How to fit nonlinear models with proc nlin?

[julesNY]

Hello! I am doing a practice problem and I need help. With the practice data below, can someone show me how I would write the proc nlin for this? The initial estimates are a=20, b=20, r=.5 and I think the model is supposed to be something like g(t) = a/(1+b(e^rt))

	Heights
ID	year1	year2	year3	year4	year5	year6	year7	year8	year9	year10	year11
1	2.12	4.8	7.4	9.1	12	13	14	15	15.5	16	16.5
2	2.95	5.5	8.4	9.6	11.5	13.5	15.5	17	17.8	18.5	19.1
3	1.95	4.6	6.6	8.75	10	11.5	13.5	14.5	15	15.5	16
4	1.85	3.3	6	8.6	10.5	13.5	17.5	19	20	20.5	21
5	2.55	4.5	5.8	7.47	9	11.5	14.5	16	16.5	17.5	18
6	3	5.5	8.5	9.85	11.5	13.5	15.5	17	18.5	19.5	20.2
7	2.72	4.1	6.6	9.32	11	13	15.5	16.5	17	17.5	17.8
9	2.05	3.3	6.6	8.7	10	12.5	14.5	16.5	17.5	18	18.6
10	2.71	5	7.1	9	10.5	13	14	15.5	17.5	18	18.5

[Rick_SAS · Accepted solution]

This is an example of fitting a "growth model" to data. For an example that uses PROC NLIN, see "Fit a growth curve in SAS."

The article uses a different model, but the PROC NLIN syntax should be almost the same for your model. 

However, in order to fit the data, the data should be in "long form." So your first step is to put the data into a data set that looks like this:

data Growth;
input ID  Time Height;
datalines;
1   1    2.12
1   2    4.8
1   3    7.4
1   4    9.1
...ETC...
1  11    16.5
2   1    2.95
2   2    5.5
...ETC...
;