Studies of social support initially focused on social environmental factors influencing personal health, meaning to extend the discussion on individual mental disturbances to society and public topics. In order to explore the relationship between social support and personal physical and mental health, we should systematically define the concept of social support, analyze its operations and study its source structures and types comprehensively so as to measure its different influences on personal health and even on social integration. Scholars have been increasingly drawn to study of social support. In psychology, sociology, psychotherapy and social work, areas in which study of social support is conducted, the application field of social support has been expanded. The purpose of this article is to summarize the basic concept and development of study of social support. We will start the discussion by briefly introducing the definition of social support, then explain its source structure, relationship and structure, dimensionalities, mechanism of action, and at last introduce tools for its measurement.

1 Definition of Social Support

Different researchers have different understandings of social support. Some describe it as emotional love and respect (Cobb 1976). Others distinguish different types of support, such as emotional, practical and propositional (Kahn and Antonucci 1980). However, although these definitions encompass a number of relational sources and functional types, they fail to establish the consistent characteristics of social support. Many researchers criticize the lack of systematic definitions in previous study of social support (House et al. 1988) as well as the lack of consensus on the various kinds of definitions of social support (Dean and Lin 1977). Some of them even jokingly think that those who propose valid definitions of social support should be awarded with a big prize for their work in terms of concept construction of social sciences so as to highlight its variability and uncertainty (Gottlieb 1985).

In earlier studies, social support was seen as equal to perceived support. From his psychological perspective, Cobb (1976) believed that social support refers to three kinds of information influencing individual mental states. The first is emotional support, or letting individuals feel that they are cared and loved; the second is respect, or letting individuals feel that they are respected and are valuable; and the third is membership, letting individuals believe that they belong to some kind of mutually beneficial communication network, which further encompasses three pieces of information, namely what is the source of these relational networks, what kinds of goods and services they provide, and how they jointly resist risks. Such a definition of social support is grounded in psychology and affirms that what really matters is how individuals perceive influence, namely the effect of perceived support (rather than the influence of reality) on individuals.

In the early 1980s, the concept of social network analysis made its entry in study of social support, and many researchers in this field have used the concept and its related measurement method to analyze traditional social support issues. However, concepts like that of social network, social support, social connections and social contact were not clearly distinguished; most of them refer to several aspects of factors such as existence, quantity and contact frequency of social connections (House et al. 1988). Some researchers have mentioned that early attempts to introduce the social network analysis methodology into study of social support research confused, to a certain extent, the concept of social network with that of social support. Specifically, variables in the social network structure were used to measure social support (Lin et al. 1985; Lin and Ensel 1989). Researchers now realize that, conceptually and empirically, social network and social support are two different phenomena: social network specifically refers to the structural features of social connections while social support refers to the emotions and practical properties of social connections.

House et al. (1988) considered “social connections” as a general concept encompassing three sub-concepts, i.e. social integration, social network and connection content (the latter includes social support). Among these, social support specifically refers to the positive and potential effects of connection content in terms of healthcare or pressure relief. According to the above definition, House et al. made a clear distinction between the two concepts of network structure and social support. Substantially they defined social support with a focus on the positive functions of social connections, believing that social support was a kind of process related to social connections (House et al. 1988).

There are many other similar cases in which social support has been defined with a focus on functionality. For example, some believe that social support refers to interpersonal interactions, i.e. people helping one another (Antonucci and Knipscheer 1990). The advantage of defining social support as the positive function of connections is that it narrows the concept of social support, making it clearer. However, such functional definition does not give due consideration to the negative effect of social connections. Some researchers affirm that the effects of social support should be considered from both positive and negative perspectives. We will further discuss this topic later.

Some researchers have defined social support from the perspective of social resources. Social support is regarded as a kind of resource possessed by individuals helping them solve problems in times of crisis. Lin Nan clearly defined social support as a component of social resources used by individuals to meet expressive or practical demands (Lin 1986).

The above summary only covers a few of the numerous definitions of social support. Some of them define social support as covering almost all aspects of social connections (Gottlieb 1983; Turner et al. 1983). Turner summarized and classified the many definitions of social support that different researchers have come up with, including social bonds, social network, meaningful social contacts, availability of confidants, and others’ companionship. Turner (1999) pointed out that despite their different denominations, these kinds of social support all refer to stable social communication between people.

2 Sources and Structure of Social Support

2.1 Sources of Social Support–Support Structure

Different perspectives result in different levels of understanding. Nonetheless, in the case of social support, most early researchers paid attention to the limited close relationships of individuals. It has been discovered that spouses and family members are the key support providers protecting personal health (Weiss 1974). Here, we may use the term primary relation, which refers to the support providers with which individuals have the most contact and on which they are most likely to have an influence. After the measurement and analysis tools of social network made their entry in the field of social support, researchers started to focus on a more extensive range in terms of the source of support. Pearlin pointed out that there are three sources of social support: social networks, active contacts and intimate relationships (Pearlin 1985). The first source, social networks, was considered by Pearlin as the sum of all social connections of individuals. These social networks define the outermost boundaries of available support. The second source, active contacts, specifically refers to active social connections amid individuals. If social networks refer to all available sources of support, active contacts are the sources likely to be pursued by individuals. The third source is intimate relationships, which stresses degrees of relational intimacy in terms of emotion and trust. Similarly to Cooley’s primary relation concept (1915), this refers to the most important, unpractical and continuous relations for individuals, such as spouses and very close friends. When it comes to this third source, people are usually able to obtain support without requesting it. Pearlin (1985) believed that these three different sources were not competitive and not mutually exclusive, each of them imbedded in each other and explaining the source and characteristics of social support.

Some researchers think that the sources of social support should extend to whole communities in which individuals live. Lin and his colleagues classified the three levels of support sources, namely community relation, social network and intimate relationship, affirming that these three levels represented the outer layer, middle layer and inner layer of social connections, each of which have different influences on health (Lin et al. 1999). It is most important to consider that, theoretically, each level can independently provide individuals with social support; nevertheless the author also points out that support from inner layers is limited by the structure of outer layers (Lin 2001).

Similar divisions of support sources include Kapferer’s three sections of individual networks (1973), namely extensional section, effective section and intimate section. Previous studies defended the view that the three sections are differently sized: the extensional section encompasses a great deal of individuals; the effective section encompasses about 10 people; and the intimate section encompasses about 3. Wellman discovered intimate bonds are usually limited to 3–6; important relational bonds to 5–15; whereas acquaintances and potential relationships may reach up to 1000 (Wellman and Gulia 1999). The above divisions of social support are not only significant in terms of clearly defining the concept. In actual studies, the differences between support provided by different sources and their effect should be clearly stated. Let us first try to understand the different kinds of social support that different relational bonds provide.

2.2 Types of Relational Bonds and Social Support

Different network structures and relational characteristics result in different types of social support. Two aspects need to be considered here. First, different types of support provided by different relationships in the network should be analyzed (such as in Wellman and Wortley 1989). Typical studies focus on the correlation between the strength of a relational bond and the type of support it provides (such as in Wellman and Wortley 1990; Lin 1982 1986), or between the mode and frequency of social communication and the type of support provided (such as in Wellman and Wortley 1990). Second, differences between types of support provided by different network structures should also be analyzed. In this section, we will discuss both aspects.

Researchers pointed out that different types of bonds provide different types of support, mentioning also that not all relational bonds actually provide support. In concrete analysis, some researchers aimed at the following major relational dimensionalities: strength, contact, interaction situation, kinship and personal characteristic. When comparing these relational characteristics and the types of support they provide, it can be found that most relations provide a specialized type of support. The link between relational characteristics and social support is closer than that between personal characteristics and social support (Lin et al. 1986; Wellman and Wortley 1990; Wellman 1992). Only a few relational bonds can provide support such as emotional support and companionship. Familial bonds or intimate kinship are the core providers of emotional and practical support (Antonucci 1990; Walker et al. 1993; Wellman and Wortley 1990).

However, the characteristics of single relational bonds aren’t the only factor to consider when analyzing the differences between different types of social support. The influence of network structures on the provisions of social support should not be neglected either (Gottlieb 1985). Researchers criticized the fact that previous social support studies only focused on support provided by single relational bonds in the network, defining social support as the existence, quantity and frequency of social connections, yet neglecting to involve the structural factors of social networks, such as the scale and density, (House et al. 1988). It has thus been pointed out that further analysis of the influence of the social network structure on the degree to which individuals apply social support is required (Hurlbert et al. 2000). If the influence of the social network structure on social support is great, then previous studies on social support that only focused on relational bonds might have ignored some significant issues (Turner and Marino 1994).

Social support researchers have gradually realized that giving consideration only to a single or to certain relational bonds isn’t enough to understand the effect of social support, thus the influence of the overall characteristics of the networks on the social support they provide should be analyzed. First, Wellman et al. considered the overall structural characteristics of individual community networks, such as amplitude, density and availability, and discovered that amplitude (which encompasses scale and diversity) is most important in determining the existence of social support. Moreover, the larger the scale of a given individual network is, the more varied the types of support it can provide. Second, they discovered that spatial connections are also important in providing support. In this sense, closely connected kinship networks make individual needs clearer and provide support more effectively. Besides, researchers believe that the influence of the above three characteristics of individual community networks on support can’t be reduced to the sum of the influence of relational bonds. This can be understood as follows: “Seen from the provision of social support, social networks exceed the sum of the relational bonds composing them.” (Wellman and Gulia 1999).

3 Dimensions of Social Support

As stated in Sect. 1, Cobb and House et al. proceeded to research from the perspective of community mentality, and mainly focused on the effect of individual sense of support on health. Such research affirms that what matters to individuals isn’t what support contends, but rather what support implies for individuals. In other words, what matters to individuals isn’t the support they actually acquire, but the support they perceive. Later developments in research gradually included the concept of multi-dimensional social support, which can roughly be divided into that of subjectively-perceived support, actually-acquired support and support structure (Sarason et al. 1990).

Barrera and Ainley (1983) held a similar opinion, affirming that the classifications of social support were multi-dimensional. In their analysis, social support was classified into three kinds: embeddedness, enacted support and perceived social support. The above classification, although it uses different denominations, basically pointing to the above three aspects of multi-dimensional social support, namely perceived support, actually-acquired support and support structure.

Distinguishing between these three aspects can help clarify the underlying analytic concept, and furthermore encourage a discussion on the inner connections of perceived support, actually-acquired support and support structure as well as a survey of the significance of different layers of support for individual health. First, we discuss the connection between support structure and process characteristic. Social support on the structural layer is closely connected with individual perceived support and actually-acquired support. Individual social networks are formed through the influence of structural constraints and personal selection. On the one hand, network structures and relational characteristics can influence the factors for individual perceived support and actually-acquired support; on the other hand, if an individual perceives that he/she lacks support or acquires less support than what he/she perceives in reality, he/she may adjust his/her existing social network, thus changing its structure and content.

Before distinguishing whether actually-acquired support and perceived support are of different significance to individuals, it should be first determined whether there is a correlation between the two. Generally, actually-acquired support and perceived support are not consistent. Two cases can be observed: one where support doesn’t exist in reality but is believed to exist, and the other where it exists in reality but is not believed to exist. It has been discovered that the correlation between perceived support and actually-acquired support isn’t remarkable (Mancini and Bliezner 1989). Still, among theoretical explanations of perceived support, there is no lack of influence of various factors including actually acquired support on perceived support. These influences mainly include that of earlier experiences from the perspective of psychology, the individual capacity factor, and the influence on the experience of actually acquired support.

There are two theoretical models giving specific explanations when it comes to the influence of earlier experiences on perceived support. The first is the cognitive theory model, which affirms that the cognitive structure of support developed in an individual’s early years influences how he/she treats his/her surroundings; in other words, this model affirms that an individual’s cognition of support is the accumulation of a series of cognitions. The second is the attachment model, which affirms that attachment formed in an individual’s early years can influence the development of his perception of support in the future (Procidano and Smith 1997; Faber and Wasserman 2002), and that the degree to which adults with different types of attachment types perceive support varies a lot (Ognibene and Collins 1997). Both models lay emphasis on the effect of earlier support experiences on perceived support. Moreover, it has been discovered through many empirical studies that individual factors and actually-acquired support have a combined effect on perceived support and that differences in an individual’s capacity and personality can cause differences in how he/she perceives support (Sarason et al. 1990; Sarason et al. 1986).

Figure 1 captures how social support works as described in the foregoing analysis:

Fig. 1
figure 1

Relationship among the different dimensions of social support

Full size image

Many empirical studies on social support have led researchers to believe that social support is a multi-dimensional concept, whose different dimensions have different significance for individual health and behaviors (Dean and Lin 1977; House 1981). The discussion about social support in the previous section focused on the structural dimensionality of social support and its connection with the effect of support. In the next two sections, we will see that the analysis of the influence of social support on individual behaviors or health should consider the different effects of actually-acquired support and perceived support as well as the differences brought by pressure on the effect of support effect.

Are there any differences between actually-acquired support and perceived support in terms of their influence on health? It has been proved by empirical studies related to perceived support that actually-acquired support is closely related to health (Antonucci and Israel 1986; Sandler and Barrera 1984; Wethington and Kessler 1986). In comparing the significance of actually-acquired support and perceived support for individual health, many have discovered that perceived support is of greater significance for mental health than actually-acquired support, whereas actually-acquired support usually exerts its influence on individual health via perceived support (Sarason et al. 1990; Turner and Marino 1994; Thoits 1995; Wethington and Kessler 1986). Therefore Pearlin insisted on his idea that the influence of social support on individuals was exerted via “perceived support” (1989). In addition, researchers have made a distinction between the perceived adequacy and the perceived availability of perceived social support according to specific situations, and thus have discovered that adequacy is of greater significance than availability to mental health (Haines et al. 2002).

When exploring the effects of perceived support and actually-acquired support, researchers should also pay attention to different types of support. The differences between emotional support and practical support may reciprocally influence the action mechanism of perceived support or that of actually-acquired support. Quite a few researchers have pointed out that emotional support is of primary significance for health (Lin et al. 1985; Lin and Ensel 1989); while other researchers have discovered that practical support may also play a vital role in specific situations (Beggs et al. 1996; Haines and Hurlbert 1992; Wellman and Wortley 1990). Results from different studies are not completely consistent, and so far it hasn’t been proved that emotional support is of greater significance than practical support for health (Troits 1995; Lin et al. 2000). As a matter of fact, a great number of social support studies remain unable to confirm the relational characteristics between different dimensionalities or between these dimensionalities and individual health. Moreover, the mechanism in which support plays its role is also limited by different environmental conditions. The next section will discuss the different action mechanisms of social support under the conditions with and without pressure on individual health.

4 Direct Effect and Buffer Effect

Many studies have pointed out to the significance of social support for mental health. However, the way in which social support helps mental health has been disputed on the basis of two aspects, namely the direct effect and buffer effect of social support. Earlier studies mostly affirm that social support plays its role indirectly as a protection mechanism; it can buffer pressure, and helps people cope with pressure (Pearlin 1989).

Studies on the mediation function of social support as a “buffer” have concentrated on analyzing the pressure given individuals faced and these given individuals’ health. Many earlier studies have verified that social support is helpful in buffering individual pressure (Caplan 1974; Cassel 1976; Cobb 1976) and researchers have discovered the remarkable relationship between the acquisition of social support and frustration, proving that social support can reduce pressure (Dean and Lin 1977). Kaplan et al. even believed that social support only had positive health effects in the pressurized environment (Kaplan et al. 1977). There are few studies on the direct effect of social support on individual health during this period. Researchers who supported the buffer effect even believed that people were in fact always under different levels of pressure, which is always reduced by social support in a process beneficial to both body and mind, though this is sometimes hard to measure (Antonovsky 1979).

In subsequent studies, Turner and his colleagues discovered while analyzing young mothers under pressure that even controlled degrees of pressure do not dampen the obvious relationship of social support with mental health (Turner 1981; Turner and Noh 1983). At the same time, Turner and Noh discovered that the effect of social support on health changes along with changes in degrees of pressure, although it may also be under the influence of classes or other individual factors. Turner and Noh further distinguished different classes, the result of which distinction showed that young lower-classes mothers see the effect of social support change along with the changes in degrees of pressure and perceived support has an obvious effect when under great pressure; whereas the effect of social support on middle-class mothers is not so influenced. Almost simultaneously, other researchers discovered the usual coexistence of the direct effect and the buffer effect of social support on health (Cohen and Wills 1985; House and Kahn 1985). It is not necessary to choose between one or the other; what does matter is that the specific support mechanisms coming into play here should be explored, as well as at what time and under what conditions they do so (House et al. 1988).

Thus, researchers assume that individuals can feel the buffer effect of social support more easily when under specific pressuring conditions or when answering questions related to pressure. The pressurized environment individuals face influences the action mechanism of support, and so may measurement tools. Some researchers have been able to further study the buffer effect of social support while measuring the availability of perceived support; on the contrary, its direct effect is most obviously when actually-acquired support is measured (Cohen and Wills 1985; Kessler and McLeod 1985). In terms of its effect, there are two aspects to social support (referring particularly to perceived support): it can serve as a resource to cope with pressure in specific times; and in general it has a direct effect on the development of individual social mentality. The two action mechanisms of social support each correspond to its short-term and long-term effect on individual health.

5 Measuring Social Support

In earlier community psychology-oriented social support studies, several specific support sources selected by researchers according to specific realities were the measurement tools most employed for simple measuring (Kessler and McLeod 1985). This method was used, for example, in studies on pregnant women who accept social support from relatives and friends, as well as in studies on the social support acquired by married men. However the range of application of these measuring tools was limited and they were not extensively applied. The subjects to be measured stemmed directly from the actual demands and restrictions of research. Therefore, there was little comprehensive and systematic evaluation of the reliability and validity of these measurement tools.

Since the 1980s, researchers have attempted to develop multi-dimensional and extensively applicable measurement tools for social support. These tools no longer aim at the specific psychological issues of specific groups, but are applicable to ordinary questionnaire survey interviewees. In this case, type measurement and role classification are basically the same despite the multi-angular definitions of social support. Classifications may vary from as few as a couple of types to as many as dozens of types, most of which follow the standards of individual demand categories. For example, the support system can be classified as such: (1) emotional support, (2) help in completion of work, and (3) advice and material help; or as such: (1) consultation and advice, (2) companionship, (3) support for important affairs, (4) support for general affairs, (5) financial support and (6) emotional communication, etc. There are nonetheless different opinions on the classification of several types of support, such as that of “advice and companionship”, which contains both a practical purpose and an emotional connection and therefore is in fact a mixed type of support. In general, researchers agree to roughly classify different types of support into the categories of emotional support, instrumental support and mixed support. These have been widely measured in trans-cultural studies (Hollinger and Haller 1990; Ruan et al. 1997; Lee et al. 2005). In terms of the classification of the different roles of different support sources, most of the measurement tools above encompass the following categories: family members, relatives, good friends, general friends, colleagues, neighbors, professionals, community organizations, governmental agencies, etc. In this discussion, we will refrain from exploring the subject in further details.

It should be noted that, besides the need for the integration and normalization of existing tools, the measurement of social support should also correspond to its theoretical development. On the one hand, researchers should develop on the theoretical architecture of social support by summarizing the results of empirical studies. For example, Lin et al. (2000) summarized the great achievements of study of social support and raised the analytical framework on the structure and process of social support. They believed that the structural characteristics of social support lay in the support networks while also affirming that perceived support and actually-acquired support are process characteristics. The process characteristics of social support can further be classified into three dichotomies: perceived support and actually-acquired support; support in times of crisis and support in ordinary times; and emotional support and practical support. On the other hand, we should also apply the issues raised during theoretical discussions to empirical studies. For example, issues like the introduction of a negative function index as a measurement tool for social support are to be further discussed in the future.