Parameter estimates included in the output for the model with age and gender are shown here.
These are expressed relative to the baseline levels of zero for Age 1 and Gender 1. Thus, the estimate for Age
2 is 0.921 higher and that for Age 3 is 0.655 higher than for Age 1. These differences are significant (P<0.01)
and (P<0.05), respectively. From the sizes of the standard errors we can detect that there is no significant
difference between age groups 2 and 3. The parameter estimate for Gender 2 (females) is 0.485 lower than that
for Gender 1 (P<0.05).
It is possible also to obtain estimates on the original observational scale by clicking the
Predict button
at the bottom right of the dialog box shown earlier. By then clicking 'Age' the table alongside is produced.
This gives predicted estimates for the three age groups adjusted for differences between males and females.
Thus, 19% of children were found to be infected with bilharzia compared with an average of 34% in adolescents and adults.
Then, by clicking 'Gender' we see that 34% of males were found to be infected compared with 25% of females.
Please note that this is a purely imaginary interpretation of the values of
the variable N_infections that have been simply used to illustrate the method of logistic regression.
The results are not applicable to the case study.
Parameter estimates included in the output for the model with age and gender are shown here.
These are expressed relative to the baseline levels of zero for Age 1 and Gender 1. Thus, the estimate for Age
2 is 0.921 higher and that for Age 3 is 0.655 higher than for Age 1. These differences are significant (P<0.01)
and (P<0.05), respectively. From the sizes of the standard errors we can detect that there is no significant
difference between age groups 2 and 3. The parameter estimate for Gender 2 (females) is 0.485 lower than that
for Gender 1 (P<0.05).
It is possible also to obtain estimates on the original observational scale by clicking the
Predict button
at the bottom right of the dialog box shown earlier. By then clicking 'Age' the table alongside is produced.
This gives predicted estimates for the three age groups adjusted for differences between males and females.
Thus, 19% of children were found to be infected with bilharzia compared with an average of 34% in adolescents and adults.
Then, by clicking 'Gender' we see that 34% of males were found to be infected compared with 25% of females.
Please note that this is a purely imaginary interpretation of the values of
the variable N_infections that have been simply used to illustrate the method of logistic regression.
The results are not applicable to the case study.
