Assessing linearity between logit and continuous variable.

[Inaamul Haq]

Hi all. The book by Vittinghoff et al (Regression Methods in Biostatistics: Linear, Logistic, Survival, and Repeated Measures Models, 2nd edition) on Page 174 includes Fig 5.4 (attached) to demonstrate the linearity of Logit CHD Risk and Age. A sample of the data is here.

* Example generated by -dataex-. For more info, type help dataex
clear
input byte(age chd69)
50 0
51 0
59 0
51 0
44 0
47 0
40 0
41 0
50 0
43 0
59 0
54 0
48 0
39 1
49 0
55 0
44 0
56 0
42 0
44 0
45 0
42 0
44 0
55 0
49 0
45 1
45 0
54 0
44 0
40 0
41 0
46 0
57 0
40 1
54 0
43 0
40 0
50 0
39 0
45 0
44 0
41 0
46 0
43 0
53 0
52 0
55 0
44 1
40 0
47 0
54 0
39 0
49 1
54 0
58 0
48 0
48 0
43 0
45 0
43 0
41 0
55 1
41 0
44 0
54 0
55 0
57 0
52 0
41 0
54 0
52 0
54 0
51 0
40 0
40 0
49 0
50 0
39 0
43 0
51 0
48 0
49 0
50 0
48 0
39 0
40 0
59 0
46 0
47 0
43 1
47 0
49 0
45 0
42 0
55 0
54 0
56 0
50 1
39 0
51 0
52 0
50 0
55 0
55 0
51 0
57 0
53 0
42 0
51 0
42 0
50 0
49 0
48 1
52 0
47 0
56 1
42 0
47 0
45 0
40 0
44 0
39 0
49 0
56 1
48 0
47 0
45 1
47 0
51 0
54 0
39 0
56 0
57 0
42 0
44 0
41 0
42 1
53 0
55 0
40 0
49 0
42 0
42 0
51 0
52 0
44 0
39 0
55 0
51 0
41 0
46 1
44 0
56 0
48 0
44 0
44 1
46 0
39 0
41 0
41 0
48 0
53 0
51 1
52 1
58 0
49 0
39 0
43 0
46 0
42 0
50 0
47 0
49 0
39 0
53 0
42 0
44 0
40 0
45 0
56 1
47 0
42 0
43 0
39 0
53 0
53 0
44 1
51 1
58 0
59 1
55 0
55 1
46 0
39 0
59 0
46 0
46 0
53 0
55 0
41 0
48 0
49 1
50 0
51 0
50 0
56 0
52 1
44 0
59 0
49 1
46 0
40 0
46 0
50 0
44 0
42 0
49 0
41 0
43 0
40 0
43 0
52 0
51 1
44 0
59 0
42 0
50 0
40 0
48 0
44 0
43 0
44 0
48 0
44 0
39 0
51 1
40 0
42 0
57 0
54 0
40 0
48 0
40 0
47 0
51 0
53 0
54 1
48 0
54 0
46 0
39 0
57 0
50 0
42 0
45 0
54 0
47 0
39 0
50 0
55 1
39 0
55 1
39 0
52 0
45 1
46 0
47 0
46 0
54 0
45 0
43 0
43 0
40 0
42 0
52 0
54 0
41 0
57 0
49 0
58 0
51 0
47 0
41 0
40 0
52 0
59 0
41 0
49 0
43 0
48 0
44 0
57 0
46 0
43 0
39 0
40 0
43 0
45 0
45 0
45 0
42 0
43 0
51 0
41 0
56 0
51 0
40 0
56 1
41 0
54 0
53 0
44 0
39 0
46 0
41 0
49 0
48 0
44 1
40 0
42 0
47 0
52 0
42 0
45 0
53 0
45 0
40 0
40 0
47 0
56 0
46 0
47 0
40 0
49 0
43 0
53 0
50 0
46 1
54 0
51 0
56 0
47 0
40 0
44 0
46 0
47 0
49 0
59 0
50 0
43 0
42 0
59 0
45 0
44 0
52 0
44 0
49 1
49 0
41 0
41 0
57 0
52 0
47 0
39 0
41 0
44 0
49 0
41 1
41 0
42 1
50 0
48 0
59 1
46 0
53 1
40 0
40 0
55 0
48 0
55 0
45 0
46 0
50 0
43 0
39 0
44 0
58 0
56 1
39 0
41 0
48 0
53 1
45 0
48 0
52 1
48 1
44 0
46 0
56 0
41 0
55 1
39 0
48 0
39 0
50 0
46 0
43 0
41 0
52 0
40 0
40 0
56 0
52 1
49 0
48 0
39 0
55 0
50 0
59 1
39 0
47 0
45 0
42 0
56 0
41 0
42 0
44 0
39 0
43 0
40 1
45 0
49 0
54 0
48 0
42 0
44 0
40 0
46 0
46 0
40 0
42 0
45 1
47 1
42 0
39 0
39 0
47 0
51 0
58 0
51 0
52 0
46 0
44 0
44 0
53 1
40 0
44 0
48 0
46 0
52 0
39 0
52 0
58 0
48 0
55 0
59 0
44 0
52 0
52 0
39 0
44 0
47 0
48 1
46 0
45 0
50 0
40 0
42 0
50 0
43 0
55 0
41 0
46 0
49 1
53 0
56 0
39 0
40 0
45 0
43 0
59 0
56 1
40 0
44 0
56 0
39 0
41 0
44 0
41 0
53 0
end
label values chd69 yesno
label def yesno 0 "No", modify
label def yesno 1 "Yes", modify

------------------ copy up to and including the previous line ------------------
The full dataset is available online at https://regression.ucsf.edu/sites/g/.../data/wcgs.dta.
I want to replicate Figure 5.4.
Can someone please help me write the code for Figure 5.4?
The plotted points are the empirical log odds of the outcome for each of the unique values of age observed in the sample.

[Justin Niakamal]

use "https://regression.ucsf.edu/sites/g/files/tkssra6706/f/wysiwyg/home/data/wcgs.dta", clear

logit chd69 i.age, or
predict phat,  xb

tw scatter phat age, sort || lowess phat age, lpattern(shortdash) ||                                 ///
 lfit phat age,  lpattern(dash) ytitle("Logit CHD Risk") xtitle("Age (years)")                       ///
 legend(order(3 "Linear Fit" 2 "LOWESS smooth") ring(0) pos(4) rows(1) region(lcolor(black)))

[Inaamul Haq]

Thank you Justin Niakamal . You saved my day.