topic Re: unknown or missing value under string variable in New SAS User

unknown or missing value under string variable

Uddin — Thu, 16 Jun 2022 11:51:38 GMT

Hi,

I have a data set regarding road fatalities. Here is an example of my data set:

crash_id, bus_involvement

1, yes

2, no

3, -9

4, no

5, no

6, yes

Could you please let me know what does it mean ‘-9’ under crash_id 3? Is it unknown or missed data? If it is then how I will fix the issue before calculation?

Looking for your kind response.

Re: unknown or missing value under string variable

PaigeMiller — Thu, 16 Jun 2022 11:55:35 GMT

I would imagine you should ask the person who created the data. We don't know why -9 shows up there.

Re: unknown or missing value under string variable

Kurt_Bremser — Thu, 16 Jun 2022 12:41:14 GMT

Since you did not post your data as recommended (data step with datalines), we have no idea what it really contains.

It might even be that bus_involvement is numeric, with a format that displays 0 as no and 1 as yes. Since -9 is not covered by the format, it is displayed as is.

Anyway, the meaning of -9 should be defined in the documentation for the data you received. You got none? Request it.

Re: unknown or missing value under string variable

Uddin — Thu, 16 Jun 2022 13:20:14 GMT

Hi,
Thanks for your kind response. Here is the data step:
data road_crash;
input crash_id $ bus_involvement$;
datalines;
20205093 Yes
20205008 No
20205050 -9
20201127 No
20202001 No
20207010 Yes
;
run;
According to data sets, bus_involvement is string variable.

Looking for your kind feedbacks.

Re: unknown or missing value under string variable

mkeintz — Thu, 16 Jun 2022 13:45:37 GMT

It appears the entry of a -9 for bus involvement was a decision made by those who provided the data to you. We could guess what it means, but we can't read their minds. Ask them.

As a technical note, you are reading the -9 as a string value, producing bus_involvement="-9". A SAS procedure like proc freq will treat "-9" as a valid value. For character variables, only a blank is treated as missing. If you learn from your data source that the -9 is intended to represent a missing value, you could reassign "-9" to a blank. But before doing so, I would check to see if there are any records in which bus_involvement is already blank. If so, and if blank really means something different than -9, then you might want to leave bus_involvement unchanged and make a new variable bus_involvement2 from it, which you can modify without losing the original information.

Re: unknown or missing value under string variable

Uddin — Thu, 16 Jun 2022 14:02:14 GMT

Hi,
Thank you so much for your kind response. I checked the whole dataset, and couldn't find any blank under bus_involvement variable. Then do you think I can make it missing or blank? Looking for your kind response. Kind Regards

Re: unknown or missing value under string variable

PaigeMiller — Thu, 16 Jun 2022 14:25:28 GMT

@Uddin wrote:
Hi,
Thank you so much for your kind response. I checked the whole dataset, and couldn't find any blank under bus_involvement variable. Then do you think I can make it missing or blank? Looking for your kind response. Kind Regards

What is the benefit of converting -9 to a blank? I see none.

But you've missed the point. Until you know what -9 means, converting it to something else may not be a good thing to do. Please go find out what -9 means.