Is attributing a behavior to some external cause or factor operating from outside the person?

1.2 Causal Attributions

Causal attributions, or beliefs regarding the causes of events, were the second major focus in The Psychology of Interpersonal Relations. Attribution theory was even more impactful than Heider's balance ideas, and became the dominant theme in social psychology for nearly fifteen years, between 1970–1985.

Heider postulated that people are motivated to understand and to master their environment—understanding is adaptive and instrumental to future behavior, and also there is a basic curiosity and desire ‘to know.’ Understanding and mastery require knowledge of the causes of events.

Pursuing the distinction in his thesis between thing versus medium, Heider stated that we are often only in contact with immediate facts or raw date (the ‘medium’) but we search for the underlying core processes or dispositional properties to explain these facts (the causal ‘things’). These perceived causes are generally enduring aspects of the world. Covariation information regarding the presence and absence of the inferred cause and the effect are important sources of causal knowledge. But when explaining the actions of persons, Heider reasoned that we often ascribe their behavior to some stable dispositional quality or trait, rather than to situational factors. The underestimation of the situation as a perceived cause of the behavior of others, and over-attribution to the person, subsequently was labeled ‘the fundamental attribution error.’

Although the topics of balance and attribution have been treated separately in the psychological literature, Heider saw them as intertwined. He reasoned that balance principles guide causal attributions. Thus, we infer that a good act was done by a good person, and ascribe success more than failure to the self (the ‘hedonic bias’), for in both of these examples the inferences result in balanced states or simple structures.

3.2 Attribution as Causal Judgment

The second major strand of attribution research was inspired by Kelley's (1967) model of causal attribution. Focusing on Heider's insight that causal judgments are pivotal in social perception, Kelley proposed that such judgments are based on a simple information-processing rule: people infer those causes that covary with the event in question. Specifically, when an agent A behaves toward object O, the cause of the behavior is perceived to lie in A (internal attribution) if few other people behave as A does (low consensus), if A behaves the same way toward O over time (high consistency), and if A behaves the same way toward other objects (low distinctiveness). By contrast, the cause of the behavior is perceived to lie in O (external attribution) if most other people behave as A does (high consensus), if A behaves the same way toward O over time (high consistency), and if A behaves differently toward other objects (high distinctiveness). For Kelley, the principles of covariation applied equally to behaviors (whether intentional or unintentional) and physical events. His model was therefore embraced by causal reasoning researchers in both social and cognitive psychology.

Empirical tests of the covariation model tended to support it, although this support was limited to experimental settings in which explicitly presented co-variation information had an effect on judgments. At the turn of the twenty-first century, no studies had demonstrated that people spontaneously, in everyday situations, seek out covariation information before answering a why question. Moreover, the covariation model makes no predictions about situations in which people lack covariation information (e.g., single observations) or in which they are not motivated to make use of the information (e.g., under time pressure). For those instances, Kelley later proposed additional rules of causal reasoning (‘causal schemata’). One of them is the discounting principle, which states that, under certain circumstances, a second cause weakens the plausibility of a first cause. For example, if a student aces a difficult exam, we might explain it by assuming that he studied hard; but upon hearing that he cheated we may no longer believe that he studied hard, thereby discounting the previous explanation.

The problem that Kelley's causal attribution models tried to address was the ‘causal selection problem’—how perceivers select particular causes for explaining a given behavior or event. Even though Kelley's models themselves did not completely solve this problem (alternative models were developed later), Kelley's work profoundly influenced attribution research by assuming that (a) people break down causes into internal ones (something about the agent) and external ones (something about the situation) and (b) that the internal–external dichotomy generally applies to all behaviors and events alike. This dichotomy proved to be a compellingly simple dependent variable that allowed researchers to explore a variety of interesting phenomena. One of these phenomena is the self-serving bias in explanations (see Miller and Ross 1975)—the tendency for people to explain their own positive and negative outcomes so as to maintain favorable self-perceptions or public impressions. For example, students would be expected to explain a good grade by citing internal causes (e.g., ability or hard work) and a bad grade by citing external causes (e.g., bad luck or an unreasonable teacher). Another important phenomenon is the actor–observer asymmetry, which is the tendency for people to explain their own behaviors and other people's behaviors in systematically different ways. Specifically, Jones and Nisbett (1972) argued, and later studies confirmed, that people tend to explain their own behaviors by reference to external factors (e.g., ‘I chose psychology as my major because it's interesting’) but explain other people's behavior by reference to internal attributes (e.g., ‘He chose psychology as his major because he wants to help people’).

Some researchers proposed alternatives to the internal–external dichotomy of causes. For example, Weiner introduced two additional distinctions—one between stable and unstable causes and one between controllable and uncontrollable causes—and thereby improved predictions for people's emotions and motivations in the wake of explaining achievement outcomes, health outcomes, and deviant social conduct (see Weiner 1995). Abramson et al. (1978) introduced another distinction, between global and specific causes, to describe a hopeless explanatory style in depression (see Depression). However, most attribution work at the beginning of the twenty-first century still uses a single internal–external (or sometimes trait–situation) distinction.

The focus on internal vs. external causal attributions to all behaviors and events alike can be contrasted with Heider's claim that ordinary people sharply distinguish between (a) attributing intentional actions to the actor's motivation and (b) attributing unintentional behaviors (e.g., failure, depression) to causal factors internal or external to the agent. Heider felt, as an interview in 1976 reveals, that people's attributions of events were adequately depicted in Weiner's model, but that attributions of actions to motives were inadequately treated by contemporary research (Ickes 1976). A second limitation of contemporary research is that it developed primarily cognitive models of the attribution process, whereas Heider demanded careful attention to the communicative functions of attributions and the language of causality. Theoretical developments of the 1990s tried to address both of these potential limitations.

3.1 Causal Attributions

Parents try to find explanations for why their children have behaved in a particular way, that is, they make ‘causal attributions.’ Although people make causal attributions in a variety of contexts, parents are probably particularly likely to make them because they need to understand their children so that they can effectively influence them. Dix and Grusec (1985) have outlined some of the features of parents' thinking in this context. In the search for explanation, parents can make internal or dispositional attributions that find the source of action in the child's personality or character. Alternatively, they can make external attributions that locate the source of action in the external situation or environment. When a negative action is attributed to dispositional factors it is most often seen as intentional and under the child's control. In this case parents have been shown to react punitively, possibly both because of their accompanying anger and because a Western ethical system dictates that intentional misdeeds should be punished. Parents who make external attributions, believing their child was tired or provoked or did not know any better, are likely to respond in a more benign way, e.g., by reasoning in an attempt to guide their child's future actions. When the attributions are accurate, parental behavior is likely to be appropriate for modifying the child's actions. When the attribution is inaccurate, however, parenting will be ineffective, given that children who lack knowledge, for example, are merely punished and do not learn what correct actions are, or children who have knowledge do not experience the negative consequences of their actions.

9.7.1 Causal attribution

Nisbett suggests that the attribution of causation differs across cultures. Westerners are inclined to attend to some focal object, analyzing its attributes and categorizing it in an effort to find out what rules govern its behavior. Causal attributions tend to focus exclusively on the object and are therefore often mistaken. East Asians are more likely to attend to a broader perceptual and conceptual field, noticing relationships and changes and grouping objects based on family resemblance rather than category membership. Causal attributions emphasize the context. Social factors are likely to be important in directing attention. East Asians (more than Westerners) report that from their perspective, they live in complex social networks with prescribed role relations. To East Asians, attention to context is important to effective functioning. More independent Westerners live in less constraining social worlds and have the luxury of attending to the object and their goals with respect to it.

This could mean that moderators from different cultures react to different events when observing a user research test and thus may report different user research problems. For instance, “Chinese people are inclined to attribute behavior to context and Americans tend to attribute the same behavior to the actor” (Nisbett, 2003, p. 114). One example reported by Nisbett showed that Americans were much more likely to attribute certain behavior to the presumed personality traits of the person; Indians emphasized that contextual factors attributed to the behavior.

Furthermore, it seems that westerners are more likely to embark on causal attribution to describe a few factors as being the reason behind some observed phenomena. If these effects hold also for user research evaluators, it may have implications. Most important, evaluators may differ to the extent to which they clearly identify a factor behind some observed difficulty; the number of problems that are attributed to users’ personality traits (e.g., being slow and inexperienced) may also differ.

2.3 Limits on Behavioral Measurement Constrain Causal Attribution

To many scientists and non-scientists alike, it may appear that behaviors are relatively simple and straightforward to measure, while neurochemistry, neuroanatomy, gene sequences, and gene expression patterns are vastly more complicated. It is frequently argued that causal attributions in neurogenetic work are limited by an ability to quantify genetic or neurobiological variation, or limitations in the statistical techniques used to relate behavioral variation to neurobiology and genetics. In reality, just the opposite is the case. Measures of neurobiological and genetic variables are typically subjected to closer critical scrutiny than behavioral measures. They are typically also direct measures. Since so many of the behaviors people study neurogenetically can only be measured via inferential linkages, the characteristics of the behavioral variables themselves are typically the factors which most constrain causal attributions.

One example is provided by work on rodent learning that uses the amount of time it takes animals to find a platform hidden under the surface of a deep pool of water as a behavioral variable. During some number of training sessions, individual animals can presumably use a constellation of spatial cues in the room surrounding the pool to memorize where the platform is. However, there are many other factors that will determine whether animals can utilize these training opportunities in the same way, and also determine the speed with which an animal familiarized with this situation will find the platform.

To take a hypothetical illustration, animals which are more prone to panic attacks may find being tossed into a pool of water quite stressful, and may spend a good deal of time swimming around the edges of the pool trying to escape. Thus, they may not have the same learning opportunities as calmer animals, and, even if this difference in opportunity to learn is controlled for, they may still be slower because it takes them a while to calm down enough and remember that there is a platform to be looked for. Other animals may have motor problems with coordinated swimming and be hesitant to leave the side of the pool or simply be slower swimmers. There are many reasonable scenarios by which animals whose genotypes differ may show reliable differences in this behavioral task, all unrelated to their spatial learning ability. Researchers must conduct numerous tests to rule out these alternative explanations before accepting that a neurobiological or genetic linkage is with learning ability rather than with stress, motor coordination, or a host of other possible explanations.

Because there is currently no systematic way of advancing reasonable alternative explanations, or of deciding how many potential alternatives there are, this open-ended process can take a long time. When one is finally sure of a linkage between a behavioral process and genetic and/or neurobiological variation, one must then make sure that individual differences in behavior can be sufficiently well quantified to enable the biological linkages between genes, neurobiology, and behavior to be adequately explored.

7 The Role of Cognition

Though we deal with empathic affect, cognition's importance is evident in two empathy arousal processes (verbally mediated association and perspective-taking): in the key role of self/other differentiation in empathy development; in the production of empathic anger and empathy-based guilt through causal attribution; and in children processing the information in inductions. Beyond that, cognitive development enables humans to form images, represent people and events, and imagine themselves in another's place; and because represented people and events can evoke affect, victims need not be present for empathy to be aroused in observers. Empathy can be aroused when observers imagine or read about victims, discuss or argue about economic or political issues that involve victims, or make moral judgments about hypothetical moral dilemmas (see Moral Reasoning in Psychology) in which there are victims. See Hoffman (1987, 2000) for discussion of the importance of cognition, including the cognitive dimension of moral principles, in empathy's development and prosocial functioning.

1.1 Evolution of the Field

Social work, psychiatry, and psychoanalysis in both Great Britain and the United States all made enormous contributions to the development of family therapy. In her influential book Social Diagnosis (1917), Mary Richmond was one of the first proponents of studying the whole family and its needs. At the same time, the dominance of the psychoanalytic model in the mental health field focused on the mother–child dyad as the source of psychopathology. This model, with its linear–causal attributions, blamed maternal deficiencies for any childhood disturbance. Adler became the first of Freud's pupils to openly challenge such an explanation of human behavior. He offered an alternative and more socially rooted theory of psychodynamics, organized child guidance clinics in Vienna, and counseled children, parents, and teachers.

By the 1930s, a strong community of analysts had developed in the United States. Among the most influential to emerge was Harry Stack Sullivan. He focused on interpersonal relations in his research into family dynamics of schizophrenia. By the time family therapy pioneers began to experiment in the 1950s, there was a well-established bias toward social explanations among American analysts. This trend was underscored by a paradigm shift that occurred with the development of General Systems Theory (GST). GST implies that family therapy is grounded in a set of assumptions about the interchange between individual, family and social processes, which operate according to certain principles that apply to all human systems (Bateson 1971). Family members are interrelated such that every individual affects each other, while the group as a whole affects the first member in a circular chain of influence, making every action in a sequence also a reaction.

Causality stability effects

In addition to attributions of locus of control, causality can be attributed to stable or unstable causes, where stability refers to changeability of the cause over time (Roesch & Weiner, 2001). Unstable causes are expected to be changeable over time, which evokes approach-centered coping, whereas stable causes evoke a sense of irreversibility, which triggers avoidance (Roesch & Weiner, 2001). Mental health conditions are susceptible to causal attributions that vary along the stable–unstable dimension because of the biogenetic and attitudinal causal explanations associated with mood disorders (Calhoun, Johnson, & Boardman, 1975; Deacon & Baird, 2009; Goldstein & Rosselli, 2003). Both biogenetic and mood-related causes are endogenous causes but differ on the dimension of stability, where attitudinal cause (e.g., negative thinking) is unstable as it is malleable, while biological cause (e.g., brain chemical imbalance) is perceived as stable (Calhoun et al., 1975; Lebowitz, Ahn, & Nolen-Hoeksema, 2013). The perception of controllability that endogenous causes evoke could be further enhanced when coupled with an unstable cause due to the malleability assigned to unstable causes. Consequently, endogenous-unstable cause of a mood disorder such as depression should trigger behavioral- and cognitive approach–oriented coping compared with endogenous-stable cause or exogenous cause.

Findings discussed so far examine how message features affect persuasive outcomes in engagement with the mental health games app, ReFocus. Given that such apps are delivered and used through devices that vary in their own features, it is important to explore how some of these channel features would affect the persuasive outcome (Oinas-Kukkonen & Harjumaa, 2018). The next section in this chapter discusses channel features, specifically screen size, in influencing engagement with the mental health app.

2.2 Influences on Other Aspects of Puberty

The example of exercise above is one in which effects have been seen on the adult shape that results from pubertal growth, in addition to delaying pubertal onset. Young ballet dancers who train many hours a day develop longer limbs—both arms and legs—compared with sisters and mothers. A similar effect has been observed among swimmers. While there surely are selection effects confounding the results for these elite athletes, the comparison with other family members provides some degree of control.

Although there are likely to be other examples of effects of behavior on pubertal outcomes, effects on gross aspects of physical development are easiest to attribute. Even here, though, there is the continuing confound of experience during development that makes causal attribution to pubertal change more difficult to establish. For example, different shapes surely result from different eating patterns (e.g., overeating versus undereating), but in this discussion of puberty the focus is on those outcomes that are pubertal rather than those that would occur at any point in development. As complicated as these issues become with physical attributes, they are even more complicated with psychological or social outcomes.

One effect of context on puberty should be mentioned. The increased life span and generally better health and nutrition of many people today have led to an earlier age at puberty, clearly among more developed nations in the world and increasingly among developing nations (Garn 1992). This changes the context for puberty in that it occurs in otherwise less mature organisms. Some had thought that all aspects of development might proceed in concert, so that earlier puberty would be accompanied by earlier cognitive, psychological, and social development. This does not appear to be the case, probably because of the experience or training (sometimes called socialization) needed to become a mature human. Given the existence of psychologically or socially immature adults, we should not be surprised about physically mature but otherwise immature adolescents.

The fact that puberty occurs a few years earlier now than even in the mid-1900s, and is now closer to 10 years of age among girls than to 20 years as was more typical around 1900, has tremendous significance for our understanding of and also the experience of childhood. Biological maturity permitting procreation occurs at ages much earlier than those typically considered appropriate for taking on adult work and family roles. Most societies have not adjusted to this disjunction, either to protect children from becoming parents too early or to change socialization practices and social institutions to support earlier entry into adult roles. Instead, some societies have blamed the victim in this situation, suggesting that early childbearing represents a moral failure on the part of the child or their families.

2 Research About Perceived Control

Perceived control and related constructs have a long-standing history in the psychology of motivation and self-regulation. Five major conceptions have guided research in this area: locus of control (Lefcourt 1981, Rotter 1990) (see Control Beliefs: Health Perspectives), causal attributions (Weiner 1985) (see Motivation and Actions, Psychology of), learned helplessness (Seligman 1975, Abramson et al. 1978) (see Control Beliefs: Health Perspectives), self-efficacy (Bandura 1977) (see Control Beliefs: Health Perspectives), and perceived own intelligence as fixed vs. malleable (Dweck and Elliott 1983). Locus of control refers to the issue whether an event is controlled or produced by factors inside or outside the individual. Causal attributions refer to the retrospective explanation of outcomes of behavior by internal factors such as effort or ability and external factors such as task difficulty and luck. Learned helplessness focuses on the issue of whether events are contingent on the organism's behavior, and further distinguishes between stable and unstable and specific and unspecific causes of events. Finally, the distinction between fixed vs. malleable intelligence addresses inter-individual differences in the tendency to view intelligence as either an unchangeable entity or as a modifiable resource for performance. All these characteristics addressed by the five theories have been found to hold important implications for behavior and the evaluation of outcomes (Weiner 1986, Heckhausen 1991, Skinner 1995, Dweck 1999). Depending on the outcome, internal, stable, and general control beliefs are conducive, in the case of success, or detrimental, in the case of failure, to effort investment, ambitious task choice, persistence, and improvement in performance. Moreover, the temporal direction of control perceptions is critical, with prospective expectations of control influencing goal choice, and persistence, and retrospective attributions of control guiding the way in which success and failure are interpreted as indicative of future control potential.

Recently, researchers in this field have called for conceptual and domain-specific distinctions to be discussed in more detail here (for more detailed information about research traditions, see Control Beliefs: Health Perspectives; Motivation and Actions, Psychology of). Perceptions about personal control involve two components, ‘agents of control, and means of control’ (Skinner et al. 1988, Skinner 1996, p. 552). These two components are often not sufficiently differentiated, but confounded in a summary construct of perceived control. First, control beliefs imply conceptions about the controllability and the relevant means to attain an outcome. This component of control beliefs is captured by various closely related concepts (see review in Skinner 1996): contingency judgment; means–ends beliefs, and strategy beliefs. Second, control beliefs involve conceptions about the agent's resources for control in terms of having access to means for bringing about an outcome. This component of control beliefs is referred to as competence judgment, agency beliefs, and capacity beliefs.

Research about perceived control falls into two broad areas: (a) normative age changes in perceptions of control across the lifespan, and (2) the association of perceived control and various functional outcomes, such as task-performance, school grades, subjective well-being, and health (see also review in Schulz and Heckhausen 1999).

Research on childhood development of perceived control has shown that control beliefs become increasingly realistic, thus moving away from illusory gross overestimations of control in early childhood (Weisz 1983). However, the extent to which perceptions of personal control become realistic in mid-childhood, adolescence, and adulthood is a function of cultural variation (Little et al. 1995). School children in the United States, for instance, hold beliefs about their competence to attain good grades (i.e., agency beliefs), which have little relation to their actual school grades, whereas children in the former German Democratic Republic reported agency beliefs which closely resembled their actual school achievements.

In addition, developmental research has revealed that the agency and the means–ends components of control beliefs, as well as beliefs about different causal factors (e.g., ability, effort, luck, teacher) were increasingly differentiated during mid-childhood and exhibit unique developmental trajectories and relations to outcomes. Skinner et al. (1998), for instance, showed that during mid-childhood, cycles of influences between perceived personal control and school achievement were focused on different components at different ages. Beliefs about effort dominated the cycles in younger school children, whereas beliefs about ability were the focus of the regulatory cycles in older children (see Self-regulation in Childhood).

As to normative age changes during adulthood the findings to date provide a mixed picture (see review by Schulz and Heckhausen 1999). While earlier reviews of the literature reported little evidence for age-related changes in generalized measures of control (Lachman 1986a), more recent research has identified decreases in perceived control across adulthood (e.g., Mirowsky 1995). However, Lachman was the first to emphasize and provide evidence for domain specific perceptions of control and their age-related trajectories (Lachman 1986b). Lachman's research shows that older adults report decreasing perceptions of control for those domains for which their more constraint social roles and declining physical fitness yield deteriorated opportunities and resources for control.

The association of perceived control and health has been studied extensively using field experimental paradigms involving interventions improving objective control opportunities (e.g., Langer and Rodin 1976, Schulz and Hanusa 1978, Baltes and Baltes 1986), quasi-experimental methods exploiting naturally occurring environmental changes of control potential (e.g., Timko and Moos 1989), and large sample surveys (e.g., Krause 1987, Menec and Chipperfield 1997). Findings from all these studies, research paradigms, and assessments of perceived control and health outcomes converge to show that having high perceptions of control benefits one's health, and having low perceptions of control can be damaging (see also review in Schulz and Heckhausen 1999) (see Control Beliefs: Health Perspectives; Health: Self-regulation; Self-regulation in Adulthood).