Question

The data below are measurements of zidovudine (AZT) dosage (mg/kg/h) and steady-state maternal plasma concentration (ng/ml) in 32 subjects:

AZT Dose	AZT Conc.	AZT Dose	AZT Conc.
2.5	832	0.6	771
2.5	672	1.8	757
0.8	904	0.9	213
2.5	554	0.6	394
2.5	600	0.9	391
1.9	500	1.3	430
0.7	815	1.1	440
1.9	340	1.4	567
1.9	592	1.1	337
0.9	391	0.8	450
1.5	710	0.7	174
1.4	591	1	470
1.4	660	1.1	426
1.5	694	0.8	395
1.8	668	1	360
1.8	601	0.9	600

1. Based on the reported r, is the relationship between AZT dose and concentration statistically significant? Use two different methods to make this determination and justify your answer using each method.
2. Perform a complete regression analysis between AZT dose and AZT concentration, then evaluate whether AZT dose can be predictive of AZT concentration using two different statistical approaches and justify your answer using each approach.

Answer 1

not statistically significant.The null hypothesis in testing of regression coefficients (for simple linear regression it is beta 1) is that linear regression is insignificant and the alternate hypothesis states it is significant.

Here p values is high which implies simple linear regression is not significant.

If p value would have been say 0.004 then we could have concluded that it is significant,that is,there exists a linear relationship between the two variables.