In recent years, the number of voluntary non-remunerated blood donations has stood still or has even declined in some countries. The regional office of the World Health Organization (WHO) for Europe states that ‘90% of individuals who are eligible to donate blood are not currently doing so’ (WHO, 2011). In contrast, the demand for blood products steadily increases. One reason for this trend is the demographic change, resulting in a growing number of golden-age individuals (Fendrich and Hoffmann, 2007; Greinacher et al., 2007; Zou et al., 2008). Furthermore, regional shortages, such as the outbreak of Escherichia coli infection (EHEC) in Germany, as well as emergencies and sudden disasters, increase the demand for large and safe blood reserves (Gilcher, 2003; McCarthy, 2007; Crawford et al., 2008). According to these challenges, all partners involved in collecting blood e.g., nonprofit blood centres, mainly the blood centre of the Red Cross, for-profit blood centres, transfusion institutes and public as well as private hospitals) are deeply interested in the various mechanisms that stimulate individuals to donate blood.
In order to stop or even change the negative trend, nonprofit blood centres in Germany have called for new approaches in blood donor management, for example, strategies to re-activate inactive or deferred donors (Seifried et al., 2011, p. 10).
Against this background, the results of a recently presented literature review on different mechanisms of charitable giving serve as a foundation for this study (Bekkers and Wiepking, 2011). The authors identified eight mechanisms as the most important forces that drive charitable giving in general. The eight mechanisms are (1) awareness of need, (2) solicitation, (3) costs and benefits, (4) altruism, (5) reputation, (6) psychological benefits, (7) values and (8) efficacy. After an isolated discussion of each mechanism, the authors concluded in their literature review: ‘We think that identifying systematic patterns in the mix of the mechanisms and interactions among them are important tasks for future research’ (Bekkers and Wiepking 2011, p. 944).
We agree strongly on this statement. Therefore, we decided to realize a literature review and an empirical study on systematic patterns in blood giving by an evaluation of the importance of altruistic values and satisfaction with the treatment on donor loyalty. The selection of the variables included in our study was taken because of the following two reasons.
First, for several decades, pure altruism has been seen as the key driver for giving blood in research studies (Titmuss, 1970; Oswalt and Napoliello, 1974). These studies reveal that the main reason for giving blood is altruism. Consequently, an altruistic motivated blood donor would respond by giving more blood in situations in which other individual give less. In the 1990s, some authors became critical and questioned whether ‘pure altruism’ really exists and concluded that the altruism model lacked predictive power (Andreoni, 1989, 1990; Piliavin, 1990; Andreoni and Rao, 2011). Today, it is widely accepted that the act of blood donation is difficult to understand and that besides pure altruism or altruistic values, other external factors influence the behaviour of blood donors (Glynn et al., 2002; Hupfer et al., 2005; Nilsson Sojka and Sojka, 2008).
Second, it is far more practical for blood centres to identify concrete instruments or management systems that are able to attract and retain blood donors than it is to induce or mobilise an altruistic motivation. This argument is congruent to the opinion of Healy (2000), who argued that psychological blood donor motives are difficult to influence. In this regard, the subsequent literature review comes to the result that the satisfaction with different components of the blood donation process (in the following labelled as ‘satisfaction with the treatment’) is a crucial factor for building blood donor loyalty. This factor can be measured and managed by a systematic satisfaction management system. Consequently, more knowledge on the positive effects of satisfaction with the treatment in comparison with other giving mechanisms is needed. Therefore, the purpose of this paper is to answer the following two research questions: How important is satisfaction with treatment, compared with altruistic values, for building blood donor loyalty? Are interactions among altruistic values and satisfaction with the treatment observable?
In the next section, we provide an interdisciplinary literature review of blood donation management. Following this, in the literature section, we compare the results of our survey of the literature on blood giving with the findings of Bekkers and Wiepking (2011) on charitable giving. We then develop a research model and derive four hypotheses. We test the hypotheses empirically by analysing survey data of 2149 blood donors of the German Red Cross, applying partial least squares (PLS) path analysis. After reporting our results, we discuss the implications of our findings for nonprofit research and practice.
Literature on blood donation management
Starting with Titmuss' (1970) work on crowding out in blood donation, a huge amount of articles related to blood donation management was published by an interdisciplinary research community. In order to systemise the findings, we conducted a systematic literature review (Tranfield et al., 2003) and analysed conceptual and empirical blood donation studies from different research disciplines between 1970 and 2010. To ensure that an interdisciplinary perspective was maintained, we conducted an online search in two databases: Business Source Complete (covering the research areas of economics, marketing and philanthropy) and Science Direct (covering the research areas of psychology, sociology and medicine). A detailed description of the review procedure is documented by Leipnitz (2010). From the 840 articles found, we identified 91 articles with a primary focus on blood donation management.
In line with the mechanisms of charitable giving described by Bekkers and Wiepking (2011), the findings of the studies on blood donation management provide evidence for a multitude of mechanisms responsible for the individual decision to donate blood. Therefore, we use their categorisation of the eight mechanisms to present our own findings of the literature review.
One of the main findings of our literature review is the predominance of altruism as a major mechanism for giving blood (Healy, 2000; Goette et al., 2010). As mentioned in the beginning, the work of Titmuss (1970) on altruism and crowding-out effects can be seen as a starting point of the whole discussion. Other authors continued Titmuss' work (1970) on crowding out effects. They found that in cases where the blood centre paid for the blood donation, the possibility to donate the money to a charity counteracted the normally resulting crowding-out effect (Mellström and Johannesson, 2008). Similarly, another study has shown that the payment of 10 Euros crowded out blood donations, but not the giving of a voucher (Lactera and Macis, 2010).
However, most of the studies identified in the literature review refer to altruistic values of blood donors (Oswalt and Napoliello, 1974; Oborne and Bradley, 1975; Oborne et al., 1978; Oswalt and Gordon, 1993). In recently published articles, the focus of the discussion changed to an analysis of the impact of altruistic values to acquire first-time blood donors (Glynn et al., 2002; Hupfer et al., 2005; Glynn et al., 2006; Nilsson Sojka and Sojka, 2008). In this regard, some studies found that altruistic values are not only important for the first blood donation but also a very important factor for building loyal behaviour (Trimmel et al., 2005; Glynn et al., 2006; Nilsson Sojka and Sojka, 2008).
Despite the fact that previous literature strongly supported the view that a blood donation is a typical altruistic act, several studies suggested that the measures of altruistic values are biased by social desirability (Ferguson et al., 2007; Ferguson et al., 2008). Ferguson et al. (2007, 2008) showed that blood donors were influenced by personal rather than altruistic beliefs.
Similar to charitable giving, studies on blood donation management confirm that different forms of solicitation can trigger blood donation behaviour (Glynn et al., 2002; Misje et al., 2005; Glynn et al., 2006; Harrington et al., 2007; Nilsson Sojka and Sojka, 2008). Studies concerning tangible forms of solicitation—for instance, personally addressed recruitment letters (Chamla et al., 2006) as well as gifts, compensatory incentives and tokens (Glynn et al., 2003)—show that they influence blood donor return rates positively. Intangible solicitation such as personal requests (Glynn et al., 2002; Glynn et al., 2006; Harrington et al., 2007), media campaigns (Misje et al., 2005; Hupfer, 2006; Nilsson Sojka and Sojka, 2008), e-mail requests (Geyer, 2005) or telephone recruitment (Lipsitz et al., 1989) have been documented as effective methods for acquiring first-time donors. Some researchers focused their analysis on lapsed blood donors and found that blood centres were able to re-activate the relationship if they made further requests for donations (Glynn et al., 2006; Harrington et al., 2007). In addition to this, Reich et al. (2006) found that e-mail recruitment was less effective than telephone requests.
Awareness of need
Several studies highlighted awareness of the need for blood donations as an important factor in an individual's decision to donate blood (Leibrecht et al., 1976; Boe and Ponder, 1981; Glynn et al., 2002; Pagliariccio et al., 2003; Hupfer et al., 2005; Harrington et al., 2007; Shaz et al., 2009). Shaz et al. (2009) found that the awareness of need varied significantly in the blood donor and the non-donor segments. The results of this study showed that 50% of the donors but only 35% of the non-donors were aware of the urgent need for blood. However, awareness is not the only powerful factor that attracts blood donors; the possibility of a close family member needing a blood transfusion in the future also influences the willingness to donate blood (Belda Suárez et al., 2004; Hupfer, 2006).
Costs and benefits
Previous literature on blood donation marketing illustrates that the mechanism of costs and benefits makes little difference in a blood donation context compared with other giving forms. Despite small expenses for attending a blood donation event (e.g., transportation costs), there are no monetary costs for the donor (Piliavin, 1990). On the basis of this reason, the term ‘costs of donating blood’ in an unpaid context is unusual. Instead, most of the researchers use the term ‘barriers’ (Piliavin, 1990). It is an accepted fact that potential blood donors need to overcome individual barriers such as fear of needles, as well as potential physical reactions such as dizziness, bruises or the possibility of diseases (Bartel et al., 1975; France et al., 2005; Harrington et al., 2007; France et al., 2008; Beerli-Palacio and Martín-Santana, 2009). Furthermore, blood donors have to deal with logistic factors such as transport connections, the limited dates of a blood event and/or limited time slots (Schreiber et al., 2006; Ringwald et al., 2007; Nilsson Sojka and Sojka, 2008; Reid and Wood, 2008). As regards the benefits of a blood donation, various authors confirm the effectiveness of blood credits or the offer of a medical check (Condie et al., 1976; Lightman, 1981; Sanchez et al., 2001; Glynn et al., 2003; Tscheulin and Lindenmeier, 2005; Mellström and Johannesson, 2008).
Furthermore, we identified studies that highlighted personal reputation as a main mechanism for giving blood. Some research found that individuals would sometimes try to fulfil the expectations of others (mostly family or friends) that he or she should join a group going to a blood donor event (Foss, 1983; Mc Combie, 1991). In this regard, other studies confirm that the reputation mechanism, which is a direct result of the influence of friends and family, is a decisive factor in the donation behaviour of a person (Condie and Maxwell, 1970; Condie et al., 1976; Foss, 1983; Mc Combie, 1991; Misje et al., 2005). Furthermore, empirical evidence exists that in the segment of first-time donors, the influence of friends or family is higher compared with the segment of regular blood donors (Condie and Maxwell, 1970; Oswalt and Hoff, 1974).
The power of the mechanism psychological benefits is evident in several studies (Callero and Piliavin, 1983; Nilsson Sojka and Sojka, 2003; Ferguson et al., 2008; Lemmens et al., 2009). Donors report feeling proud or experiencing a warm glow after or during the blood donation (Callero and Piliavin, 1983; Lemmens et al., 2009). This is in line with findings that donors exhibit more selfish than social reasons for their donations (Ferguson et al., 2008). They report positive effects such as the feeling of individual satisfaction, being more alert and feeling generally better than before the blood donation (Nilsson Sojka and Sojka, 2003).
Investigations into the values of donors as a mechanism for giving blood are relatively small because we already mentioned that most of the studies examined altruistic values, which are discussed under altruism. We identified in our review only three additional studies on values. The results of Burnett's research studies (1981, 1982) indicate that blood donors show lower levels of self-esteem, are family persons, more religious and more conservative compared with non-donors. Another study in this context confirms that blood donors tend to have high levels of prosocial characteristics such as altruistic values, empathy and social responsibility (Steele et al., 2008).
The efficacy mechanism is based on the donor's belief that his or her donation makes a remarkable difference for the beneficiaries and for society in general. The literature on efficacy as a mechanism in blood donor management is very scarce. However, we identified the following studies in our review. Andaleeb and Basu (1995) found that trust in the quality standards of the blood donation centre was an important prerequisite for efficacy. Moreover, the results of a large experimental study in which first-time donors were confronted with different messages regarding the use of their blood donation confirmed efficacy as a mechanism for blood giving. The authors noted that ‘a recruitment message appealing to the donor's empathy was almost 20 percent more effective than one appealing to self-esteem’ (Reich et al., 2006, p 1093).
Satisfaction with treatment
In addition to the eight previously mentioned mechanisms, we identified another factor as important on the donor's decision to donate blood: the satisfaction with the treatment during the donation process. Some studies found that a high satisfaction, affected by the staff friendliness, staff involvement/competence, the quality of information given by the blood centre to the donor in preparation for the blood event and the medical examination in general, is a crucial factor for a successful donation experience (Newman and Pyne, 1997; Daigneault and Blais, 2004; Moog, 2009). Other studies furthermore show that the satisfaction with the treatment has a strong impact on the intention to donate again in the future (Paulhus et al., 1977; Newman and Pyne, 1997; Daigneault and Blais, 2004; Nguyen et al., 2008; Moog, 2009). For instance, Nguyen et al. (2008) found that the satisfaction of donors varies among demographic and donation history subgroups but was always positively related with the intent to return for future donations. In line with such findings, Newman and Pyne (1997, p. 592) show that the quality of the service, offered by the blood donation organisation, becomes a major hygiene factor for the continuance of donating blood and state that an imperfect performance can reduce the number of returning blood donors. Therefore, the satisfaction with the treatment is an important factor for building and maintaining donor loyalty, which needs to be addressed by blood donation centres (Newman and Pyne, 1997; Nguyen et al., 2008; Moog, 2009).
Mix of blood giving mechanisms
To the best of our knowledge, we identified only one study that examines a mix of selected mechanisms in blood giving. Beerli-Palacio and Martín-Santana (2009, p. 207) developed a model of the predisposition to blood donation and conceptualised information (media campaign), experience, fear and motivation as the main mechanisms that drive people to give blood. The authors tested their model on the basis of a sample of 303 potential blood donors from Spain and found that the construct information (media campaign) had a significant and much stronger influence on the predisposition to donate blood compared with the construct motivations.
We summarise the literature review as follows: For decades, altruistic values have been seen as a major mechanism in blood donation, but contradictory results evoke doubts about its relative influence. The argument of previous reviews (Piliavin, 1990) that a blood donation is a complex act, which seems to be triggered by multiple mechanisms, is also supported by our review. Therefore, a deeper understanding of how the mechanisms are related to each other is needed.
Research model and hypotheses
Our research model is shown in Figure1. The basic rationale of the model is that both altruistic values and satisfaction with treatment have a positive impact on blood donor loyalty. Furthermore, we argue that these constructs are not isolated factors but connected to each other. However, no indication about the causal direction is given in the literature. Therefore, we tested both causal directions. Finally, our research model includes barriers that affect blood donor loyalty negatively.
Within the research model, we derived four research hypotheses. The first relationship we conceptualised is the path between altruistic values and blood donor loyalty. Drawing on the theory of warm glow giving (Andreoni, 1989) and on the empirical research findings discussed earlier, we argue that altruistic values are still one of the most important drivers for regular blood donation behaviour (Glynn et al., 2002; Hupfer et al., 2005; Nilsson Sojka and Sojka, 2008). For example, during the EHEC crisis in Germany, the willingness of first-time and sporadic blood donors to give blood again increased heavily. Especially remarkable is the fact that during the crisis, the highest donation rates were observed in the segment of unpaid blood donors, and not in the segment of paid blood donors. Thus, we offer our first hypotheses.
H 1. The more salient the altruistic values for blood donors, the higher the blood donor loyalty.
Furthermore, we conceptualised a path between satisfaction with treatment and blood donor loyalty. This argumentation is based on nonprofit marketing results, which show that a highly satisfied donor is more likely to be loyal than a less satisfied donor. This relationship is also evident in the blood donation context (Nguyen et al., 2008). Newman and Pyne (1997), for example, argue that the social interaction between a blood donor and a blood centre during the blood donation process (e.g., a warm welcome, a clean room or competent answers of the staff on questions) is of high importance for the satisfaction of blood donors and is a prerequisite for future donation intentions (Gotlieb et al., 1994; Storbacka et al., 1994; Brady and Robertson, 2001; Nguyen et al., 2008). Therefore, the enhancement of blood donor satisfaction with treatment is crucial for fostering the loyalty of donors (Newman and Pyne, 1997). This assumption holds especially true for first-time donors, who are not familiar with the donation process. Several studies show that prolapses such as vasovagal reactions deter donors and, consequently, reduce the return rates among first-time donors (Ferguson and Bibby, 2002; France et al., 2004; France et al., 2005). Therefore, our second hypothesis is as follows:
H 2. The higher the blood donor satisfaction with treatment, the higher the blood donor loyalty.
Furthermore, we assume that altruistic values and the satisfaction with treatment are not isolated factors but linked to each other. This is based on the proposition that several mechanisms and factors interact in the donation process (Frisch and Gerrard, 1981; Bekkers and Wiepking, 2011). However, limited knowledge is given about the direction of the effects. On the one hand, it seems logical that a high level of altruistic values may help to get over a bad experience of the treatment because altruistic-driven donors put first the needs of beneficiaries (Newman and Pyne, 1997). On the other hand, a positive experience of the treatment during the donation might overcome little pronounced altruistic values. Thus, we hypothesise both directions with the following:
H 3a. There is a direct effect of altruistic values on the satisfaction with treatment.
H 3b. There is a direct effect of the satisfaction with treatment on altruistic values.
The fourth path in our research model is the relationship between blood donor barriers and blood donor loyalty. Medical science research proves that there are a huge number of eligible donors who never donate or do not repeat their first donation because of the fears of blood donation or the inconvenience caused. Consistent with this assertion, we argue in our study that barriers play an important role in the decision process and negatively influence one's loyalty for a blood donation organisation (Cacioppo and Gardner, 1993; Cacioppo et al., 1997). This is in line with articles in which reasons why people avoid blood donations are referred to as barriers (Hupfer et al., 2005). Hence, we derive the last hypothesis:
H 4. The higher the individual barriers of a blood donor, the lower the blood donor loyalty.
Data collection and sample
To answer our two research questions, we analysed a dataset from the German Red Cross. The German Red Cross is recognised as being the most important nonprofit organisation in the country, which collects voluntary non-remunerated blood donations and holds approximately 70% of the market share. Additionally, the German Red Cross satisfies 70% of the German demand for blood products and relies on a broad donor base. The organisation regularly evaluates motives and barriers of blood donors as well as relationship factors such as blood donor satisfaction and blood donor loyalty. The evaluation is based on large-scale surveys, which are constantly conducted. This permanent evaluation of their activities, among other factors, indicates a high degree of donor orientation and marketing expertise. Therefore, the German Red Cross is regarded as an excellent source for providing reliable datasets on blood donors.
The dataset that we obtained from the German Red Cross was collected from first-time donors, classified as donors with a first blood donation within the last 12 months. In spring of 2009, 8000 randomly selected first-time donors were invited by letter, accompanied by an information leaflet and a self-administered questionnaire regarding participation in the survey. The survey was closed after a 2-month period (April to May 2009). Finally, 2899 first-time donors returned usable questionnaires resulting in a satisfactory return rate of 36.2% (which can be seen as good compared with other postal questionnaires in Germany; see Nienaber and Tietmeyer, 2010). Because of missing values of demographic data, we deleted 750 cases of our overall sample. Finally, our sample size comprised 2149 answers of first-time donors. Table1 shows the sample characteristics displaying age, gender, education and profession of the respondents.
Table 1. Sample characteristics (N = 2149)
Currently in school
Junior high school
High school graduation
A comparison of this sample with the basic population of German blood donors—regularly surveyed by the Robert-Koch Institute (Ritter et al., 2008)—reveals that the respondents' gender and age are different from the overall blood donor population. Specifically, 60.4% of the respondents are women (51.6% in the German blood donor population) and 39.6% are men. In terms of age, 41.0% of the sample are 18–24 years of age (51.7% in the German blood donor population), 17.7% are 25–34 years (18.3%), 21.2% are 35–44 years (16.1%), 16.0% are 45–54 years (10.6%) and 4.1% are 55–68 years (3.3%). Non-response bias was assessed by comparing the first 25% respondents and the final 25% respondents (Sargeant and Lee, 2004). The rationale is that the last 25% of respondents are more likely to resemble nonrespondents (Armstrong and Overton, 1977). The t-test indicates that early responders do not differ statistically (p > 0.05) from late respondents except to the items ‘competence’ and ‘donations next year’.
In order to measure our variables, we adopted those items of the German Red Cross questionnaire that fit to accepted scales used in the literature on blood donation (see the questions in Table2). A full adoption of accepted scales was not always possible. All measures used in this study are documented in Table 2.
Table 2. Operationalisation of blood donation marketing constructs
Scale; related literature
I donate blood because I want to help others.
5-point Likert scale (1 = strongly agree to 5 = strongly disagree); Glynn et al., 2006
Satisfaction with treatment
How satisfied have you been with the following aspects of your first blood donation?
5-point Likert scale (1 = very satisfied to 5 = not satisfied at all); Parasuraman et al., 1985 (SERVQUAL)
1. Cleanness of facilities.
3. Staff competence.
To what extent do you agree with the following statements regarding your first blood donation?
5-point Likert scale (1 = strongly agree to 5 = strongly disagree); Fernández-Montoya et al., 1998; Glynn et al., 2006
1. I was afraid of the blood withdrawal.
2. I had to overcome my convenience.
Blood donor loyalty
1. Would you recommend donating blood to acquaintances in your personal environment?
11-point scale (1 = sure to 11 = by no means); Swanson et al., 2007; Sargeant and Lee, 2004
2. How many times are you planning to donate next year?
4-point scale (>3; 2–3; 1; not at all); Sargeant and Woodliffe, 2007; Bennett, 2006
Generally, two measurement approaches for latent variables are possible (Diamantopoulos and Winklhofer, 2001). Whereas a formative measurement implies that indicators define the latent variable, a reflective measurement means that indicators are manifestations of the construct (Jarvis et al., 2003). In this study, the altruistic value of a blood donor is specified by one single reflective indicator (Rossiter, 2002) because other items to measure altruistic values such as ‘I enjoy helping others’ or ‘I believe that donating blood is a duty’ (see Glynn et al.2006) were not included in the questionnaire of the Red Cross.
To measure satisfaction with treatment, we used three formative indicators: satisfaction with the cleanliness of facilities, reception and staff competence. These indicators are commonly used in service marketing research (Parasuraman et al., 1985; Newman and Pyne, 1997; Daigneault and Blais, 2004). On the basis of the criteria of Jarvis et al. (2003), we specified these three items by a formative measurement approach because of the action-driven character of the questions. The appropriateness of a formative measurement approach is further indicated as the correlations of the three indicators are low (from 0.37 to 0.53) (Jarvis et al., 2003).
For the same reason, we measured barriers including two formative indicators: fear of the donation process and personal convenience. We identified these indicators in several medical studies on blood donation as highly relevant in explaining why individuals do not donate blood (Hupfer et al., 2005; Glynn et al., 2006; Ringwald et al., 2007).
The blood donor loyalty was specified by two reflective indicators, namely the intention to recommend blood donations (Sargeant and Lee, 2004; Swanson et al., 2007) and planned blood donation in the next year (Bennett, 2006; Sargeant and Woodliffe, 2007).
Two general methodological approaches—the covariance-based structural equation modelling and the PLS (Henseler et al., 2009; Hair et al., 2011)—could be applied to measure causal relationships with latent variables. We considered the latter to be the appropriate method for the empirical test of our hypotheses for three reasons. First, the PLS approach is nonparametric in nature and does not make any assumptions about the distribution of the data. As our dataset is not normally distributed, covariance-based structural equitation modelling is not applicable. Second, the PLS analysis is robust with different scale types, which is given in our study once. Third, this approach allows the measurement of reflective as well as formative measurement models (Hulland, 1999; Henseler et al., 2009). As our measurement incorporates formative as well as reflective variables measured by different scale types, the PLS approach is most appropriate for our analysis. We analysed our PLS model using the software SmartPLS 2.0 (SmartPLS; Hamburg; Germany) (Ringle et al., 2005). The absence of an established global goodness of fit criterion for evaluating PLS estimates affords use of a catalogue of nonparametric criteria to assess the model (Chin, 1998).
The assessment of a PLS model typically follows a two-step approach where first the measurement model and second the structural model are assessed (Hair et al., 2011). We applied this procedure and found the following results presented in Figure2.
Evaluation of the measurement model
Altruism, blood donor satisfaction and blood donor loyalty are specified by a reflective measurement approach and, therefore, are assessed here with regard to reliability and validity of the constructs (Hair et al., 2011). To evaluate construct reliability, the factor loadings should reach a threshold of at least 0.70 (Henseler et al., 2009). All loadings achieve the required threshold except for the factor loading of ‘donations next year’ (0.62). Furthermore, we tested internal consistency, which is ensured if the composite reliability accomplishes at least 0.70 (Henseler et al., 2009). Hair et al. (2011, p 145) considered that in exploratory studies, values from 0.60 to 0.70 are acceptable. This level is reached by all variables. Convergent validity is assured if the average variance extracted (AVE) achieves at least 0.5 (Fornell and Larcker, 1981). The AVEs all exceed 0.50 and, therefore, confirm satisfactory convergent validity (results available from authors upon request).
We analysed discriminant validity to guarantee that each construct is significantly different from other variables. We assessed the square root of the AVE of each reflective latent variable. This value is determined to be higher than the constructs' highest correlation with any other construct (Fornell and Larcker, 1981). Discriminant validity is evident as this condition is satisfied (results available from authors upon request).
The constructs satisfaction with treatment and barriers are measured by a formative specification. The estimated weights, documented in Figure 2, show the relative importance of each indicator. We determined the significance of the estimated weights by bootstrapping with 5000 iterations (Chin, 1998; Tenenhaus et al., 2005). At this point, it is important to note that significant results are easily achieved with such a sample size of 2149 cases. This fact is taken into consideration when interpreting our results later in the discussion section. As shown in Figure2, all weights (from 0.39 to 0.76) are at significant levels. To check whether indicators are redundant, we furthermore investigated the variance inflation factor (VIF) for each variable to test for multicollinearity. The VIF values lie below the required threshold of 10 (Henseler et al., 2009), suggesting that multicollinearity is not an issue in our study. Overall, we interpret our measurement model as valid and reliable.
Evaluation of the structural model
The structural model evaluates the relationships between the integrated blood giving mechanisms. The essential criteria for evaluating the structural model are the individual path coefficients (see Figure2 and Table3) and the coefficient of determination (R2) of the endogenous latent variables. The individual path coefficients of the PLS model could be interpreted as standardised beta coefficients of an ordinary least squares regression. The path coefficients can be used to evaluate the hypotheses. The results are shown in Table 3.
Table 3. Results of the structural model
Supported by the data
H1: Altruistic values → blood donor loyalty
H2: Satisfaction with treatment → blood donor loyalty
H3a: Altruistic values → satisfaction with treatment
H3b: Satisfaction with treatment → altruistic values
H4: Barriers → blood donor loyalty
Hypothesis 1 suggests that altruistic values are positively related to blood donor loyalty. The results in Table 3 show that the path coefficient of 0.19 with a t-value of 6.55 is statistically significant, supporting hypothesis 1. Furthermore, we found a significant path coefficient of 0.25 (t-value of 9.60) for the relationship between satisfaction with treatment and blood donor loyalty. Thus, hypothesis 2 is supported by the data.
This is the most interesting and, overall, important result of the study because the factor satisfaction with treatment is higher compared with altruistic values. In addition, we assessed the correlation between altruistic values and satisfaction with treatment. First, we applied a correlation analysis but find only weak correlations between the indicators of the two constructs with a highest value of 0.15. Then, we estimated the direct path between the latent factors for altruistic values and satisfaction with treatment in both directions. The path coefficient is for both directions 0.16 at significant levels, supporting hypothesis 3a and 3b. This result indicates that the relationship could be bi-directional.
Hypothesis 4 suggests that barriers have a negative impact on the loyalty of first-time donors. The weak path coefficient of −0.10 is in the hypothesised direction and is at a significant level with a t-value of 4.23. Therefore, we find support for hypothesis 4.
Beyond the path coefficients, we evaluated the R2 values, which determine the construct variance explained by the model. Chin (1998) recommends interpreting the R2 values as follows: values less than 0.19 are weak; up to 0.33, moderate; and 0.67 or higher, substantial. However, in disciplines such as consumer behaviour, R2 values of 0.20 are considered high (Hair et al., 2011). The R2 of the latent endogenous variable blood donor loyalty in our model shows that only 13% of its variance can be explained by the latent exogenous variables altruistic values, satisfaction with treatment and barriers. Consequently, the model should be expanded in future research projects with other mechanisms for blood giving to increase the quality of explanation for loyalty.
Implications, limitations and future research
The results of this study show that satisfaction with treatment has a stronger relationship on blood donor loyalty than altruistic values. Therefore, we recommend changing the research perspective and considering different mechanisms together. The following implications are of importance in regard to the findings of our first research question.
Blood donation services should focus more intensively on the acquisition and retention of blood donors by ensuring a high satisfaction with the treatment during the donation process. With regard to our model, the interaction between blood centre staff and blood donor was shown as important for gaining satisfied and loyal first-time donors. Therefore, blood donation services should enhance their service quality by identifying and correcting quality shortfalls. The reception and the competence of the service employees are considered to be the strongest indicators of a satisfying service experience. Additionally, securing hygienic and clean donation environments is of great importance for blood donation services. Moreover, well-educated and competent staff is needed to assure donors that their donation is safe, as well as to reduce potential incidents such as vasovagal reaction or bruises. Because of a more challenging environment, where some organisations pay for donations but most of the existing nonprofit blood donation organisations disapprove of payments, the service experience should be regarded as an important incentive against paying competitors. We call for further research that empirically tests different combinations of blood giving mechanisms to determine which is most appropriate in different settings.
In respect to research question 2, we found only a weak correlation between altruistic values and satisfaction with treatment and therefore conclude as a managerial implication that both mechanisms should be managed and prioritised in different situations. Blood donation centres could focus on altruistic values in media or advertising campaigns to acquire first-time donors. To retain first-time donors, however, the organisations should focus more on the management of satisfaction during the blood giving process. The implementation of a satisfaction measurement and management system could be an appropriate action in this regard. The low but significant direct effect of satisfaction with treatment on altruistic values furthermore implies that a segmentation of the donor base, for example in a cluster with high-satisfied or low-satisfied donors, could be a good way to better understand blood donors and their behaviour.
Our study exhibits several limitations that provide potential avenues for further research. First, our quality of explanation is low because we evaluated only two factors. Future research could expand the model or even try to test all mechanisms identified in our literature review in one comprehensive framework to learn more about the relationships of different blood giving mechanisms.
Second, we drew on a dataset collected by the German Red Cross. Therefore, the adopted items for measuring our constructs were limited to those indicators chosen by the German Red Cross to evaluate their marketing activities. For this reason, altruistic value is measured in our study by a single-item approach instead of measuring the construct by several items. Further research might refine the measures of the construct satisfaction with treatment and improve related quality criteria. Because of the given dataset by the German Red Cross, we could not include additional items controlling for acquiescence bias (Paulhus, 1991; Robinson et al., 1991). Therefore, results might be biased because some respondents tend to evaluate everything positively.
Third, we did not evaluate further moderating effects that weaken or strengthen the effects of mechanisms such as personal characteristics. In particular, the status that distinguishes between first-time donors and regular donors with a longer donation history is responsible for different perceptions of satisfaction with treatment (Ibrahim and Mobley, 1993; Fernández-Montoya et al., 1998; Misje et al., 2005). This indicates the need to treat these donor categories differently. Furthermore, the mix of different mechanisms needs to be examined more closely. For instance, it stands to reason that solicitation is present in settings, such as student campuses, where direct social pressure is more successful than in anonymous settings. Thus, further investigations on the mix of different mechanisms are important for future research.
The authors would like to thank the anonymous reviewers as well as the editors René Bekkers and Pamala Wiepking for their valuable comments and constructive feedback on an earlier version of this paper. We also acknowledge the support of the German Red Cross Blood Services (North and East), which provided the data of our study.
Silke Boenigk is a Professor of Business Administration, in particular Management of Public, Private & Nonprofit Organizations, at the University of Hamburg; Germany. Her research interests are in the areas of relationship fundraising, blood donor management, partnerships and measurement issues. In 2010, she founded a health care marketing research cooperation with the German Red Cross.
Sigrun Leipnitz is a research assistant at the Chair of Business Administration, in particular Management of Public, Private & Nonprofit Organizations, at the University of Hamburg. She is part of the health care marketing research cooperation with the German Red Cross. Her scientific work is focused on blood donation marketing and the analyses of factors that influence blood donors on their decision to donate blood.
Christian Scherhag is a research assistant at the Chair of Business Administration, in particular Management of Public, Private & Nonprofit Organizations, at the University of Hamburg. His current research interests include relationship fundraising, especially about the success of donor prioritisation on the fundraising performance, culture marketing and blood donation marketing.