Question

You have estimated the correlation between exposure to noise (independent variable) and hearing loss (dependent variable) among factory workers. Both exposure and outcome measures were collected as continuous variables. The statistical analysis showed a Pearson r of 0.4 and the correlation between the two variables was highly significant at p < 0.001.

Question: Can one argue that the relationship between the investigated exposure and outcome is causal? why and How likely is the possibility that the results were a mere chance finding? Why? If You were requested to again investigate the association between exposure to noise and hearing loss while also accounting for each worker’s age, level of education, past history of a hearing ailment, and presence of comorbidity. What statistical analysis would provide an answer and after you have conducted the analysis previously, the association between exposure to noise and hearing loss weakened and became statistically insignificant. Explain this finding

Answer 1

The relationship can not be taken completely casual .there could be other factors that are contributing to hearing loss and correlation between exposure to noise and hearing loss may be nonsense correlation.

If the statistical analysis involves other explanatory variables such as workers age,hearing ailment,level of education and presence of comorbidity. We can apply multiple linear regression and test for the significance of regression coefficients.

If the analysis reveals that the association between exposure to noise and hearing loss is insignificant this indicates that any of the newly added variables is correlated to exposure to noise while exposure to noise is not directly correlated to hearing loss and exposure to noise has nothing to do with hearing loss.