Introduction

Locus of control (LOC) is defined as the extent to which the individual judges outcomes to be contingent on his or her behavior.1 Health locus of control (HLC) is the degree to which the individual believes health outcomes are the result of his own actions (internal LOC), luck or chance, or the influence of significant other people (external LOC).

A study of LOC in an spinal cord injury (SCI) sample indicated that persons with recent injuries manifested a more external LOC than an age, gender and education matched able-bodied control group.2 The researchers reported that 39% of the SCI sample had an external score on the LOC of behavior scale,3 as compared with 10% of the control group. A follow-up study of a number of the original participants showed by contrast to the 12-month data that there was only a trend for the SCI group to be more externally focused on LOC of behavior scores at 24 months.4 In an additional paper on the 24-month data, regression analysis found the experience of pain 2 years after injury and feeling out of control of one's life before hospital discharge were predictive of depression 2 years after injury.5

Cluster analysis using the Levenson Internal External Scale,6 showed that the first of three groups who attributed control more externally to chance had significantly higher depression and role dissatisfaction and lower life satisfaction. Those in the other two clusters had more internal attributions or a mix of internal and external attributions and they had better outcomes.7 Correlations between types of LOC and outcomes from studies cited here appear in Table 1.

Table 1 Studies of spinal cord injury and locus of control
Full size table

It appears that individuals with an internal LOC in the sense of a general attributional style seem to be better adjusted than those with an external style. Health-specific measures of LOC have also been researched. Cluster analysis with the Ways of Coping Scale,8 showed that the first of the two groups showing higher self-blame and wishful thinking on the Ways of Coping Scale were less well adjusted.9 People falling into that group endorsed the Multidimensional Health Locus of Control scale (MHLC),10 evenly in terms of internality and externality and reported higher levels of depression and distress. By contrast, the second group, who were better adjusted, relied more on internal attributions than on external ones on the MHLC.

In a study using multiple regression with the MHLC, internality was positively correlated with overall satisfaction (r=0.36) whereas chance was negatively correlated with satisfaction (r=−0.22). In addition, internality was positively correlated with subjective well-being and powerful others were negatively correlated with health indicators.11

A similar study using the MHLC scale indicated internality was positively related to purpose in life, and that purpose in life was related to adjustment. Internal HLC was also related to good adjustment.12 Chance MHLC, was positively related to the negative outcome of neurosis and anxiety (see Table 1).

A clear body of research exists to show that for either general or health-specific measures of LOC, internality is correlated with better outcomes and externality in its various forms is associated with pathology. Seeing ones world as being amenable to change on the basis of one's own actions is evidence of a sense of personal control, and is healthy in our lives. A question remains around the relationship between personal adjustment and blame attributed either internally or externally for the accident that caused the SCI.

In an early study (1977), participants rated (on a 0–5 scale) the extent to which they blamed themselves, and the extent to which they believed they could have avoided what happened.13 Participants also assigned a percentage of blame to themselves, others, environment and chance. These ratings were combined into a single self-blame score. Staff members rated the coping of the participants in the study. A multiple regression suggested that blaming oneself was paradoxically a successful predictor of staff ratings of good coping. Endorsement of self-implicating avoidability, or blaming another, was related to staff ratings of poor coping. Individuals were more likely to blame themselves if they believed they could have avoided the accident.

In a subsequent 1988 study,14 the same procedure for creating a composite self-blame score was used. A comparison was drawn between those who attributed up to 10% of the blame internally (low self-blame group) and those who attributed 11% or more of the blame internally (high self-blame group). In contrast to the 1977 study, analysis found that self-blame correlated with poor coping (see Table 2). However, individuals were again more likely to blame themselves if they believed they could have avoided the accident.14

Table 2 Studies on spinal cord injury and attributions of cause and blame
Full size table

A European 1988 study compared a group of people injured for <18 months with a group injured for more than 24 months.15 Participants rated their level of concern with attributions of cause and avoidability. For people in the recently injured group, self-rated concern with avoidability correlated with medical staff ratings of poor coping (r=0.85). However, there was little correlation with the rating of coping from the psychologist (r=0.17). For the longer injured group, self-rated concern with avoidability was correlated with good coping as rated by a psychologist (r=0.79), yet, there was little correlation with medical staff ratings (r=0.19). Concern with cause was not significantly correlated with any ratings of coping, and blaming oneself was not addressed in the study. Several previous studies rely on ratings of coping by staff members rather than examining the self-rated coping of participants. This may introduce additional uncontrolled factors or biases and is perhaps a more questionable methodology than self-report.

A 1996 study,16 distinguished between attributions of cause and attributions of blame. The researchers investigated causality using a set of probes that did not use the word blame. Items such as ‘My own behavior caused my accident’ were rated on a 1–7 scale. Avoidability was addressed by rating the statement ‘I feel I could have avoided my accident and therefore my SCI’. Items relating to blame but not using the terms ‘cause’ or ‘caused’ such as ‘I am to blame for my accident’ were also rated. Results again showed a correlation between blaming oneself and self-avoidability (r=0.54). This was of the same order as the correlation between blaming oneself and perceiving self-causation (r=0.66). This 1996 study began to separate out some of the confounding variables of earlier research yet suffered for not addressing the link between adjustment and self-rated attributions of cause and blame. Correlations for attributions of cause and blame and outcomes from studies cited here are listed in Table 2.

It is noteworthy that to a large degree, and unlike studies of LOC, research on attributions of cause or blame in SCI has failed to reach a consensus on several questions. These include whether self-blame or particular attributions for one's own role in causing or avoiding the SCI are adaptive.

Materials and methods

Participants

All participants (N=30) were wheelchair users and had a traumatic SCI. Injuries were the result of a specifiable acute onset rather than disease or congenital factors. Ten people had sustained a complete injury and 20 people had sustained an incomplete injury. None of the participants had suffered either moderate or severe head injuries. There were 25 men and 5 women. The sample had an average age of 45.37 (range 24–63). They had an average time since injury of 11.63 years (range 2–30). They had an average age at injury of 33.73 years (range 12–58).

Instruments

The SCL-90-R,17 is a self-report inventory that screens for a range of psychopathology. Ninety items are rated on a five-point distress scale ranging from 0 (not at all) to 4 (extremely). There are nine symptom dimensions; somatization, obsessive-compulsive, interpersonal sensitivity, depression, anxiety, hostility, phobic anxiety, paranoid ideation and psychoticism. Test–retest reliability for 1 week ranges from 0.78 to 0.90 and internal reliability ranges from 0.77 to 0.90. The SCL-90-R was used to assign participants to one of two adjustment groups. If the respondent has a Global Severity Index score (on Norm B, the non-patient norm) ⩾T score of 63, or if any two primary dimension scores are ⩾T score of 63, then the individual is considered a positive risk or case.

The MHLC Form C,18 assesses the degree to which health-related outcomes are observed by the individual as a result of his own actions, luck or chance, or the influence of significant other people such as health professionals or family members. Form C is designed to be condition specific unlike the earlier scale.10 The 18 items are rated on a six-point scale ranging from 1 (strongly disagree) to 6 (strongly agree). The four subscales are internally consistent with alpha reliabilities of 0.79–0.82 for the chance subscale, 0.70–0.71 for the powerful others subscale, 0.71 for the doctors subscale, and 0.85–0.87 for the internal subscale.18 Test–retest reliability over 1 month was 0.80 for the internal subscale, 0.72 for the chance subscale, 0.58 for the doctors subscale and 0.40 for the other people subscale. Factor analysis of 290 patients with arthritis or pain or diabetes or cancer confirmed a four-factor solution. In this study, participants rated items based on beliefs regarding general health after SCI rather than the unchanging SCI itself.

Attributions of cause and blame and perceptions of avoidability. Participants rated seven items similar to those used in a previous study,16 on a seven-point scale ranging from 1 (not at all true) to 7 (extremely true): ‘I believe that I caused the incident that led to my injury’. ‘I believe that another person or persons caused the incident that led to my injury’. ‘I believe that chance or fate caused the incident that led to my injury’. ‘I believe that I could have avoided the incident that led to my injury’. ‘I believe that another person or persons could have avoided the incident that led to my injury’. ‘I blame myself for my injury’. ‘I blame another person or persons for my injury’.

Procedure

Participants were sent a letter by Spinal Injuries Ireland, or the Irish Wheelchair Association, that informed them of the study aims, and requested co-operation. Subsequent to obtaining informed consent each person was interviewed on a prearranged date at his or her home. Each question was read out and responses were recorded. All interviews were conducted in person; there were no telephone interviews or postal surveys.

We certify that all applicable institutional and governmental regulations concerning the ethical use of human volunteers were followed during the course of this research.

Results

The SCL-90-R operational definition of caseness was used to define the two adjustment groups

There was no evidence for a significant difference in gender between the two adjustment groups (χ2=0.028, df=1, P>0.05) or that type of injury (complete or incomplete) differed between groups (χ2=1.89, df=1, P>0.05). Similarly, there was no evidence that age differed between the groups (t=−0.39, df=28, P>0.05), or that age at injury differed between groups (t=−1.24, df=28, P>0.05), or that the time since injury differed between the two groups (t=1.84, df=28, P>0.05).

T-tests were also carried out on the data to test hypotheses with respect to HLOC scores. Kruskal–Wallace analysis of variances were carried out on the data for attributions of cause, blame and avoidability. Cronbach's α for the current sample was calculated for each subscale and is documented along with means, s.d. and t-values in Tables 3 and 4.

Table 3 Reliability and t-test results for the multidimensional health locus of control
Full size table
Table 4 Kruskal–Wallace ANOVA results for attributions of cause and blame
Full size table

The well-adjusted group had a significantly lower others HLC score (t=−2.24, df=28, P=0.016). The well-adjusted group endorsed self-cause at a somewhat higher level (H=2.85, df=1, P=0.091) and endorsed self-avoidability at a significantly higher level than the less well-adjusted group (H=6.52, df=1, P=0.011).

Discussion

Results from the MHLC (Form C) indicate that the adjustment groups did not differ in beliefs regarding the effect of their actions, chance or their doctor's actions on their health. The less well-adjusted group endorsed their belief in the effect of other's actions at a greater level. The well-adjusted group was therefore less external in perceptions of control over health. Those people saw less of a role for their family and friends in the maintenance of their health after SCI than the people who were poorly adjusted. This is consistent with general LOC theory and HLC theory.

This research also investigated participants’ endorsement of attributions of cause and blame. We failed to observe a difference in self-blame between the adjustment groups. However, participants were asked to separate cause, blame and avoidability. The well-adjusted participants endorsed the belief that they could have avoided their accident at a significantly greater level than the poorly adjusted group. Another attribution ‘self-cause’ was different to the level of a trend. The trend in the data suggests that the well-adjusted group tended more toward the belief that they caused their own accident than the poorly adjusted group.

Further investigation is required to clarify if taking responsibility for the injury reflects a type of internal LOC with respect to the accident. If interpreted in this way, the results suggest that the well-adjusted group manifested a greater internal LOC with respect to aspects of the accident. They denied blaming themselves, yet weakly acknowledged a causal role, and strongly endorsed the possibility they could have avoided the accident. A longitudinal methodology would be required to establish any causal relationship between attributions and adjustment.

This study has a number of limitations. A drawback with the methodology is that there was no mechanism to objectively measure the extent to which persons who said they caused/could have avoided their accident actually caused/could have avoided their accident. There is a question of false positives and false negatives in this regard. This should be addressed as a key factor in any future study. In addition, the extremely small sample size makes generalization difficult. The sample size represents around 2 years worth of incidences of SCI in the Republic of Ireland in that around 12–15 people have a SCI annually. The sample of 30 was a self-selected volunteer sample, raising the possibility that people who were particularly poorly adjusted, or who were encountering significant personal challenges, chose not to participate.