stata - Interpreting log-log regression results where the original values of one IV have all been increased by 100% -
sorry cryptic title, not know how describe problem in way , i'm sure solution rather straightforward.
i running regression dependent , independent variables transformed natural logarithms. problem have 1 iv, dc-ratio, has lot of values of 0% , log(0) undefined , results in missing value in stata. if use value, stata drop missing values , sample biased since uses funds have dc-ratio >0%. issue more half of dataset have dc-ratio of 0%. idea turn values of 0% dc-ratio 0.0000001%.
my councelors solution, according him used in research, add 1 values of dc-ratio, turning 0% dc-ratio in 100% , 30% in 130%. have followed solution.
besides fact wondering @ possible when in original values dc-ratio of 100% maximum, can't figure out how interpret results. need subtract +1 somewhere?
i familiair interpreting log-log results, can't figure out how deal adjustment of original values. in picture, regression number 9 preferred , know 1% increase in dc-ratio lead increase of 1.01^7.126= 1.0735-1= 7.35% increase in geometric mean of dependent variable. can tell right not make sense in economic context.
i hope shed light on problem.
many in advance!
regression results:

Comments
Post a Comment