Effectiveness of regional advertising in the health insurance market: the influence of regional attachment

(1)

Effectiveness of regional advertising in the

health insurance market: the influence of

regional attachment

By

Laura Hooijman

June 2020

University of Groningen Faculty of Economics and Business

(2)

2

Effectiveness of regional advertising in the health

insurance market: the influence of regional

attachment

By Laura Hooijman Wellerzandweg 9 8314PW Bant (+31)6 30887409 l.s.hooijman@student.rug.nl / laurahooijman@hotmail.com Student number: S3830977 June 2020 University of Groningen Faculty of Economics and Business

Msc. Marketing Management

(3)

3

Abstract

The health insurance industry has a vital role in society. Health care expenses are rising and the Dutch government is stimulating to apply a more regional focus. This study focuses on a health insurance case company in the Netherlands that is considering adjusting its marketing efforts to be regionally focused. The goal of the study was to generate insights into the moderating role of regional attachment on the relationship between advertising type and advertising effectiveness. Moreover, predictors of place attachment are investigated as predictors of regional attachment. This study aspired to answer the following research question: Does regionalized advertising increase advertising effectiveness in the health insurance market, and to what extent is this effect influenced by regional attachment?

The research consisted of a survey experiment: subjects were assigned to a regional advertising or national advertising condition using stratified sampling. It was found that regional attachment does not influence the relationship between advertising type and advertising effectiveness. However, regional attachment influences advertising effectiveness, meaning that respondents with a high degree of regional attachment evaluated the advertisement more positively than respondents with low regional attachment, irrespective of advertising type. Moreover, it was found that social ties, residence length, and age are predictors of regional attachment.

(4)

4

Acknowledgements

(5)

5

Chapter 1. Introduction

The health insurance industry has a vital role in society. In the Netherlands, this industry is competitive. Given comparable products, health insurance companies mostly differentiate on price. However, there is a shift toward differentiation through customer value creation.

The Dutch government is stimulating this movement. Due to improving medical technologies and wellness that directly affect aging and population increases, health care expenses in the Netherlands are rising (RIVM, n.d.). Therefore, the Dutch government is working toward “accessible and affordable health care, now and in the future, tailored to the needs of people who receive this care” (Tweede Kamer, 2019, p. 5). One means of facilitating this mandate is a more local/regional focus (Tweede Kamer, 2019).

This study focuses on a health insurance case company in the Netherlands that is considering adjusting its marketing efforts to be regionally focused, because people feel connected to the region in which they live. This connection is phrased as regional attachment. Regional attachment builds on the concept of place attachment, which is a “positive connection or bond between a person and a particular place” (Altman & Low, 1992). To test the added value of a regional focus, the researcher formulated the following research question:

Does regionalized advertising increase advertising effectiveness in the health insurance market, and to what extent is this effect influenced by regional attachment?

The outcome of this research has both practical and theoretical relevance. The unique contribution of this research to literature is the effect of regional attachment on the relationship between advertising type and advertising effectiveness in the health insurance market. While there has been previous research about regional attachment and advertising effectiveness, none has explicitly focused on the moderation effect of regional attachment on the relationship between advertising type and advertising effectiveness in the health insurance market. Therefore, this research leads to new, valuable insights.

(7)

7 identify advertising themes, customer pain points, and product service adjustments. Thus, this research focuses on the influence of regional attachment on the relationship between regional advertising and advertising effectiveness to provide new and valuable insights about the circumstances under which regional marketing may be effective.

(8)

8

Chapter 2. Theoretical framework

This chapter provides an overview of the relevant literature, starting with literature about advertising in general. Next, literature about regional advertising and place attachment appears. Finally, this chapter concludes with existing literature on predictors of place attachment.

2.1 Advertising

The core of this research is advertising effectiveness in the health insurance market. Much research has proven that advertising is effective, leading to affective, behavioral, and cognitive consumer responses (Fennis & Stroebe, 2015). Petty, Cacioppo, and Schumann (1983) found that advertising effectiveness depends heavily on the level of involvement of the receiver. Involvement is the “extent of perceived personal relevance of a brand, product or product category” (Fennis & Stroebe, 2015, p. 415). Petty et al. (1983) studied advertising effectiveness and found two routes to persuasion: central and peripheral. This discovery led to the Elaboration Likelihood Model (ELM), where elaboration of the advertisement is the amount of thought the consumer gives an advertisement (Petty & Cacioppo, 1986). The highly-involved customer is also likely to have high elaboration, with logical reasoning and the formation of an attitude leading to a logical central route in decision making (Petty & Cacioppo, 1986). Moreover, Petty et al. (1983) found that personal relevance profoundly affects the motivation to process stimuli.

The findings of Petty et al. (1986) support the theory of how consumers acquire, represent, and encode advertising information by Greenwald and Leavitt (1984). There are four stages: preattentive analysis, focal attention, comprehension, and elaborative reasoning (Greenwald & Leavitt, 1984). How consumers proceed through these four stages is heavily influenced by outcome-relevant involvement (Greenwald & Leavitt, 1984). Outcome-relevant involvement is where the outcome of the purchase has significant personal consequences for the consumer (Johnson & Eagly, 1989). In the case of low involvement, information is processed outside the conscious control in the preattentive analysis (Greenwald & Leavitt, 1984). In the case of high involvement, information is processed in the elaborative reasoning stage, leading to new beliefs, evaluations, and conclusions (Greenwald & Leavitt, 1984).

(9)

9 positively influences the effectiveness of advertising. Additionally, Te’eni-Harari, Lehman-Wilzig, and Lampert (2009) found that product involvement is essential for young people to enhance advertising effectiveness. One explanation for this positive relationship between involvement and advertising effectiveness is the theory of processing fluency, which “assumes that the increased ease of processing is experienced as pleasant and that this positive affect will be used as information in the evaluation of the stimulus” (Fennis & Stroebe, 2015, p. 413).

Considering the literature above, the perceptual fluency of the advertisement and the motivation to process the stimuli are critical when considering how to optimize advertising effectiveness.

2.2 Regional focus

Perceptual fluency and involvement matter in personalized advertising, because Sundar and Marathe (2010) found that a personalized or tailored advertisement is more relevant to the receiver. Shanahan, Tran, and Taylor (2019) studied advertisement personalization in the social media context and found that “perceived personalization positively impacts consumer brand engagement and brand attachment, both enhancing perceived quality and brand loyalty” (Shanahan et al., 2019, p. 57).

The reasoning behind regionalized advertising and personalized advertising is partly similar. Regionalized advertising is not as relevant as personalized advertising but is more relevant than national advertising, so one might expect that regionalized advertising leads to a more positive attitude toward the advertised brand than national advertising. In this manner, regionalized advertising leads to more perceived personal relevance, which leads to better marketing performance. The researcher expects that regionalized advertising leads to a more positive evaluation of the advertisement than national advertising. Hence, a regional focus is beneficial.

A regional or local focus might not benefit all companies, depending on how the customer base experiences regionalized advertisements. Considering the regional component, it is to be expected that regional attachment, the connection people have to the region they live in, influences the relationship between regional advertising and how the advertisement is evaluated by customers.

(10)

10 neighborhoods, and cities (Hidalgo & Hernandez, 2001). Moreover, Toruńczyk-Ruiz and Martinović (2020) argued that place attachment could apply on various scales, from apartments to cities to even larger regions. Therefore, place attachment is a useful concept to explain regional attachment.

Hashemnezhad, Heidari, and Hoseini (2013) reviewed the literature about a sense of place and place attachment and found that place attachment is a subset of a sense of place. With the concept of sense of place, people feel, perceive, and attach meaning to places; place attachment is the positive emotional and functional bond between places and people (Hashemnezhad et al., 2013). Thus, in people-place interactions, a sense of place is a general connection to a place, and “place attachment is a positive emotion which people have about the place” (Hashemnezhad et al., 2013, p. 11). According to Scannell and Gifford (2017), place attachment leads to psychological benefits in the following categories: “memories, belonging, relaxation, positive emotions, activity support, comfort-security, personal growth, freedom, entertainment, connection to nature, practical benefits, privacy, and aesthetics.”

Place attachment manifests in various levels and forms (Williams & Vaske, 2003). After extensive review, Scannell and Gifford (2010) clarified place attachment with a tripartite framework of person, place, and process, leading to the following definition: “a bond between an individual or group and a place that can vary in terms of spatial level, degree of specificity, and social or physical features of the place, and is manifested through affective, cognitive, and behavioral psychological processes” (Scannell & Gifford, 2010, p. 5).

The focus of this research is the potential of a regional focus in marketing by comparing regional and national advertising and their differences in advertising effectiveness based on the concept of place attachment. The researcher anticipates that a positive relation between a person and region can positively influence the evaluation of regional advertising as more relevant compared to national advertising. In sum, the effect of advertising type on advertising effectiveness is moderated by the degree of regional attachment, such that the effect is stronger when the degree of regional attachment is higher, which leads to the first hypothesis:

(11)

11

2.3 Factors predicting regional attachment

It is essential to identify who is sensitive to regional marketing. This researcher hypothesizes that people who demonstrate higher regional attachment are more sensitive to regional marketing. Therefore, the next step is to identify the factors that influence the degree of regional attachment. Given that place attachment is a useful concept to explain regional attachment, predictors of place attachment can indicate factors that influence the degree of regional attachment (predictors of regional attachment).

Lewicka (2011) reviewed the literature on place attachment, including a summary of its predictors divided into three categories: sociodemographic, social, and physical-environmental (Lewicka, 2010). These predictors lead to the second hypothesis:

(H2): Predictors of place attachment influence the degree of regional attachment.

The following section elaborates on these three categories of predictors, which leads to sub-hypotheses.

2.3.1 Sociodemographic predictors

Hidalgo and Hernandez (2001, p. 273) found that “the degree of attachment varies with age and sex,” which is in line with the findings of Pretty, Chipuer, and Bramston (2003) that implied that age and sex influence the degree of regional attachment. However, it is still unclear which age groups or sex have the highest degree of place attachment. Moreover, Lewicka (2005) found a negative relationship between education and place attachment; thus, education is a negative predictor of place attachment. In the literature, socioeconomic status influences attachment (Lewicka, 2011), and Fried (1984) found that socioeconomic status influences the relative importance of physical and social factors. However, there is no consensus in the literature about the effect of socioeconomic status on the degree of place attachment.

Given the erratic patterns of the predictors above, Lewicka (2011) argued that these predictors are probably mediated or moderated by additional factors. Nevertheless, the hypotheses include these factors to avoid discarding valuable information as these predictors are essential in many different studies. The researcher also values the study of these predictors in the context of the health insurance market.

(12)

12 example, sustained periods of absence can decrease attachment but also “make the heart grow fonder,” leading to an increased attachment (Gustafson, 2009; Van der Klis and Karsten, 2009; Lewicka, 2011). Thus, mobility is both a positive and negative predictor of place attachment.

A further predictor of place attachment is home ownership, which is a very consistent predictor and often used as a proxy measure (Lewicka, 2011). The findings of Lewicka (2010) show that this predictor holds for different types of housing, from pre-war tenements to family houses.

When studying a sense of place among adolescents and adults, Pretty et al. (2003) found that the life stage of residents is vital to consider when linking a resident with the community. Based on many studies, Lewicka (2011, p. 216) concluded that residence length is “the most consistent positive predictor of attachment to residence places.” Length of residence predicts place attachment both directly and indirectly by affecting the strength of local ties, which is a social predictor of place attachment (Lewicka, 2010).

The abovementioned literature leads to an extension of the second hypothesis with the following sub-hypothesis:

(H2a): Age, sex, socioeconomic status, educational background, mobility, home ownership, and residence length are predictors influencing the degree of regional attachment.

2.3.2 Social predictors

The main social predictor of place attachment is neighborhood and community ties, which is operationalized in different manners (Lewicka, 2011): the strength of neighborhood and community ties, the extensiveness of neighborhood and community ties, and involvement in informal social activities in the neighborhood and community (Lewicka, 2011). Given the similarity in the designations of the strength of neighborhood ties, the strength of community ties, and the extensiveness of neighborhood and community ties, the researcher combined these predictors into one item measuring community ties.

(13)

13 approach than this research that studies attachment to an entire region, where the size of the region might be less relevant. Therefore, the hypotheses in this study do not include this predictor.

Based on several studies, Lewicka (2011) concluded that sense of security is a consistent positive predictor of place attachment. Brown, Perkins, and Brown (2004) found that weak place attachment is a predictor of crime. Additionally, Lewicka (2011) confirmed that sense of security is a significant predictor of place attachment.

This literature leads to an extension of the second hypothesis with the following sub-hypothesis: (H2b): Community ties and sense of security are predictors of regional attachment.

2.3.3 Physical-environmental predictors

Physical predictors of place attachment are difficult to operationalize and measure (Lewicka, 2011). One approach uses the Index of Perceived Residential Environment Quality (PREQ) designed by Fornara, Bonaiuto, and Bonnes (2009), which includes an extensive array of physical parameters, such as recreational services and the presence of green areas. However, the focus of this research is on the effectiveness of regional marketing, identifying who is sensitive to regional marketing. Therefore, this study focuses on individual-level factors: sociodemographic predictors and social predictors. Physical-environmental predictors do not fit the scope of this study and are not discussed further.

The hypotheses formulated above appear graphically in a conceptual model in figure.

(14)

14

Chapter 3. Research methods

This chapter discusses the research design and methods of the study, developed to test the hypotheses of the theoretical framework, as well as the procedures, operationalization of the variables, and plan of data analysis.

3.1 Data collection

3.1.1 Participants

The researcher sent the survey to a subset of the company’s database; consequently, this study exclusively uses data from people insured at this health insurance company. The subset was selected randomly. To create as much diversity as possible, the researcher included people of all age ranges, sexes, socio-economic classes and educational levels. Only people from the regions of Noord, Midden, Achterhoek, Twente, and West appear in the subset, because the insurance company focuses on these regions. Because the researcher sent the survey to participants living in specific regions of the Netherlands, the language of the survey and the stimuli is Dutch.

The initial sample size is 5.000 people, with an expected a completion rate of at least 10%, this would lead to a response rate of 500. The final response rate was 478 (N = 478).

The gender distribution was 51,7% male and 48,3% female respondents. The mean age of the respondents was 62 years old, with the youngest respondent being 18 and the oldest respondent being 97. The age distribution appears in figure 2.

(15)

15 The respondents were from all regions but slightly more from Achterhoek (25,3%) and Noord (25,1%) as compared to Twente (18,5%), Midden (17%), and West (14,2%). The respondents were from all types of educational backgrounds; however, 35% of the respondents had a secondary vocational (MBO) education level, as appears in table 1. Socioeconomic class is the division of society based on education, profession, and income. Social class B1 (26,3%) and B2 (33,9%), which are the middle and upper-middle classes, are over-represented in the data, as appears in table 2. Frequency Percent Missing 1 0,2 Primary education 56 11,5 LBO / VMBO 79 16,2 MAVO / MULO 68 14,0 HAVO / VWO / HBS 28 5,7 MBO 172 35,3 HBO 61 12,5 WO / MBA / Postdoctoral 22 4,5 Total 487 100,0

Table 1. Distribution of educational background, based on the Dutch education system.

Frequency Percent Missing 1 0,2 Sociale klasse D 72 14,8 Sociale klasse C 47 9,7 Sociale klasse B2 165 33,9 Sociale klasse B1 128 26,3 Sociale klasse A 74 15,2 Total 487 100,0

Table 2. Distribution of socioeconomic class.

(16)

16 3.1.2 Research design

The researcher created a survey according to the structure of the conceptual model. The survey started with questions about the predictors of regional attachment as found in the literature and is followed with questions about the subjects’ degree of attachment to the region in which they live. The second part of the survey consisted of a survey experiment. The independent variable received treatment by assigning subjects to the regional advertising (version A) or national advertising (version B) condition using stratified sampling, such that all regions are represented in both conditions (Boeije, t Hart & Hox, 2009). Subjects read one advertising message (version A or B), after which the survey continued. The dependent variable was advertising effectiveness, measured by questions about positive/negative evaluations of the message based on the studies of MacKenzie and Lutz (1989), and Currás-Pérez, Bigné-Alcañiz, and Alvarado-Herrera (2009).

The researcher sent the survey link via email. When participants opened the link, they read an introduction to the study. To ensure that all participants understood the specific regions in the survey, the researcher defined each region at the start of the survey. After which the questions followed. The full survey appears in Appendix A.

3.1.3 Materials

The 10 predictors of place attachment, as found in the theoretical framework, were operationalized. The dataset of the health insurance company already included designations for age, sex, place of residence, socioeconomic status, and level of education. This resulted in five concepts: length of residence, home ownership, mobility, sense of security, and community ties that the researcher operationalized into statements. Participants indicated on a five-point Likert scale how much they agree with the statement (Dawes, 2008). Examples of the questions include “I greet my neighbors when I see them,” and “I own a house.”

The researcher measured community ties with five items on a five-point Likert scale, which are computed in one summed variable: m=3,818, SE=0,034, α=0,672.

(17)

17 Lewicka (2008), which consisted of 12 items on a five-point Likert scale. The scale was highly reliable and was summed in one regional attachment variable: m=3,934, SE=0,028, α=0,848.

The last part of the conceptual model is the moderating influence of the degree of regional attachment on the relationship between advertising type and advertising effectiveness. To measure the relationship between advertising type and advertising effectiveness, the researcher manipulated an out-of-home (OOH) advertisement and used it as stimuli. Participants saw either version A (a regional advertisement) or version B (a national advertisement). A free translation of the regional advertisement for Twente is: “Twente, you are beautiful...when we stand strong together.” A free translation of the national advertisement is: “Human, you are beautiful...when we stand strong together.” The stimuli of version A were used as an OOH advertisement in 2019. The stimuli of version B is designed for this study.

To measure the effect of the manipulation, the researcher operationalized the dependent variable, advertising effectiveness, by measuring nine items on a five-point Likert scale and one item on a scale of 1-10. Examples of the questions measuring advertising effectiveness are “the advertisement is convincing” and “this advertisement positively affects my opinion about the company.” Moreover, respondents rated the advertisement on a scale of 1 to 10. The 10 items are significantly correlated with high internal consistency (α = 0,909). However, if the researcher deleted the rating of the advertisement from 1-10, Cronbach’s alpha increased from 0,909 to 0,917. Therefore, the summed variable of advertising effectiveness only consisted of nine items (m=3,79, SE=0,028, α=0,917). The full questionnaire appears in Appendix A.

One last remark about the survey: only respondents who saw version A of the stimuli (regional advertisement) received the manipulation check question.

3.2 Data analysis

The data analysis appears in this section. The researcher analyzed the collected data using software Microsoft Excel and IBM SPSS Statistics.

3.2.1 Preparing data for analysis

(18)

18 received a reminder email and could stop and continue at any time, lengthy response time was not an issue. However, the researcher did not deem respondents who spent a short time reading questions as serious and deleted them. When using a baseline of three minutes, the researcher deleted 29 responses.

Secondly, the researcher checked for missing values that can lead to inaccurate predictions from biased/inefficient estimates (Field, 2013). Moreover, certain values were missing due to routing in the survey, but the researcher did not have to account for these. Missing values are not included in the analyses.

Third, the researcher checked for irregularities in the length of residence variable by computing a new variable (age – length of residence). Values, where the length of residence was equal to or greater than two years longer than the participant’s age, were assigned as missing values since the researcher could not trust those values. No other irregularities appeared.

Fourth, the researcher checked for outliers, which are observations that are quite different from most others, because outliers can affect descriptives such as the mean, the standard error, and confidence intervals (Field, 2013). The researcher discovered most outliers in the length of residence variable and the mobility variable, given that these variables are the only questions where respondents could type their answers. Each identified unlike value was individually evaluated using logical reasoning and deleted when necessary. For example, the researcher deleted values claiming the respondent moved more than 20 times, given that the average times a person moves in a lifetime is approximately eight times.

(19)

19 3.2.2 Hypotheses testing

To test hypothesis H1, the researcher performed a moderation analysis to determine how the degree of regional attachment moderates the relationship between advertising type and advertising performance. The first step was checking the Cronbach’s alpha, thereafter, the researcher performed the moderation analysis using model 1 of the PROCESS macro in SPSS (Hayes, 2012; Hayes, 2020).

(20)

20

Chapter 4. Results

This chapter presents the results of the data analysis.

4.1 Introduction to the data

An essential element of this research is regional attachment. The average regional attachment over the full dataset is 3,93 out of 5. Regional attachment is the strongest in Achterhoek, with a value of 4,09. West is the weakest region, with a value of 3,63.

Only 60% of the respondents affirmatively answered the manipulation check about noticing that the advertising was specific for the region. This question did not apply to respondents who saw the national advertisement and did not receive the question. To verify whether the answer to this question skews the results, the researcher performed the moderation analysis twice: the first time with the full dataset and the second time with only cases who answered the question correctly or where the question was not applicable, which is 362 cases.

4.2 Testing hypothesis 1

To analyze whether regional attachment influences the relationship between advertising type and advertising performance, the researcher performed a regression analysis using the PROCESS macro of Hayes (2012; 2020). The regression analysis was not significant (R2=0,0350, F=5,4914, p=0,001), as appears in table 3. Advertising type (regional or national) is not significantly predicting advertising effectiveness and the interaction effect of advertising type and regional attachment is not significantly predicting advertising effectiveness. Thus, no moderation effect of regional attachment on the relationship between regional advertising and advertising performance exists. The degree of regional attachment does not influence the relationship between advertising type and advertising effectiveness.

Note. R2 = 0,04; F = 5,49; p = 0,00

Table 3. Moderation analysis using PROCESS.

b SE B T p

Constant 3,38 0,03 119,70 0,00

Advertising type -0,00 0,03 -0,11 0,91

Regional attachment 0,19 0,05 4,03 0,00

Regional attachment x advertising type

(21)

21 Besides rejecting the hypothesis, the researcher found a different main effect. To analyze whether regional attachment influences advertising effectiveness, the researcher performed a simple regression with regional attachment as the independent variable and advertising effectiveness as the dependent variable. The regression analysis was significant: R2=0,035, F=16,465, p=0,000. Respondents with a high degree of regional attachment evaluated the advertisement more positively than respondents with low regional attachment, irrespective of the type of advertising. Thus, regional attachment directly influences advertising effectiveness (B=0,188, t=4,058, p=0,000).

Moreover, the researcher performed a regression analysis using the PROCESS macro of Hayes (2012; 2020) to analyze whether regional attachment influences the relationship between advertising type and advertising performance when only using cases that answered the manipulation check correctly or where the manipulation check was not applicable. The regression analysis was not significant (R2=0,0449, F=5,9711, p=0,001), as appears in table 4. Advertising type (regional or national) does not significantly predict advertising effectiveness, and the interaction effect of advertising type and regional attachment does not significantly predict advertising effectiveness. Therefore, no moderation effect of regional attachment on the relationship between regional advertising and advertising performance exists. The degree of regional attachment does not influence the relationship between advertising type and advertising effectiveness. Only including cases that correctly answered the manipulation question or where the question was not applicable does not lead to a different outcome of the moderation analysis.

Note. R2 = 0,05; F = 5,97; p = 0,00

Table 4. Moderation analysis using PROCESS, only including cases which answered the manipulation check correctly or where this question was not applicable.

Moreover, the researcher performed a simple regression with regional attachment as the independent variable and advertising effectiveness as the dependent variable, only using cases that answered the manipulation question correctly or where this question was not applicable.

b SE B T p

Constant 3,41 0,03 113,60 0,00

Advertising type -0,02 0,03 0,77 0,44

Regional attachment 0,21 0,05 4,13 0,00

Regional attachment x advertising type

(22)

22 The regression analysis was significant: R2=0,043, F=17,211, p=0,000. Respondents with a high degree of regional attachment evaluated the advertisement more positively than respondents with low regional attachment, irrespective of the type of advertising. Thus, regional attachment directly influences advertising effectiveness: B=0,210, t=4,149, p=0,000.

Notably, except from small changes in the output between the first and second time performing the moderation analyses, the analyses produced the same result. Regional attachment does not influence the relationship between advertising type and advertising effectiveness. Regional attachment does influence advertising effectiveness.

4.3 Testing hypothesis 2

Hypothesis H2 consists of several predictors of place attachment expecting to influence the degree of regional attachment. These predictors are age, sex, socioeconomic status, educational background, mobility, home ownership, residence length, community ties, and sense of security.

(23)

23 b SE B Beta p VIF (Constant) 3,426 ,273 ,000 Social ties ,151 ,038 ,186 ,000 1,084 Sense of security ,012 ,026 ,021 ,659 1,093 Residence length ,014 ,002 ,458 ,000 1,480 Age -,009 ,003 -,199 ,002 1,955

Mobility (times moved) -,003 ,011 -,011 ,821 1,128

Home ownership -,134 ,066 -,115 ,041 1,583

Sex -,053 ,056 -,043 ,351 1,047

Socioeconomic class ,020 ,028 ,042 ,459 1,566

Educational background -,009 ,016 -,025 ,576 1,028

Note: R2=0,215, F=10,725, p=0,000 Table 5. Multiple regression.

Moreover, the researcher performed a multiple regression with the predictors of place attachment as the independent variables and regional attachment as the dependent variable, only using cases that answered the manipulation question correctly or where this question was not applicable. The regression analysis was significant: R2=0,181, F=6,723, p=0,000. The regression analysis showed that sense of security, mobility, home ownership, sex, socioeconomic class, and educational background were not significantly predictive of regional attachment. Additionally, the regression analysis showed that social ties, residence length, and age significantly predicted regional attachment. Therefore, social ties, residence length, and age influence the degree of regional attachment. Moreover, the regression analysis showed that residence length is influencing regional attachment the strongest. The outcome of the multiple regression analysis only using cases that answered the manipulation question correctly or where this question was not applicable, appears in table 6.

b SE B Beta p VIF (Constant) 3,291 ,311 ,000 Social ties ,162 ,043 ,201 ,000 1,068 Sense of security ,012 ,030 ,022 ,683 1,084 Residence length ,013 ,002 ,424 ,000 1,460 Age -,009 ,003 -,200 ,002 1,478

Mobility (times moved) -,002 ,013 -,008 ,881 1,136

Home ownership -,073 ,077 -,064 ,345 1,728

Sex -,018 ,065 -,015 ,783 1,055

Socioeconomic class ,002 ,033 ,004 ,958 1,690

Educational background -,001 ,019 -,004 ,944 1,037

Note: R2=0,181, F=6,723, p=0,000

(24)

24 The first multiple regression analysis leads to different results than the multiple regression analysis only including cases which answered the manipulation check correctly or where this question was not applicable. Home ownership significantly predicted regional attachment the first time performing the analysis. However, the second time performing the analysis showed that home ownership is not significantly predictive of regional attachment. The researcher deemed respondents who answered the manipulation question affirmative as more serious. Therefore, the researcher trusts the result of the multiple regression only including cases which answered the manipulation check correctly or where this question was not applicable, meaning that home ownership is considered as not significantly predicting regional attachment.

The result of testing hypothesis 2 appears in table 7.

Table 7. Overview of predictors of hypothesis H2.

Accepted/ Rejected

Social ties (summed variable) Accepted

Sense of security Rejected

Residence length Accepted

Age Accepted

Mobility (times moved) Rejected

Home ownership Rejected

Sex Rejected

Socioeconomic class Rejected

(25)

25

Chapter 5. Discussion

The goal of the study was to generate insights into the moderating role of regional attachment on the relationship between advertising type and advertising effectiveness. Moreover, the researcher identified predictors of place attachment and investigated them as predictors of regional attachment. This study aspired to answer the following research question:

Does regionalized advertising increase advertising effectiveness in the health insurance market, and to what extent is this effect influenced by regional attachment?

The first hypothesis, the moderating effect of regional attachment on the relationship between advertising type and advertising performance, was rejected in the results section due to not having a significant effect of advertising type and interaction effect. However, the researcher did find that regional attachment influences advertising effectiveness, which means that respondents with a high degree of regional attachment evaluated the advertisement more positively than respondents with low regional attachment, irrespective of advertising type.

These findings do not support the reasoning that regionalized advertising, just as with personalized advertising, is perceived as more relevant by the receiver and, therefore, more positively evaluated (Shanahan et al., 2019). Considering that involvement is the “extent of perceived personal relevance of a brand, product or product category” (Fennis & Stroebe, 2015, p. 415), this reasoning assumed that the receiver of the regionalized advertisement has at least a moderate level of involvement. However, the survey did not measure the respondents’ involvement with the advertised brand. Therefore, the question why regional attachment influences advertising effectiveness and regional attachment does not influence the relationship between advertising type and advertising effectiveness remains unanswered.

(26)

26 Residence length was found as the strongest predictor of regional attachment. When studying the literature, Lewicka (2011) found that residence length was the strongest and most consistent predictor of place attachment. Thus, the finding that residence length is the strongest predictor in this study is in line with the literature discussed in the theoretical framework.

The practical relevance of these findings is that practitioners can use this information to identify (potential) customers with high regional attachment. Considering the conclusion that respondents with high regional attachment evaluate the advertisement more positively, identifying people with high regional attachment can be valuable for practitioners.

5.1 Limitations

This research has several limitations. First, the manipulation was constrained by available advertisements. The six advertisements differed in backgrounds, people in the picture, and text (see Appendix A3). The differences between age, sex, and appearance of the two people in the picture may lead to more or less recognition by the receiver, which could influence the perceived relevance by the receiver. Therefore, differences in the evaluation of the advertisement cannot be traced back with certainty to the manipulation.

Second, only 60% of the respondents affirmed the manipulation check that asked if respondents noticed that the advertising was specific for the region, which threatens the validity of this study. Performing the analyses with only those cases that answered the manipulation question affirmatively or where this question was not applicable, lead to significantly different results for the multiple regression analysis and did not lead to significantly different results for the moderation analysis. Hence, the researcher suggests to repeat this study with additional questions about the manipulation on the page where the manipulation appears. In this way, the researcher expects that respondents will take more time to read and inspect the advertisement, which in turn will result in more meaningful answers about the advertising effectiveness questions that follow.

(27)

27 researcher speculates that one reason might be that the involvement level for this health insurance brand is low.

5.2 Recommendations

Some questions stay unanswered and new questions arise from this study. A question that remains unanswered is why people with a high level of regional attachment evaluated the advertisement more positively than people with a low level of regional attachment, given that the type of advertising did not result in significant differences. The researcher suggests future research to include involvement level in the questionnaire, and to evaluate different types of advertising, which might lead to different results.

Moreover, the exploration of whether a regional focus might positively impact brand engagement and brand loyalty for customers with high regional attachment in the same manner as personalized advertising, as found by Shanahan et al. (2019), might lead to new, relevant insights.

Lastly, future research can expand this study to different regions and different types of brands. Regions with different cultures might have other predictors of regional attachment. Moreover, some brands with a long history in the region might already be embedded in the mind of residents as locally focused.

5.3 Conclusion

(28)

28

References

Low, S. M., & Altman, I. (1992). Place attachment. Place attachment (pp. 1-12). Springer, Boston, MA.

Boeije, H., t Hart, H., & Hox, J. (2009). Onderzoeksmethoden (8e druk). Den Haag: Boom Uitgevers.

Brown, B. B., Perkins, D. D., & Brown, G. (2004). Incivilities, place attachment and crime: Block and individual effects. Journal of environmental psychology, 24(3), 359-371.

Currás-Pérez, R., Bigné-Alcañiz, E., & Alvarado-Herrera, A. (2009). The role of self-definitional principles in consumer identification with a socially responsible company. Journal of business ethics, 89(4), 547.

Dawes, J. (2008). Do data characteristics change according to the number of scale points used? An experiment using 5-point, 7-point and 10-point scales. International journal of market research, 50(1), 61-104.

Fennis, B. M., & Stroebe, W. (2015). The psychology of advertising. Psychology Press.

Field, A. (2013). Discovering statistics using IBM SPSS statistics. Sage.

Fornara, F., Bonaiuto, M., & Bonnes, M. (2010). Cross-validation of abbreviated perceived residential environment quality (PREQ) and neighborhood attachment (NA) indicators. Environment and Behavior, 42(2), 171-196.

Fried, M. (1984). The structure and significance of community satisfaction. Population and Environment, 7(2), 61-86.

Gordon, M. E., McKeage, K., & Fox, M. A. (1998). Relationship marketing effectiveness: the role of involvement. Psychology & Marketing, 15(5), 443-459.

(29)

29 Gustafson, P. (2001). Roots and routes: Exploring the relationship between place attachment and mobility. Environment and behavior, 33(5), 667-686.

Gustafson, P. (2009). More cosmopolitan, no less local: The orientations of international travellers. European Societies, 11(1), 25-47.

Hashemnezhad, H., Heidari, A. A., & Mohammad Hoseini, P. (2013). Sense of place” and “Place Attachment. International Journal of Architecture and Urban Development, 3(1), 5-12.

Hayes, A. F. (2012). PROCESS: A versatile computational tool for observed variable mediation, moderation, and conditional process modeling.

Hayes, A. F. (2020). PROCESS macro for SPSS and SAS. Retrieved May 17, 2020, from http://www.processmacro.org/index.html

Hidalgo, M. C., & Hernandez, B. (2001). Place attachment: Conceptual and empirical questions. Journal of environmental psychology, 21(3), 273-281.

Johnson, B. T., & Eagly, A. H. (1989). Effects of involvement on persuasion: A meta-analysis. Psychological bulletin, 106(2), 290.

Lewicka, M. (2005). Ways to make people active: The role of place attachment, cultural capital, and neighborhood ties. Journal of environmental psychology, 25(4), 381-395.

Lewicka, M. (2008). Place attachment, place identity, and place memory: Restoring the forgotten city past. Journal of environmental psychology, 28(3), 209-231.

Lewicka, M. (2010). What makes neighborhood different from home and city? Effects of place scale on place attachment. Journal of environmental psychology, 30(1), 35-51.

(30)

30 Low, S. M., & Altman, I. (1992). Place attachment. In Place attachment (pp. 1-12). Springer, Boston, MA.

MacKenzie, S. B., & Lutz, R. J. (1989). An empirical examination of the structural antecedents of attitude toward the ad in an advertising pretesting context. Journal of marketing, 53(2), 48-65.

Petty, R. E., Cacioppo, J. T., & Schumann, D. (1983). Central and peripheral routes to

advertising effectiveness: The moderating role of involvement. Journal of consumer research, 10(2), 135-146.

Petty, R. E., & Cacioppo, J. T. (1986). The elaboration likelihood model of persuasion. In Communication and persuasion (pp. 1-24). Springer, New York, NY.

Pretty, G. H., Chipuer, H. M., & Bramston, P. (2003). Sense of place amongst adolescents and adults in two rural Australian towns: The discriminating features of place attachment, sense of community and place dependence in relation to place identity. Journal of environmental psychology, 23(3), 273-287.

RIVM (n.d.). Volksgezondheid Toekomst Verkenning 2018: Trendscenario zorguitgaven. Retrieved February 24, 2020, from https://www.vtv2018.nl/zorguitgaven.

Scannell, L., & Gifford, R. (2010). Defining place attachment: A tripartite organizing framework. Journal of environmental psychology, 30(1), 1-10.

Scannell, L., & Gifford, R. (2017). The experienced psychological benefits of place attachment. Journal of Environmental Psychology, 51, 256-269.

(31)

31 Sundar, S. S., & Marathe, S. S. (2010). Personalization versus customization: The importance of agency, privacy, and power usage. Human Communication Research, 36(3), 298-322.

Te’eni-Harari, T., Lehman-Wilzig, S. N., & Lampert, S. I. (2009). The importance of product involvement for predicting advertising effectiveness among young people. International Journal of Advertising, 28(2), 203-229.

Toruńczyk-Ruiz, S., & Martinović, B. (2020). The bright and dark sides of length of residence in the neighbourhood: Consequences for local participation and openness to newcomers. Journal of Environmental Psychology, 67, 101383.

Tweede Kamer. (2019). Inleiding: zorg voor de toekomst - Vaststelling begroting Ministerie van Volksgezondheid, Welzijn en Sport 2020. Retrieved March 12, 2020, from

http://www.rijksbegroting.nl/2020/voorbereiding/begroting,kst264861_5.html

Van der Klis, M., & Karsten, L. (2009). Commuting partners, dual residences and the meaning of home. Journal of Environmental Psychology, 29(2), 235-245.

Kotler, P. (2013) Marketing highlight: Het Mentality-model van Motivaction en persona's. Principes van marketing (p. 251-252). Amsterdam: Pearson Benelux

(32)

32

Appendices

Appendix A. Survey

A1. Invitation email

Onderwerp: Hoe verbonden bent u met uw regio?

(Onderwerp herinneringsmail: Herinnering: Hoe verbonden bent u met uw regio?) [[AANHEF]]

Hierbij nodigen wij u uit om deel te nemen aan het onderzoek naar de binding met uw regio. Graag horen wij uw mening over de onderwerpen die in dit onderzoek aan bod komen. Met de resultaten van dit onderzoek kunnen wij onze dienstverlening nog beter op uw behoeften afstemmen.

Door deel te nemen aan dit onderzoek geeft u toestemming om uw antwoorden te gebruiken voor analysedoeleinden. Uw antwoorden worden anoniem en vertrouwelijk behandeld en niet met derden gedeeld.

Het invullen van de vragen kost u slechts 5 minuten. U kunt het onderzoek bereiken via onderstaande link: Klik hier om direct deel te nemen

Met hartelijke groet, Menzis

A2. Reminder email

Onderwerp: Herinnering: Hoe verbonden bent u met uw regio? [[AANHEF]]

Hierbij nodigen wij u uit om deel te nemen aan het onderzoek naar de binding met uw regio. Graag horen wij uw mening over de onderwerpen die in dit onderzoek aan bod komen. Met de resultaten van dit onderzoek kunnen wij onze dienstverlening nog beter op uw behoeften afstemmen.

Door deel te nemen aan dit onderzoek geeft u toestemming om uw antwoorden te gebruiken voor analysedoeleinden. Uw antwoorden worden anoniem en vertrouwelijk behandeld en niet met derden gedeeld.

Het invullen van de vragen kost u slechts 5 minuten. U kunt het onderzoek bereiken via onderstaande link: Klik hier om direct deel te nemen

(33)

33 Menzis

A3. Questionnaire

1 Deze vragenlijst gaat over uw verbondenheid met de

regio [[REGIO]]. Met de regio Achterhoek worden de volgende gemeenten bedoeld: Aalten, Berkelland, Bronckhorst, Doetinchem, Montferland, Oost Gelre, Oude IJsselstreek en Winterswijk.Klik op de knop ‘Verder‘ om naar de eerste vraag te gaan.

Tussenpagina

VRAAG 1 ALLEEN TONEN ALS AAN DE ONDERSTAANDE VOORWAARDEN WORDT VOLDAAN, INDIEN NIET VOLDAAN SPRING NAAR: >> VOLGENDE VRAAG

de waarde uit verrijkingskolom REGIO is Achterhoek

regio [[REGIO]]. Met de regio Arnhem worden de volgende gemeenten bedoeld: Arnhem, Doesburg, Duiven, Lingewaard, Overbetuwe, Renkum, Rheden, Rozendaal, Westervoort en Zevenaar.Klik op de knop ‘Verder‘ om naar de eerste vraag te gaan.

Tussenpagina

de waarde uit verrijkingskolom REGIO is Arnhem

regio [[REGIO]]. Met de regio Gelderse Vallei worden de volgende gemeenten bedoeld: Barneveld, Ede, Renswoude, Rhenen, Veenendaal en

Wageningen.Klik op de knop ‘Verder‘ om naar de

(34)

34

eerste vraag te gaan.

de waarde uit verrijkingskolom REGIO is Gelderse Vallei

regio [[REGIO]]. Met de regio Rivierenland worden de volgende gemeenten bedoeld: Buren, Culemborg, Neder-Betuwe, Tiel en West Betuwe.Klik op de knop ‘Verder‘ om naar de eerste vraag te gaan.

Tussenpagina

de waarde uit verrijkingskolom REGIO is Rivierenland

regio [[REGIO]]. Met de regio Noord wordt de gehele provincie Groningen plus de gemeenten Noordenveld en Tynaarlo bedoeld.Klik op de knop ‘Verder‘ om naar de eerste vraag te gaan.

Tussenpagina

(35)

35

regio [[REGIO]]. Met de regio Twente worden de volgende gemeenten bedoeld: Almelo, Borne, Dinkelland, Enschede, Haaksbergen, Hellendoorn, Hengelo, Hof van Twente, Losser, Oldenzaal, Rijssen-Holten, Tubbergen, Twenterand en

Wierden.Klik op de knop ‘Verder‘ om naar de eerste vraag te gaan.

Tussenpagina

de waarde uit verrijkingskolom REGIO is Twente

regio [[REGIO]]. Met de regio West worden de volgende gemeenten bedoeld: Leidschendam-Voorburg, Pijnacker-Nootdorp, Rijswijk, ‘s

Gravenhage, Wassenaar en Zoetermeer.Klik op de knop ‘Verder‘ om naar de eerste vraag te gaan.

Tussenpagina

de waarde uit verrijkingskolom REGIO is West

8 De volgende stellingen gaan over uw relatie met uw

directe woonomgeving. Geef aan in hoeverre u het eens bent met de volgende stellingen.

(36)

36 Toon subvragen in willekeurige volgorde

Helemaal mee oneens Mee oneens Niet eens/niet oneens Mee eens Helemaal mee eens Ik herken mijn buren Ik groet mijn buren als ik ze op straat zie Ik erger mij aan mijn buren Mijn

reservesleutel ligt bij de buren Ik heb buren die mij kunnen helpen als ik ziek ben Ik maak mij zorgen om de criminaliteit in mijn buurt

9 Hoeveel jaar woont u al in de regio [[REGIO]]?

Deze vraag gaat over de gehele regio [[REGIO]] en dus niet alleen uw directe woonomgeving.

Open vraag (klein)

10 Hoe vaak bent u binnen de regio [[REGIO]] verhuisd?

Deze vraag gaat over de gehele regio [[REGIO]] en dus niet alleen uw directe woonomgeving.

(37)

37

11 Bent u in het bezit van een koopwoning?

Single-responsevraag

Ja Nee

Weet ik niet / Wil ik niet zeggen

12 De volgende stellingen gaan over de regio [[REGIO]].

Geef aan in hoeverre u het eens bent met de volgende stellingen.

Tabelvraag (single response)

Toon subvragen in willekeurige volgorde

(38)

38 Ik verlaat deze regio met plezier Ik zou niet graag ergens anders willen wonen Ik ben in deze regio opgegroeid Ik voel mij verbonden met deze regio

13 Hieronder ziet u een advertentie van Menzis.

Neem de tijd om deze te lezen. Het vervolg van de vragenlijst staat op de volgende pagina.

Tussenpagina

(39)

39

14 Hieronder ziet u een advertentie van

Menzis. Neem de tijd om deze te lezen.

Het vervolg van de vragenlijst staat op de volgende pagina.

(40)

40

Alle onderstaande voorwaarden zijn waar:

+ en de waarde uit verrijkingskolom NAT0_REG1 is 1 + en de waarde uit verrijkingskolom REGIO is Twente

(41)

41

Alle onderstaande voorwaarden zijn waar:

+ en de waarde uit verrijkingskolom NAT0_REG1 is 1 + en de waarde uit verrijkingskolom REGIO is Achterhoek

Tussenpagina

(42)

42 Alle onderstaande voorwaarden zijn waar:

+ en de waarde uit verrijkingskolom NAT0_REG1 is 1 + en de waarde uit verrijkingskolom REGIO is Noord

Tussenpagina

(43)

43 + en de waarde uit verrijkingskolom NAT0_REG1 is 1

+ en minstens één van onderstaande voorwaarden is waar: - of de waarde uit verrijkingskolom REGIO is Gelderse Vallei - of de waarde uit verrijkingskolom REGIO is Arnhem

- of de waarde uit verrijkingskolom REGIO is Rivierenland

Tussenpagina

(44)

44 Alle onderstaande voorwaarden zijn waar:

+ en de waarde uit verrijkingskolom NAT0_REG1 is 1 + en de waarde uit verrijkingskolom REGIO is West

19 Wat is uw algemene oordeel over deze advertentie,

uitgedrukt in een rapportcijfer?

Hierbij is 1 de laagste beoordeling en 10 de hoogste beoordeling die u kunt geven.

(45)

45 6 7 8 9 10

20 De volgende stellingen gaan over de advertentie van

Menzis die u zojuist heeft gezien. Geef aan in

hoeverre u het eens bent met de volgende stellingen.

Tabelvraag (single response) Helemaal mee oneens Mee oneens Niet eens/niet oneens Mee eens Helemaal mee eens De advertentie

spreekt mij aan De advertentie is overtuigend De advertentie komt geloofwaardig over De boodschap van de advertentie is duidelijk voor mij De advertentie past goed bij Menzis

Mijn beeld van Menzis is positief veranderd door deze advertentie Ik heb een positief gevoel bij Menzis Deze advertentie versterkt mijn band met Menzis Ik voel mij thuis bij een

(46)

46 advertenties

maakt

21 Was het u opgevallen dat de advertentie specifiek

voor de regio [[REGIO]] is?

Single-responsevraag

de waarde uit verrijkingskolom NAT0_REG1 is 1

Ja Nee

Weet ik niet zeker

22 Heeft u nog opmerkingen? Open vraag

Effectiveness of regional advertising in the health insurance market: the influence of regional attachment

Effectiveness of regional advertising in the

health insurance market: the influence of

regional attachment

Effectiveness of regional advertising in the health

insurance market: the influence of regional

attachment

Abstract

Acknowledgements

Table of contents

Chapter 1. Introduction

Chapter 2. Theoretical framework

Chapter 3. Research methods

Chapter 4. Results

Chapter 5. Discussion

References

Appendices