I Need This Because
MGT 2530
Professor Levi
THE CORRELATION-CAUSATION ISSUE: AN OVERVIEW
Many findings in organizational behavior and personnel are based on correlational research, in which variables are measured and the statistical association between them is assessed. A statistically significant correlation between two variables indicates that the variables are associated, and that this association is unlikely to have arisen by chance. For example, a significant positive correlation between stress and absenteeism means that in general, individuals who are experiencing higher stress tend to be absent from work more often than individuals who are experiencing lower stress. Interpreting correlations
A correlation is a summary description of reality, and is therefore a fact. For such a fact to be useful, however, it must be given a valid causal interpretation. Often, more than one causal interpretation of a correlation is possible. For example, the correlation between stress and absenteeism can be interpreted causally in three ways. If A = stress, B = absenteeism, and C = some third variable such as age, the basic causal interpretations, or models, are:
1. Stress causes absenteeism: A B
This interpretation assumes that stress causes (i.e., produces, brings about, increases the likelihood of) absenteeism. A fuller explanation of this causal interpretation will usually include an intervening variable that specifies the mechanism by which stress leads to absenteeism. For example, a researcher might propose that individuals who are under a great deal of stress may become ill more often, and therefore take off from work more, than those who are under less stress. Other researchers may propose other intervening variables—for example, that stress leads individuals to become more depressed, or more accident-prone, which in turn leads to higher absenteeism. The key point with respect to this causal interpretation (A—>B), however, is
Professor Levi
THE CORRELATION-CAUSATION ISSUE: AN OVERVIEW
Many findings in organizational behavior and personnel are based on correlational research, in which variables are measured and the statistical association between them is assessed. A statistically significant correlation between two variables indicates that the variables are associated, and that this association is unlikely to have arisen by chance. For example, a significant positive correlation between stress and absenteeism means that in general, individuals who are experiencing higher stress tend to be absent from work more often than individuals who are experiencing lower stress. Interpreting correlations
A correlation is a summary description of reality, and is therefore a fact. For such a fact to be useful, however, it must be given a valid causal interpretation. Often, more than one causal interpretation of a correlation is possible. For example, the correlation between stress and absenteeism can be interpreted causally in three ways. If A = stress, B = absenteeism, and C = some third variable such as age, the basic causal interpretations, or models, are:
1. Stress causes absenteeism: A B
This interpretation assumes that stress causes (i.e., produces, brings about, increases the likelihood of) absenteeism. A fuller explanation of this causal interpretation will usually include an intervening variable that specifies the mechanism by which stress leads to absenteeism. For example, a researcher might propose that individuals who are under a great deal of stress may become ill more often, and therefore take off from work more, than those who are under less stress. Other researchers may propose other intervening variables—for example, that stress leads individuals to become more depressed, or more accident-prone, which in turn leads to higher absenteeism. The key point with respect to this causal interpretation (A—>B), however, is