Solve missing data - reduced observations in regression

[Fairouz Mustafa]

Good day,

I have a problem with reducing the number of observations after running the regression, the main dataset is 256 obs, it was reduced to 106 after the regression. I checked this link and I know why, but how to solve it and run the regression with full data

https://www.statalist.org/forums/for...gression-model


The command is

regress SDG6Indx_w LogBTC_w OLogStock_w OLogMCAP_w OLogGold_w GDPCptaUSD_w LogInfRate_w Gindx_w HofstCultIndx_w BusDisclosIndx_w Innov_indx_w RDexpofGDP_w AnnGDPGroth_w UnemployOfLabr_w PRemGDP_w,robust


SEQ Country CountryID Year SDG6Indx SDG7Indx LogSDG7 SDG12Indx LogSDG12 SDG13CO2Em logSDG13 BITinUSD LogBTC
8 Argentina 1 2020 .4712 1.570e+08 8.195839 893159.00 5.9509288
9 Australia 2 2013 .83165 1 1 .01768606 -1.7523689 3.932e+08 8.5945811 147966.68 5.1701639
11 Australia 2 2015 .8419 1 1 .01092886 -1.9614253 4.033e+08 8.6056061 496334.31 5.6957743
12 Australia 2 2016 .84675 1 1 .01357889 -1.8671356 4.159e+08 8.6190361 635201.85 5.8029118
13 Australia 2 2017 .7791626 1 1 .01685388 -1.7733 4.148e+08 8.6177876 1501758.33 6.1766
15 Australia 2 2019 .86125 1 1 .01785285 -1.7482924 4.145e+08 8.6175422 621640.15 5.7935391
16 Australia 2 2020 .8661 3.919e+08 8.5931663 506940.83 5.7049573
24 Brazil 3 2020 .57513333 4.674e+08 8.6696734 355980.05 5.5514257
27 Canada 4 2015 .867525 1 1 .00111631 -2.9522148 5.731e+08 8.7582011 111516.10 5.0473376

[Rich Goldstein]

based on the small amount of info provided, I see two options: (1) use -sem- with the "method(mlmv)" option; (2) use multiple imputation to impute the data and then estimate the results; see

h sem
h mi

note, however, that with this much missing data, there will continue to be doubt about your results

maybe one (or a small number) of your variables is responsible for most of the missing data? see

h misstable