Hi, everyone.

I am trying to use the multiple assignment method in SPSS Statistics Missing Values to deal with missing values.

However, negative values are assigned to the continuous scale variables (minutes) which should only take positive values. When I constrain the minimum value to be 0, an error message "Please check the minimum value" appears, and I cannot perform the multiple assignment method properly.

What should I do in such a case?

The syntax is as follows

MULTIPLE IMPUTATION age sex BMI Weight Height Race Education Grades_science Grades_art Study_minutes Sports_minutes Sleep_minutes

/IMPUTE METHOD=AUTO NIMPUTATIONS=20 MAXPCTMISSING=NONE MAXCASEDRAWS=9999 MAXPARAMDRAWS=2 MAXMODELPARAM =1000

/CONSTRAINTS age( RND=1.0 MIN=8.0)

/CONSTRAINTS BMI( RND=1.0 MIN=10.0)

/CONSTRAINTS Weight( MIN=20.0 RND=1.0)

/CONSTRAINTS Height( MIN= 100.0 RND=1.0)

/CONSTRAINTS Grades_science( RND=1.0 MIN=0.0)

/CONSTRAINTS Grades_art( RND=1.0 MIN=0.0)

/CONSTRAINTS Study_minutes( RND=1.0 MIN=0.0)

/CONSTRAINTS Sports_minutes( RND=1.0 MIN=0.0)

/CONSTRAINTS Sleep_minutes( RND=1.0 MIN=0.0)

/MISSINGSUMMARIES NONE

/IMPUTATIONSUMMARIES MODELS DESCRIPTIVES

I apreciate it and thanks in advance.

Best regards.

According to the Algorithms doc for multiple imputation,

Using the linear regression method, a continuous variable may have an imputed value well outside

the range of observed values, so the imputed values of continuous variables can be restricted to

fall within a user-specified range, R. When an imputed value falls outside R, the algorithm draws

another imputed value until a value is drawn within R or r1 draws have been made (the maximum

number of tries allowed for drawing each missing case under the given parameter). If the r1 limit

is reached, a new set of parameters are drawn from the posterior distributions (discarding any

successfully imputed values for this variable during this imputation) and the process of imputing

missing values is repeated until a set of imputed values is obtained for this variable and this

imputation or r2 sets of parameters have been drawn (the maximum number of tries allowed for

drawing parameters).

So I suspect that you are imposing a restriction that cannot be satisfied given the models available. You might experiment by removing contstraints to see where the problem lies.

If the limit is reached, the algorithm stops and issues an error.

Thank you very much for your very detailed answer.

I tried removing the constraint of zero minimum values for study time, sports time, and sleep time, and the algorithm proceeded. However, as a result, SPSS assigned negative values (ex. -60 minutes) to those variables. Therefore, I am not able to use the result of this assignment and I am having trouble again. How can I get the algorithm to proceed without errors and the result of the time variable assignment to be greater than zero? And, I will try to adjust the constraints of the other variables a little more.

Best regards.

You might be able to solve this by increasing the default value for the number of draws...

"If minimum or maximum values are specified for imputed values of scale variables (see CONSTRAINTS

subcommand), the procedure attempts to draw values for a case until it finds a set of values that are

within the specified ranges. By default, a maximum of 50 sets of values are drawn per case. To override

the maximum, specify a positive integer for MAXCASEDRAWS. If a set of values is not obtained within the

specified number of draws per case, the procedure draws another set of model parameters and the

repeats the case-drawing process. By default, a maximum of 2 sets of model parameters are drawn. To

override the maximum, specify a positive integer for MAXPARAMDRAWS. An error occurs if a set of values

within the ranges is not obtained within the specified number of case and parameter draws."

If that doesn't work, you might try a different imputation method.

If that still doesn't work, try removing the constraint and then studying the properties of the cases where negative values were imputed to see why such values were imputed.

Finally, if nothing works, you might have to just leave those values as missing or use the single imputation procedure (MVA) to fill them in after running MULTIPLE IMPUTATION.

Thank you very much for your prompt and detailed answer.

​

I have tried to put the maximum value in MAXPARAMDRAWS, but it failed. Therefore, when I checked the distribution of the variables that the algorithm did not proceed with when the constraint was attached, I got the impression that the values inserted by multiple assignment were inserted in such a way as to maintain a normal distribution.

So there are two things I would like to know.

①Is this the correct understanding of the reason for the error in multiple substitutions?

②What other substitution methods are available in SPSS? If so, how can I use the methods?

​

​Best regards.

​