Q1: Why does the line "Contains data from XZY" in the results of the -describe- command show empty name in XYZ instead of my file, for example: Contains data from instead of Contains data from C:\DATA\MyData.sav A: This line shows the name of the last dataset you opened with -use-. Since the file was in a foreign format, it is desirable to prevent the user from accidentally overwriting the foreign data with Stata's dataset if the -save- command is issued without the filename parameter. This behavior is standard in Stata: e.g. -insheet- exhibits the same behavior. sysuse auto,clear tempfile t outsheet using `t' insheet using `t', clear describe Q2: My value labels have changed, why? A: -usespss- will change the labels for values that it converts to Stata's extended missing values. The nature of the problem is that SPSS file format allows some more flexibility in defining the missing values, consider the following example: Numerical variable X; Missing values in X: 7 8 9 Labels of values in X: 1 "one" 2 "two" 7 "no response" 8 "don't know" 9 "not applicable" After conversion the values 7, 8, and 9 will be replaced in the data with the extended missing values: .a, .b, and .c. The documentation coming with the dataset would document the values 7,8,and 9, because these were the values in the supplied dataset, but the user who imported the data would not know the correspondence between {7,8,9} and {.a,.b,.c}. Two approaches are possible. First approach prioritises retaining original values and indicates "missingness" in the variable label or variable characteristics. The resulting value labels could be: 1 "one" 2 "two" 7 "MISSING: no response" 8 "MISSING: don't know" 9 "MISSING: not applicable" Second approach gives priority to "missingness" and indicates the original values in the labels: 1 "one" 2 "two" .a "[7] no response" .b "[8] don't know" .c "[9] not applicable" The second approach is exactly what -usespss- is doing. The reason for that is that the mean of variable X computed in Stata would then coincide with the mean computed in SPSS without further involvement. There are two [minor] inconveniences associated with this approach: the original value labels might have had the actual value inside already. For example 7 "{7} no response", then after conversion the label for .a becomes "[7] {7} no response", which does not look perfect in the tabulation. Second is that if your program used to report the values in the alphabetical order of the labels, it might now report the results differently (in different order). If you don't care about the nature of the "missingness", simply replace X=. if missing(X) which will use system missing (dot) instead of all missings. Labels might have changed for other reasons. If the help file and this guide do not explain this, please let the author know about the problem. Q3: My variable/value labels became unreadable, but the program reported no error. What's wrong? A: Your labels might be using a codepage. See: http://en.wikipedia.org/wiki/Code_page -usespss- reports the codepage if it is saved in the file. Look for a string like "windows-1251" in the output of -usespss-. It will contain the type of the alphabet that needs to be selected. To set that alphabet, right-click the output window, select , then select the corresponding option for