Following a successful imputation run, my data was organized in the browse window as follows:
m0 farthest to the left, the unfilled data set, had variables from left to right as v1,v2,v3,...vn The next m1 imputed set also had variables from left to right v1,v2,v3...vn, and so forth. My patients were in successive rows as normal and expected,going downwards.
That data set had been set as MI wide.
Some months later, i loaded a saved copy of this same data set. I may have added some imputation sets.
A query showed MI set at wide.
My patients were correct, going downward in rows. But the order of the variables was different: v1,v1,v1 then v2,v2,v2 then v3,v3,v3. In this case the first v1 was in m0, the second v1 was in m2 and the third was in m3. And so forth for the other variables.
Have tried several mi convert commands to recover the first variable order, but no luck. The flongsep command requires a new file name, and that is confusing and unsuccessful.
How do I get back the original, that is the first variable order described above? Thanks, Pete
m0 farthest to the left, the unfilled data set, had variables from left to right as v1,v2,v3,...vn The next m1 imputed set also had variables from left to right v1,v2,v3...vn, and so forth. My patients were in successive rows as normal and expected,going downwards.
That data set had been set as MI wide.
Some months later, i loaded a saved copy of this same data set. I may have added some imputation sets.
A query showed MI set at wide.
My patients were correct, going downward in rows. But the order of the variables was different: v1,v1,v1 then v2,v2,v2 then v3,v3,v3. In this case the first v1 was in m0, the second v1 was in m2 and the third was in m3. And so forth for the other variables.
Have tried several mi convert commands to recover the first variable order, but no luck. The flongsep command requires a new file name, and that is confusing and unsuccessful.
How do I get back the original, that is the first variable order described above? Thanks, Pete
