|
Article Excerpt
'Taking a sickie' is a very common phenomenon in organizations. Recent evidence from the UK, for example, shows that sickness absence accounts for 4% of working time and costs the UK economy 10-12 billion pounds annually (CBI, 2004). It is little wonder, therefore, that absence is cited among the top three concerns of HR professionals. In a recent survey of management policy and practice linked with absence from work (CIPD, 2004), most employers (84%) reported that they seek to reduce these figures by interventions that improve work-related employee attitudes such as job satisfaction, job involvement or organizational commitment. In support of this idea, prior research has indeed documented that these work attitudes are negatively correlated with absence from work (Johns, 1997). In a meta-analysis of the available literature, Hackett (1989) found a relatively small mean correlation of r = -.23 between general job satisfaction and time lost measures of absence. Moreover, Harrison and Martocchio (1998) reported that corresponding correlations for job involvement--defined as the extent to which an individual identifies psychologically with his/her job--are even lower with r = -.14. For organizational commitment, findings are very similar (e.g. r = -.15 in the meta-analysis of Meyer, Stanley, Herscovitch, & Topolnytsky, 2002). These low correlations obviously contradict common assumptions regarding the impact of work attitudes as substantial predictors of withdrawal behaviour (CIPD, 2004). So why do these attitude concepts seem to have so little impact? Our study aims to explore this question using a large sample of employees working in a public organization.
To date, a number of explanations have been suggested for the findings reported above. For example, it has been proposed that range restrictions in the criterion data (low base rate, non-normal distributions of absence data) are responsible for weak relationships (Johns, 1991; Steel, 2003). Moreover, it is often argued that absence measures do not differentiate between voluntary absence (e.g. absence due to low work motivation, i.e. when employees 'take a sickie') and involuntary absence (e.g. absence due to genuine illness). However, employees' work attitudes should primarily affect voluntary behaviour (Sagie, 1998; see Steel, 2003 for a recent discerning discussion of this issue). In this paper, we tested a third explanation that was first proposed by Blau and Boal (1987). These authors suggested that simple attitude-behaviour correlations often will underestimate the importance of attitudes because absence behaviour is best explained by looking at interactions between central work attitudes.
Blau and Boal (1987) elaborated this basic idea by presenting a conceptual model of how job involvement might interact with organizational commitment to affect absence behaviour and turnover. In this model, they assume that different levels and forms of absenteeism result from a combination of both attitudes. Let us briefly consider the four combinations of high and low values of both work attitudes with respect to their impact on absenteeism. First, absence for Individuals high on both variables (these persons are called 'institutional stars') is expected to be a rare event and mainly owing to genuine illness. The reason for this assumption is that these persons are highly motivated to be at work because they value both their job and being part of their organization. Individuals with high job involvement and low organizational commitment ('lone wolves') should be mainly absent for problems linked to their immediate work environment and career enhancing purposes whereas employees with low job involvement and high organizational commitment (so-called 'corporate citizens') value their co-workers and strongly identify with organizational goals. These individuals are most sensitive to norms and absence climates and, accordingly, are mainly absent from work in accordance with organization's legitimate absence rules. Finally, incidents of absence for individuals low on both job involvement and organizational commitment ('apathetic employees') are likely to be highest compared with all other groups. These employees are absent in the sense that they take every chance to withdraw from work wherever this is not penalized by organizational policies. In summary, this model predicts that the combination of work attitudes with different attitude targets (job vs. organization) might be important in at least two different ways. First, the combination of attitudes can have an impact on mean differences in absence behaviour as high/high individuals should be typically less absent from work than low/low individuals in many situations. Second, combinations of attitudes are systematically linked with specific forms of absenteeism (sickness, career enhancing, normative and calculative absence). Therefore, it can be expected that--depending on the specific composition of a sample or depending on the specific organization under investigation--some combinations of attitudes are more or less important in predicting absence from work.
There are a few studies that have examined the importance of such interactions in predicting absence behaviour in organizations and we will now briefly summarize this work. Blau (1986) confirmed in a study on nurses that interactions between job involvement and organizational commitment are indeed significant predictors for indicators of unauthorized lateness and unauthorized absence. For both absence indicators it was found that particularly those nurses in the low/low condition had the highest absence. Mathieu and Kohler (1990) tried to replicate this result in a study with 192 bus drivers from a large public authority in the USA. They also found a significant interaction between organizational commitment and job involvement for an indicator of voluntary absence. Contrary to expectations, this interaction, however, was disordinal. This means, no significant main effect of independent variables on dependent variables exists and the regression lines cross over within the range of scale values (for a detailed discussion regarding the distinction between disordinal and ordinal forms of interactions, see Aiken & West, 1991, p. 22f). As hypothesized, lowest absence was found among bus drivers who expressed both high job involvement and high organizational commitment. However, drivers with low organizational commitment and high job involvement exhibited the highest absence.
The most recent empirical study was conducted by Sagie (1998). He adopted the interaction hypothesis for analysing the potential interplay of job satisfaction and organizational commitment. In reviewing the available evidence he concludes (p. 139) that
voluntary...
|
