。。。。。。。。。。。。。

You should solve the problem with linear regression instead of logistic regression which is designed for 0/1 situations.

You can either regression original population or natural log of population against year. Here is a starting point for you.

data Pop;
input Year Pop ;
datalines;
1990        114.333
1991        115.823
1992        117.171
1993        118.517
1994        119.85
1995        121.121
1996        122.389
1997        123.626
1998        124.761
1999        125.786
2000        126.743
2001        127.627
2002        128.453
2003        129.227
2004        129.988
2005        130.756
2006        131.448
2007        132.129
2008        132.802
;
run;

proc reg data =pop;
  model Pop = Year;
run;

data pop;
set pop;
logPop=log(pop);
run;

proc reg data =pop;
  model logPop = Year;
run;
quit;

楼上正解。