It should not take a long time before the benefits of

(15)

Table 3: Psychological barriers

Norm barriers “The new product should match my values and

norms” (Laukkanen, 2016)

Economic risk barriers “The new product should offer a good price-

performance ratio” (Kleijnen, Lee and Wetzels, 2009)

Social risk barriers “Overall, my friends should respond positively if

I purchase the product” (Kleijnen, Lee and Wetzels, 2009)

Usage barriers “The use of the new product is completely

compatible with my needs” (Rudolph, Rosenbloom and Wagner, 2004)

Information barriers “I am well informed about the new product” (Talke and Heidenreich, 2014)

Personal risk barrier “Using the product is no threat to my physical

condition” (Klerck and Sweeney, 2007)

Innovation Benefits (α= .717). Participants were asked how likely they believe each type of

benefit (ecological welfare, animal welfare and personal health) to be a motivation or reason for them to adopt a product innovation. Respondents were asked to rank each motivation on a scale from 1 to 10 reflecting their importance. The Cronbach’s Alpha (α) of both the scale (>.70) and separate items (>.30) fall within the acceptable values (table 7).

Personality traits – Taste for novelty (α= .729). This construct is measured using the ‘Kirton

Adaption Innovation (KAI)” inventory (Kirton, 1976). Other scales have been considered, yet the KAI is selected as is has been extensively tested for generalizability, content validity and reliability in several context (Goldsmith, 1984). Additionally, research shows that: “it is highly correlated with other innovativeness scales in the context of consumer adoptions of innovations” (Goldsmith, 1984). The “originality” sub-dimension (others capture conformity and efficiency) of the KAI inventory following Keller et al., (2002) and Goldsmith (1984) is used. It measures personality traits by asking respondents, on a 5 point Likert scale, the extent to which they agree on certain statements. The eleven items of the originality sub dimension were, due to the length of the survey, shortened to the first 5 items of the list (table 4). An example item is “I often risk doing things differently” (Keller and Holland, 1978) (Goldsmith, 1984). Refer to table 4 for the full list of items.

Lastly, the Cronbach’s Alpha (α) of both the scale (>.70) and separate items (>.30) fall within the acceptable values. This means that the measure is consistently reflecting the construct that it is measuring (table 7) (Field, 2013).

(16)

Table 4: Taste for novelty Q1

I often risk doing things differently

Q2 I often come up with original ideas

Q3 I cope with several ideas at the same time

Q4 I need the stimulation of frequent change

Q5 I will always think of something when stuck

Q6 has fresh perspectives on old problems

Q7 is stimulating

Q8 can stand out in disagreement against a group

Q9 would sooner create than improve

Q10 likes to vary set routines at a moment’s notice

Q11 proliferates ideas

Personality traits – Taste for variety (α=.759). Consumers’ taste for variety was measured

through the consumer-specific Exploratory Acquisition of Products (EAP) scale to measure “a consumer’s tendency to seek sensory stimulation on product purchase through risky and innovative product choices and varied and changing purchase or consumption experiences” (Trijp, Hoyer and Inman, 1996). The ten EAP items are scored on 5 point Likert scale (1= strongly disagree and 5= strongly agree) (Baumgartner and Steenkamp, 1996). Yet, the ten EAP items were, due to the length of the survey, shortened to the first 5 items of the list. An example item is: “even though certain food products are available in a number of different flavors, I tend to buy the same flavor” (Baumgartner and Steenkamp, 1996). Nevertheless, items 1, 2, 3 and 5 were negatively-keyed items which could result in acquiescence bias (Field, 2013). Therefore, these items were reverse-coded before analyzing.

Lastly, the Cronbach’s alpha (α) of both the scale (>.70) and separate items (>.30) fall within the acceptable values (table 7).

(17)

Table 5: Taste for variety

Q1 _{“Even though certain food products are available in a number of different flavors, I tend to buy}

the same flavor”.

Q2 “I would rather stick with a brand I usually buy than try something I am not very sure of” Q3 “I think of myself as a brand-loyal consumer”

Q4 “When I see a new brand on the shelf, I'm not afraid of giving it a try”

Q5 “When I go to a restaurant, I feel it is safer to order dishes I am familiar with” Q6 “If I like a brand, I rarely switch from it just to try something different” Q7 “I am very cautious in trying new or different products”

Q8 “I enjoy taking chances in buying unfamiliar brands just to get some variety in my purchases” Q9 “I rarely buy brands about which I am uncertain about how they will perform” Q10 “I usually eat the same kinds of foods on a regular basis”

Personality traits – General resistance to change (GRTC) (α=0.762). The resistance to change

scale was designed following Oreg (2003) who through an exploratory analysis indicated four reliable factors: (1) Routine Seeking, (2) Emotional Reaction to Imposed Change, (3) Cognitive Rigidity and (4) Short-Term Focus (Oreg, 2003). The scale can be used to predict reactions to change (Oreg, 2003). An example item of the Routine Seeking factor is: “I generally consider changes to be a negative thing”. An example item of the Emotional Reaction to Imposed Change factor is: “When things don’t go according to plans, it stresses me out”. An item that measures the Short-term focus is: “I sometimes find myself avoiding changes that I know will be good for me”. Cognitive Rigidity can be measured through for example the following item: “I often change my mind”. However, this is a counter-indicative item which might result in acquiescence bias (Field, 2013). For that reason, “cognitive rigidity” was reversed-coded before analyzing. All items were measured according to a five point Likert scale ranging from 1= strongly disagree to 5= strongly agree. Yet, due to the limited length of the survey, solely the first two items of each factor were measured. Refer to table 6 for the full list of items.

Lastly, the Cronbach’s alpha (α) of both the scale (>.70) and separate items (>.30) fall within the acceptable values, meaning that the measure is consistently reflecting the construct that it is measuring (table 7) (Field, 2013).

(18)

Table 6: General resistance to change Routine seeking

Q1 _{I generally consider changes to be a negative thing}

Q2 I’ll take a routine day over a day full of unexpected events any time Q3 I like to do the same old things rather than try new and different ones.

Q4 Whenever my life forms a stable routine, I look for ways to change it

Q5 I’d rather be bored than surprised.

Emotional reaction

Q6 _{If I were to be informed that there’s going to be a significant change regarding the}

way things are done at work, I would probably feel stressed.

Q7 When things don’t go according to plans, it stresses me out. Q8 When I am informed of a change of plans, I tense up a bit.

Q9 If my boss changed the criteria for evaluating employees, it would probably make me feel uncomfortable even if I thought I’d do just as well without having to do any extra work.

Short-term thinking

Q10 Once I’ve made plans, I’m not likely to change them.

Q11 I sometimes find myself avoiding changes that I know will be good for me.

Q12 Often, I feel a bit uncomfortable even about changes that may potentially improve my life.

Q13 Changing plans seems like a real hassle to me

Q14 When someone pressures me to change something, I tend to resist it even if I think the change may ultimately benefit me.

Cognitive Rigidity

Q15 I often change my mind.

Q16 Once I’ve come to a conclusion, I’m not likely to change my mind. Q17 I don’t change my mind easily.

(19)

Innovation adoption (IA) (α=.647). Participants were asked how likely it is that they would

purchase a certain product innovation which is meant as a substitute for a highly appreciated product. In this study was chosen for the following three “substitutes”: (1) a plant- based steak, (2) a non-diary ice cream and (3) a sugar-free version of respondents’ favorite cake. The likelihood of innovation adoption was measured on a 7 point Likert scale (1=very unlikely, 7=very likely). Nevertheless, the Cronbach’s alpha (α) of the scale (<.70) is questionable (table 7). This might be the case due to the low number of questions measuring this construct (3). For more information on this matter please refer to the limitations chapter. The Cronbach’s alpha of the separate items (>.30) do fall within the acceptable values (Field, 2013).

Control variables. Lastly, factors such as age, gender, education and income have an impact on

the adoption of a new product (Roberts, 1996; Claudy, Garcia and O’Driscoll, 2015). Hence, these were included as control variables. Since healthier and more environmental friendly products are generally more expensive, respondents with a relatively high income might adopt such products easier compared to participants having a relatively low income (Claudy, Garcia and O’Driscoll, 2015). Besides that, gender is included as a control variable since women appear to be more willing to pay extra for environmentally friendly foods compared to men (Loureiro, Mccluskey and Mittelhammer, 2002). In addition to that, age is included as 40% of worlds vegan population is between 25 and 34 years old (Veganbits, 2018), which increases the likelihood of this age group adopting the new product innovations measured within this model. Lastly, education was included as consumers with a relatively high level of education are more likely to respond to products that are environmentally friendly (Chan, 2008). While answering each of these “control questions” the option “I prefer not to share” was offered. While analyzing the collected data this option was categorized as “missing values”.

(20)

Table 7: Descriptive Statistics, Reliability

Measure Items α M SD

Innovation adoption 3 .647 3.265 1.026

Analytic thinking 6 .298 1.234 .199

Taste for novelty 5 .729 3.568 .677

Taste for variety 5 .759 2.966 .751

General resistance to change 6 .762 3.039 .672

Barriers 10 .721 7.132 1.200

(21)

Results

There were 341 respondents that participated in the final survey. Out of these respondents 48 respondents answered less than half of the survey; therefore they were removed from the sample. Besides that, 74 respondents finished the survey in 220 seconds or less and as the average response time was 440 seconds, it was decided to remove these respondents from the sample. This resulted in a total of 219 valid respondents. Moreover, 41.2% of the sample was female and 58.8% was male. Furthermore, 71.2% of the sample was between 18 And 34 years old, meaning that the millennial generation was highly represented. 56.6% of the respondents were collected through Amazon Mechanical Turk (Mturk) and 43.4% via Facebook and LinkedIn. In order to make sure there are no significant difference between the data of both collection methods, the analysis is replicated on split samples (0=Facebook/Linked, 1=Mturk). See table 8 to 10 for the demographics related to sample 0 and table 11 to 13 for the demographics related to sample 1.

Table 8: Age sample 0

Frequency Percent (%) Cumulative Percent (%)

Age <18 1 1,1 1,1 18-24 36 38,7 39,8 25-34 44 47,3 87,1 35-44 4 4,3 91,4 45-54 4 4,3 95,7 55-64 4 4,3 100,0

Table 9: Gender sample 0

Frequency Percent Cumulative Percent

Gender Male 49 52,7 53,3

Female 43 46,2 100,0

Table 10: Education sample 0

Frequency Percent (%) Cumulative Percent (%)

Education No schooling completed 2 2,2 2,2

High school graduate 20 21,5 23,9

Bachelor’s degree 42 45,2 69,6

Master’s degree 27 29,0 98,9

(22)

Table 11: Age sample 1

Age <18 13 10,2 10,3 18-24 64 50,0 61,1 25-34 30 23,4 84,9 35-44 10 7,8 92,9 45-54 8 6,3 99,2 55-64 1 ,8 100,0

Table 12: Gender sample 1

Gender Male 80 62,5 63,5

Female 46 35,9 100,0

Table 13: Education sample 1

Education High school graduate 32 25,0 25,6

Bachelor’s degree 74 57,8 84,8

Master’s degree 15 11,7 96,8

(23)

A correlation matrix

Before determining the relationship between all variables, the data was checked for normal distribution. The Shapiro-Wilk and the Kolmogorov-Smirnov normality test rejected the normal distribution hypothesis, meaning the data is significantly different from normal (Field, 2013). The data was then analyzed visually through histograms as well as Q-Q plots in which a clear pattern was observed for every variable within the data set. After that, all outliers were analyzed using the 3 multiplier method by (Field, 2013). This resulted in 6 outliers that fell outside this range; consequently these values were removed from the data set.

As a next step a correlation matrix was compiled for the dependent variable, independent variable and control variables (table 14 and 15) (Field, 2013). As can be seen in the Pearson’s correlations, several items are significantly correlated (** and *). For example, the personality trait “general resistance to change (GRTC)” is significantly and negatively correlated to the personality trait “taste for variety (TFV)”. In other words, the more consumers are resistant to change, the less their taste for variety. Secondly, there is a significant and positive relationship between innovation adoption (IA) and the personality trait “taste for novelty (TFN)”. This means that consumers with a taste for novelty are more likely to adopt a product innovation. Thirdly, innovation adoption is significantly and positively related to the benefits (with the exception of personal health) related to a new product. Put differently, the higher consumers expect to experience improved animal- or ecological welfare related to a new product, the more likely they are to adopt such innovations. Additionally, there is a significant and positive relationship between education and innovation adoption. This means that consumers with a higher level of education are more likely to adopt a product innovation. Lastly, there is a significant and negative relationship between analytic thinking (IQ) and the personality trait “general resistance to change”. This means that consumers with a relatively higher IQ are less likely to be generally resistant to change with regards to product innovation.

(24)

Table 14: Correlations 1 2 3 4 5 6 7 1.IA 1 2.IQ -.075 1 3.GRTC -,048 -.173* 1 4.TFV ,007 .090 -,523** 1 5.TFN ,307** .075 -,196** ,124 1 6.Benefits ,322** .029 -,013 ,026 ,315** 1 7.Barriers -.013 -.128 ,000 -,041 ,209** ,376** 1

**. Correlation is significant at the 0.01 level (2-tailed) *. Correlation is significant at the 0.05 level (2-tailed) Innovation Adoption (IA), Analytic thinking (IQ), General resistance to change (GRTC), Taste for novelty (TFN), Taste for variety (TFV)

Table 15: Correlations incl. specified barriers, benefits and control variables

**. Correlation is significant at the 0.01 level (2-tailed) *. Correlation is significant at the 0.05 level (2-tailed). Innovation Adoption (IA), Analytic thinking (IQ), General resistance to change (GRTC), Taste for novelty (TFN), Taste for variety (TFV), Functional barriers (FB), Psychological barriers (PB), Animal welfare (animal), Ecological welfare (Eco), Personal health (health)

1 2 3 4 5 6 7 8 9 10 11 12 13 14 1.IA 1 2.IQ -.075 1 3.GRTC -,048 .173* 1 4.TFV ,007 -,090 -.523** 1 5.TFN ,307** .075 -.196** ,124 1 6.FB -,050 -.143* -.053 -,011 ,182* 1 7.PB ,025 -.088 .053 -,061 ,218** ,564** 1 8.Animal ,310** -,081 .010 -,023 ,302** ,150* ,236** 1 9.Eco ,312** -,053 -.080 ,185* ,292** ,204** ,263** ,678** 1 10.Health ,119 ,103 .045 -,118 ,155* ,353** ,467** ,326** ,360** 1 11.Age -,196** ,042 .048 ,038 -,154* ,133* ,114 ,005 -,022 ,023 1 12.Gender ,137* .035 .109 -,054 ,033 ,149* ,026 ,144** ,104 ,119 -,087 1 13.Income -,089 -.053 -.061 ,011 -,045 ,112 ,048 ,067 ,006 ,045 ,324** -,084 1 14.education .271** -.015 -,028 -,053 ,134* -,038 -,105 -,044 ,012 -,006 -,108 ,166* ,119 1

(25)

Split sample correlation matrix

In order to make sure there are no significant difference between the data of both collection methods, the correlation matrix is replicated on split samples (table 16 and 17) (0=Facebook/Linked, 1=Mturk). The results demonstrate several differences between the split samples and the complete sample. First of all, solely sample 0 shows a significant negative relationship between analytic thinking (IQ) and innovation adoption (IA), meaning that the more consumers are able to think analytically, the less likely they are to adopt a product innovation. Secondly, the relationship between gender and innovation adoption shows to be much larger in sample 0. The same goes for the relationship between ecological welfare and innovation adoption, meaning that sample 0 values ecological welfare over animal welfare and personal health. Besides that, the relationship between age and innovation adoption is not significant in sample 0. Lastly, sample 1 demonstrates a negative significant relationship between taste for variety (TFV) and analytic thinking, meaning that the higher consumers’ “taste for variety” the lower their level of analytic thinking. Yet, sample 1 does not show additional significant dissimilarities suggesting that it is representative of the overall sample.

Table 16: Correlations Sample 0

1 2 3 4 5 6 7 8 9 10 11 12 13 14 1.IA 1 2.IQ -,281** 1 3.GRTC ,108 ,142 1 4.TFV -,051 ,150 -,371** 1 5.TFN ,033 ,015 -,209* ,091 1 6.FB ,026 -,057 -,049 ,096 ,220* 1 7.PB ,077 -,107 ,021 ,029 ,202 ,557** 1 8.Animal ,375** -,073 ,067 -,145 -,075 ,058 ,321** 1 9.Eco ,502** ,003 ,050 ,114 -,039 ,054 ,167 ,647** 1 10.health ,124 ,166 -,001 -,076 ,045 ,183 ,302** ,444** ,378** 1 11.age -,079 ,111 -,089 ,079 -,128 ,130 ,057 ,057 ,117 ,060 1 12.gender ,378** -,049 ,096 ,017 -,112 -,016 ,042 ,195 ,147 ,100 -,215* 1 13.income -,122 ,066 -,097 -,033 ,061 ,076 -,035 ,087 -,023 ,081 ,462** -,267* 1 14.education ,234* -,075 ,109 -,093 ,220* ,031 -,098 -,074 ,061 ,047 -,017 ,337** -,052 1

(26)

Table 17: Correlations sample 1 1 2 3 4 5 6 7 8 9 10 11 12 13 14 1.IA 1 2.IQ ,085 1 3.GRTC -,057 ,150 1 4.TFV -,013 -,219* -,594** 1 5.TFN ,414** ,131 -,121 ,095 1 6.FB -,117 -,181* -,008 -,101 ,155 1 7.PB ,011 -,091 ,021 -,087 ,210* ,606** 1 8.Animal ,272** -,085 -,019 ,048 ,451** ,183* ,186* 1 9.Eco ,195* -,087 -,169 ,233** ,433** ,302** ,346** ,686** 1 10.health ,124 ,061 ,070 -,144 ,219* ,486** ,568** ,253** ,343** 1 11.age -,203* -,054 -,065 ,133 -,103 ,189* ,119 -,029 -,095 -,013 1 12.gender -,065 ,112 ,210* -,150 ,070 ,231** ,091 ,131 ,092 ,144 ,052 1 13.income -,030 -,162 -,108 ,085 -,068 ,115 ,008 ,039 ,015 -,014 ,197* ,067 1 14.education ,287** ,044 -,090 -,058 ,074 -,095 -,102 -,026 -,027 -,044 -,145 ,007 ,279** 1

(27)

Independent sample t-test and ANOVA

The differences between sample 0 and 1 in terms of age, gender, education and income are tested statistically using an independent sample t-test (in the case of gender) and ANOVA (in the case of education, income and age). These results characterize both samples demographically which might help to clarify future outcomes.

First of all, the results in sample 0 demonstrate a significant difference between gender and innovation adoption F(88, 83.791) = .187, p < .05. Specifically, females are more likely to adopt a product innovation that is meant as a substitute for a highly appreciated product.

Secondly, the results of both samples show a statistically significant effect of education levels on innovation adoption F(5,85) = 2.467, p<.05 (sample 0), F(3, 114) = 5.670, p <.05 (sample 1). The relatively higher educated consumers appear to be more likely to adopt a product innovation. Besides that, the results of both samples show a statistically significant effect of age on innovation adoption F (5, 85) =2,467, p<.05 (sample 0), F (5,113) =2.478, p<.05 (sample 1). Lastly, the results of both samples demonstrate that income does not have a significant effect on innovation adoption.

(28)

Regression analyses

A linear regression analysis was performed to investigate the ability of analytic thinking to predict innovation adoption, after controlling for age, gender, education and income (table 18). In the first step of the hierarchical multiple regression, four predictors were entered: gender, age, education and income. This model was statistically significant F (4, 196) =6.726; P<.05 and explained 12.1% of variance in innovation adoption. The results of this model showed that only

education predicts innovation adoption significantly well (β=.267, p<0.05).

After entry of analytic thinking (IQ) at step 2 the total variance explained by the model as a whole was 12.4% F (1, 195) = .810; P>0.05. Put differently, the introduction of IQ to the model only explained an additional of 0.3% variance in innovation adoption, after controlling for age, gender, education and income (R2 change=.004; F(5.538)=.810; P>0.05). This result demonstrates that there is more than a 5% chance that this F-ratio would happen if the null hypothesis were true (Field, 2013). It can be concluded that the regression model does not predict innovation adoption significantly well. This means that H1 is rejected.

In step 3, the barriers and benefits consumers expect to experience while adopting a product innovation were added to the multiple regression analysis. These independent variables specify the reasoning behind consumers’ choice regarding innovation adoption. Results show that the F- value is solely significant for the benefits related to a product innovation (R2 change=.105; F (8.203) =13.141; P<0.05). In order to see which of the benefits contributed most to predicting the dependent variable, the same analysis was conducted for each type of benefit (animal welfare, ecological welfare and personal health). The results demonstrate that solely animal welfare predicts innovation adoption significantly well. Put differently, consumers who value the improved animal welfare related to the consumption of a new product most (in comparison to ecological welfare and personal health) are more likely to adopt a product innovation. Besides that, it can be concluded that the functional as well as the psychological barriers do not predict innovation adoption significantly well.

In short, the improved animal welfare related to the consumption of a product innovation and consumers’ level of education seemed to be statistically significant in predicting innovation adoption.

(29)

Table 18: Hierarchical regression model of innovation adoption R R² R² Change B SE β t Sig. Step 1 .347 .121* .121 Age -.125 .068 -.133 -1.833 .068 Gender .121 .141 .059 .853 .395 Education .360* .094 .267 3.827 .000 Income -.035 .028 -.090 -1.237 .218 Step 2 .353 .124 .004 Age -.121 .068 -.128 -1.767 .079 Gender .124 .141 .060 .880 .380 Education .360* .094 .267 3.823 .000 Income -.037 .028 -.094 -1.292 .198 IQ .309 .343 .061 .900 .369 Step 3 .479 .229* 0.105 Age -.105 .065 -.111 -1.613 .108 Gender .041 .136 .020 .305 .761 Education .369* .089 .273 4.127 .000 Income -.043 .027 -.111 -1.613 .108 IQ 0.334 .328 .065 1.016 .311 Barriers -.112 .060 -.130 -1.879 .062 Benefits .197* .038 .346 5.119 .000 Animal Welfare .087* .033 .232 2.619 .010 Ecological .064 .040 .139 1.583 .115 Personal health .003 .044 .005 .064 .949

* Correlation is significant at the 0.05 level

Split sample Regression analysis

In order to make sure there are no significant difference between the data of both collection methods, the linear regression analysis (OLS) is replicated on split samples (0=Facebook/Linked, 1=Mturk).

(30)

Sample 0

The results of the hierarchical multiple regression analysis for sample 0 demonstrate that apart from model 1 and 3, also model 2 is statistically significant F(1, 77) = 4.610; p<0.05 and explained 23% of variance in innovation adoption. Since model 1 explained 16.4% of the variance, the introduction of analytic thinking (IQ) in model 2 explained additional 6.7% variance in innovation adoption, after controlling for age, gender, income and education. Specifically, for 1 standard deviation increase in analytic thinking, the likelihood of innovation adoption decreases of .261 standard deviations. Therefore, it can be concluded that H1 is also

rejected in this sample. In addition to that, another main difference between the OLS analysis of

the complete sample and sample 0 is that gender instead of education predicts innovation adoption significantly well F (4, 78) =3.818; P<0.05). Lastly, in this sample ecological welfare instead of animal welfare predicts innovation adoption significantly well (table 19).

Sample 1

As a next step, a hierarchical multiple regression analysis was conducted for sample 1. The results demonstrate no significant differences compared to the original analysis including the complete sample, meaning the Mturk sample is representative of the original sample (table 20).

Additionally, a probit/logit analysis was conducted to check for any indifference. Yet, this analysis demonstrated no significant dissimilarities.

(31)

Table 19: OLS analysis of innovation adoption – sample 0 R R² R² Change B SE β t Sig. Step 1 .405 .164 .164 Age .048 .122 .046 .393 .695 Gender .554* .221 .287 2.506 .014 Education .194 .133 .162 1.454 .150 Income -.041 .041 -.119 -.994 .324 Step 2 .480 .230 .067 Age .081 .119 .079 .685 .495 Gender .540* .213 .279 2.531 .013 Education .179 .129 .149 1,390 .169 Income -.040 .039 -.116 -1.009 .316 IQ -1.326* .513 -.261 .2,583 .012 Step 3 .653 .426 .196 Age -.024 .108 -.023 -.224 .824 Gender .410* .195 .212 2,099 .039 Education .196 .118 .164 1.667 .100 Income -.019 .037 -.056 -.516 .608 IQ -1.177* .481 -.232 -2.445 .017 Barriers -.059 .091 -.060 -.645 .521 Benefits .208* .055 .362 3.762 .000 Animal Welfare .052 .055 .132 .955 .343 Ecological .182* .059 .389 3.070 .003 Personal health -.078 .062 -.132 -1.262 .211

(32)

Table 20: OLS analysis of innovation adoption – sample 1 R R² R² Change B SE β t Sig. Step 1 .347 .121* .121 Age -.143 .089 -.148 -1.613 .110 Gender -.120 .188 -.056 -.638 .525 Education .421* .136 .291 3.088 .003 Income -.035 .041 -.082 -.856 .394 Step 2 .353 .124 .004 Age -.143 .089 -.149 -1.613 .110 Gender -.138 .190 -.065 -.728 .468 Education .411* .137 .285 3.003 .003 Income -.029 .041 -.068 -.696 .488 IQ .360 .455 .072 .790 .431 Step 3 .479 .229* 0.105 Age -.113 .088 -.118 -1.292 .199 Gender -.248 .188 -.116 -1.320 .190 Education .434* .132 .301 3.300 .001 Income -.035 .040 -.081 -.876 .383 IQ .442 .452 .088 .977 .331 Barriers -.065 .092 -.082 -.713 .477 Benefits .189* .052 .343 3.643 .000 Animal Welfare .106* .042 .297 2.537 .013 Ecological .001 .054 .003 .024 .981 Personal health .071 .061 .127 1.173 .243

(33)

PROCESS model for moderator

The PROCESS tool by Andrew Hayes and his colleague Kristopher Preacher was used to test the entire model including moderators (Field, 2013). PROCESS is an observed variable OLS regression path analysis modeling tool and used to test how much the effect of analytic thinking on innovation adoption is different between personality traits (figure 2) (processmacro, 2018). Again, in order to make sure there are no significant difference between the data of both collection methods, the PROCESS model for moderator is replicated on split samples (0=Facebook/Linked, 1=Mturk).

Figure 2: PROCESS model for moderator

The regression analysis demonstrated that in sample 0, gender appeared to be the only significant control variable for innovation adoption. Hence, only gender was included as a control variable. As can be seen in table 21, the moderation effect is not taking place for any of the personality traits in sample 0. This means that the relationship between analytic thinking and innovation adoption is not moderated by “taste for novelty”, “taste for variety”, or “general resistance to change”. Based on this information H2, H3 and H4 are rejected in this sample (table 23).

Besides that, the regression analysis demonstrated that in sample 1, education appeared to be the only significant control variable for innovation adoption. For that reason, only education was included as a control variable in sample 1. As can be seen in table 22, the moderation effect is not taking place for any of the personality traits in this sample either. Based on this information H2, H3 and H4 are also rejected in this sample (table 23).

(34)

Table 21: Unstandardized OLS regression coefficients with confidence intervals to estimate innovation adoption with personality traits as moderator and gender as a control variable

Innovation Adoption Coeff. SE t p

IQ (X) -1.3701 .6084 -2.2518 .0269

Taste for Novelty (W) .1788 .2909 .6145 .5405

Gender .7143 .2153 3.3178 .0013

X*W -.6265 1.6372 -.3826 .7029

IQ (X) -1.3310 .5555 -2.3962 .0188

Taste for Variety (Z) -.0300 .1718 -.1748 .8616

Gender .7129 .1983 3.5908 .0006

X*Z -.3489 .7745 -.4505 .6535

IQ (X) -1.3016 .5622 -2.3152 .0230

General resistance to Change (U) .2136 .1799 1.1876 .2383

Gender .6994 .1864 3.7528 .0003

X*U -1.7950 .9671 -1.8560 .0669

(35)

Table 22: Unstandardized OLS Regression Coefficients with Confidence Intervals to estimate Innovation Adoption with personality traits as Moderator and Education as a Control Variable

Innovation Adoption Coeff. SE t p

IQ (X) -.3446 1.8120 -.1902 .8495

Taste for Novelty (W) .3484 .6373 .5467 .5857

Education .3786 .1189 3.1841 .0019

X*W .1379 .4935 .2793 .7805

IQ (X) -1.7321 1.7282 -1.0022 .3184

Taste for Variety (Z) -.9053 .7174 -1.2620 .2095

Education .3986 .1292 3.0851 .0026

X*Z .7348 .5715 1.2858 .2011

IQ (X) 3.8080 2.4103 1.5799 .1169

General resistance to Change (U) 1.2585 .9262 1.3589 .1769

Education .3975 .1296 3.0662 .0027

X*U -1.0373 .7265 -1.4279 .1561

(36)

Table 23: hypothesis testing

Hypothesis Statement Overall

Outcome

sample 0

Outcome

sample 1 H1 “Analytics” are more likely to adopt a product

innovation which is meant as a healthier and/or more environmental friendly alternative to a highly appreciated product.

Rejected Rejected Rejected

H2 Consumers’ “taste for novelty” will positively influence the relationship between analytic thinking and the likelihood of adopting a product innovation.

H3 Consumers’ “taste for variety” will positively influence the relationship between analytic thinking and the likelihood of adopting a product innovation.

H4 Consumers’ “general resistance to change” will negatively influence the relationship between analytic thinking and the likelihood of adopting a product innovation.

(37)

Discussion

Today the economy can be typified by continuous innovation resulting in myriad new product launches every day (Cornescu and Adam, 2013). In such an ever-changing business environment, it is essential that these continuously developing new products are effectively adopted (Makkonen, Johnston and Javalgi, 2016) as innovation adoption is a vital determinant of organizational competitiveness (Makkonen, Johnston and Javalgi, 2016). For that reason, it is of critical importance that stakeholders involved in innovations understand why consumers adopt or resist certain new product innovations (Arts, Frambach and Bijmolt, 2011). The findings of this study reveal several important insights in this regards. The overall aim was to explore whether certain behavioral or personality traits give reason for the barriers that results in consumer innovation resistance (Claudy, Garcia and O’Driscoll, 2015). To be precise, this paper aimed to contribute overcoming the pro-change bias by delivering first empirical evidence on how consumers’ personality traits influence the relationship between analytic thinking and the likelihood of adopting a product innovation that is meant as a substitute for a highly appreciated product. The launch of several innovative food products that are healthier and/or more environmental friendly, are seen as the “alternatives” to highly appreciated products such as meat, dairy and sugar. In order to make sure the two separate data collection methods used in this analysis did not result in significant differences, the analysis is replicated on split samples (0=Facebook/Linked, 1=Mturk).

The results of this study demonstrate that for the overall sample there was first of all no significant relationship between analytic thinking and innovation adoption. This surprisingly contradicts numerous studies that emphasize the increased capacity of analytics to estimate new products utility within consumers’ day-to-day existence (Ram and Sheth, 1989). Regarding the finding that analytic thinking does not predict innovation adoption significantly well, one explanation may be that analytics are not able to see only one side of a controversial issue (Norton, 2011). Analytics clearly evaluate the pros and cons of every issue, meaning the glass can be both half full and half empty simultaneously. Put differently, analytics are better able to understand what to purchase in order to safeguard their health and the environment (Quintais, 2006) but also critically numerate the concessions related to the consumption of alternative products (Pennycook, Fugelsang and Koehler, 2015). These concessions might become barriers and outweigh the improved ecological welfare, animal welfare and personal health related to the adoption of a new product. In addition to that, it can be hard for analytics to break their habits as

(38)

they are generally more resistant to change compared to non-analytics (Norton, 2011). An additional and more general explanation may be that as the IQ scale produced low reliability, the level of analytic thinking (IQ) may not have been reflected property and thus led to insignificant results.

Secondly, the results of the overall sample showed that education predicts innovation adoption significantly well. This is in line with Roberts (1996) who concluded that green behaviors are affected by education (Roberts, 1996). Regarding this finding a possible explanation might be that today’s education programs often teach consumers about environmental issues, health care matters, nutritious, product safety etc. (Bloom, 1976). The more people are exposed to such programs, the less likely they are to purchase products which could produce health or environmental related problems (Bloom, 1976). Put differently, higher educated consumers buy fewer products that are potentially harmful to their own health or the environment such as sugar- laden foods, inefficient (polluting) cars and cigarettes. This is in line with three different studies saying that those with a higher level of education are more likely to respond to products that are environmental friendly (Roberts, 1996; Chan, 2008; Arminda M Finisterra Do Paço, Mário Lino Barata Raposo and Walter Leal Filho, 2009).

Thirdly, Sample 1 demonstrated no significant dissimilarities compared to the overall sample, meaning that sample 1 is representative of the overall sample. Yet, sample 0 deviated from the original results as the linear regression analysis demonstrated that analytic thinking predicts innovation adoption significantly well. Surprisingly, for 1 standard deviation increase in analytic thinking, the likelihood of innovation adoption decreases .261 standard deviations. A possible explanation for this result might be the previously appointed challenge of analytics stating they evaluate every issue on the basis of both pros and cons (Norton, 2011), compared to non-analytics who only see one side to a controversial issue. For instance, during the consumption of a certain innovative food alternative, analytics recognize the added value to their personal health and the environment but might conclude these benefits do not outweigh the potential loss in taste or texture. The fact that this relationship is only significant in sample 0 might be due to the fact that this sample consists of 46% women compared to 36% in sample 1; knowing that female consumers buy generally less products that are potentially harmful to their own health or the environment (Laroche, Bergeron and Barbaro-Forleo, 2001). Moreover, approximately 50% of sample 0 is between 25 and 34 years old while women between the age of 25 and 34 years old are most likely to adopt a vegan diet (Fisher, Bashyal and Bachman, 2012) (Elkington, 2011). This also explains the outcome that sample 0 demonstrate a significant difference between gender and innovation adoption. The fact that in this study education and gender predict

Understanding behavioral responses to product innovation : how do consumers’ personality traits influence the relationship between analytic thinking and the likelihood of innovation adoption?