Hello Steve.  I suspect the problem is that your Year variable has more than 2 levels.  In support of that hypothesis,  I believe you are using the NPTESTS command like this:

*Nonparametric Tests: Independent Samples. 
NPTESTS 
  /INDEPENDENT TEST (Y1 Y2 Y3 Y4 Y5 Y6) GROUP (Year) MANN_WHITNEY 
  /MISSING SCOPE=ANALYSIS USERMISSING=EXCLUDE
  /CRITERIA ALPHA=0.05  CILEVEL=95.

The Mann-Whitney test requires that Year have only 2 values.  If there are more than 2 values, the Decision will be "Unable to compute".  Changing MANN_WHITNEY to KRUSKAL_WALLIS will very likely make your code run and produce results (that would ignore the ordinal nature of Year).  But as I write that, I am reminded of the so-called XY problem.  In other words, I think you should back up and tell us about the nature of your data, and the research questions you are trying to address.  Here are some initial questions that come to mind for me:

1. How many years are there?
2. Does each year represent an independent sample?  Or do some participants have data across multiple years?
3. Can you show us basic descriptive stats for each year?
4. What is the nature of the 6 dependent variables?  Is each one a Likert-type item?  Or are they "scales" for which means & SDs could be legitimately used for description? 

Thanks for clarifying.

Cheers,
Bruce

I cant figure out why the Mann Whitney U test cant be computed.

 

N sizes are different but not extreme

There are frequency distributions...but they look doable

 

Anyone have a reason why?

 

 

 

 

Steven "No Vibe Coding" Wong

Research Data Specialist

Board of Registered Nursing

 

 

Phone: (916) 574-7681

Email: steven.wong@dca.ca.gov

 

Hello Steve.  I suspect the problem is that your Year variable has more than 2 levels.  In support of that hypothesis,  I believe you are using the NPTESTS command like this:

*Nonparametric Tests: Independent Samples. 
NPTESTS 
  /INDEPENDENT TEST (Y1 Y2 Y3 Y4 Y5 Y6) GROUP (Year) MANN_WHITNEY 
  /MISSING SCOPE=ANALYSIS USERMISSING=EXCLUDE
  /CRITERIA ALPHA=0.05  CILEVEL=95.

The Mann-Whitney test requires that Year have only 2 values.  If there are more than 2 values, the Decision will be "Unable to compute".  Changing MANN_WHITNEY to KRUSKAL_WALLIS will very likely make your code run and produce results (that would ignore the ordinal nature of Year).  But as I write that, I am reminded of the so-called XY problem.  In other words, I think you should back up and tell us about the nature of your data, and the research questions you are trying to address.  Here are some initial questions that come to mind for me:

1. How many years are there?
2. Does each year represent an independent sample?  Or do some participants have data across multiple years?
3. Can you show us basic descriptive stats for each year?
4. What is the nature of the 6 dependent variables?  Is each one a Likert-type item?  Or are they "scales" for which means & SDs could be legitimately used for description? 

Thanks for clarifying.

Cheers,
Bruce

I cant figure out why the Mann Whitney U test cant be computed.

 

N sizes are different but not extreme

There are frequency distributions...but they look doable

 

Anyone have a reason why?

 

 

 

 

Steven "No Vibe Coding" Wong

Research Data Specialist

Board of Registered Nursing

 

 

Phone: (916) 574-7681

Email: steven.wong@dca.ca.gov

 

Can you provide a small dataset that allows others to reproduce the result you are getting?  

Hello Steve.  I suspect the problem is that your Year variable has more than 2 levels.  In support of that hypothesis,  I believe you are using the NPTESTS command like this:

*Nonparametric Tests: Independent Samples. 
NPTESTS 
  /INDEPENDENT TEST (Y1 Y2 Y3 Y4 Y5 Y6) GROUP (Year) MANN_WHITNEY 
  /MISSING SCOPE=ANALYSIS USERMISSING=EXCLUDE
  /CRITERIA ALPHA=0.05  CILEVEL=95.

The Mann-Whitney test requires that Year have only 2 values.  If there are more than 2 values, the Decision will be "Unable to compute".  Changing MANN_WHITNEY to KRUSKAL_WALLIS will very likely make your code run and produce results (that would ignore the ordinal nature of Year).  But as I write that, I am reminded of the so-called XY problem.  In other words, I think you should back up and tell us about the nature of your data, and the research questions you are trying to address.  Here are some initial questions that come to mind for me:

1. How many years are there?
2. Does each year represent an independent sample?  Or do some participants have data across multiple years?
3. Can you show us basic descriptive stats for each year?
4. What is the nature of the 6 dependent variables?  Is each one a Likert-type item?  Or are they "scales" for which means & SDs could be legitimately used for description? 

Thanks for clarifying.

Cheers,
Bruce

I cant figure out why the Mann Whitney U test cant be computed.

 

N sizes are different but not extreme

There are frequency distributions...but they look doable

 

Anyone have a reason why?

 

 

 

 

Steven "No Vibe Coding" Wong

Research Data Specialist

Board of Registered Nursing

 

 

Phone: (916) 574-7681

Email: steven.wong@dca.ca.gov

 

Sure Bruce attached is the sample dataset

 

Steven "No Vibe Coding" Wong

Research Data Specialist

Board of Registered Nursing

 

 

Phone: (916) 574-7681

Email: steven.wong@dca.ca.gov

 

Can you provide a small dataset that allows others to reproduce the result you are getting?  

Hello Steve.  I suspect the problem is that your Year variable has more than 2 levels.  In support of that hypothesis,  I believe you are using the NPTESTS command like this:

*Nonparametric Tests: Independent Samples. 
NPTESTS 
  /INDEPENDENT TEST (Y1 Y2 Y3 Y4 Y5 Y6) GROUP (Year) MANN_WHITNEY 
  /MISSING SCOPE=ANALYSIS USERMISSING=EXCLUDE
  /CRITERIA ALPHA=0.05  CILEVEL=95.

The Mann-Whitney test requires that Year have only 2 values.  If there are more than 2 values, the Decision will be "Unable to compute".  Changing MANN_WHITNEY to KRUSKAL_WALLIS will very likely make your code run and produce results (that would ignore the ordinal nature of Year).  But as I write that, I am reminded of the so-called XY problem.  In other words, I think you should back up and tell us about the nature of your data, and the research questions you are trying to address.  Here are some initial questions that come to mind for me:

1. How many years are there?
2. Does each year represent an independent sample?  Or do some participants have data across multiple years?
3. Can you show us basic descriptive stats for each year?
4. What is the nature of the 6 dependent variables?  Is each one a Likert-type item?  Or are they "scales" for which means & SDs could be legitimately used for description? 

Thanks for clarifying.

Cheers,
Bruce

I cant figure out why the Mann Whitney U test cant be computed.

 

N sizes are different but not extreme

There are frequency distributions...but they look doable

 

Anyone have a reason why?

 

 

 

 

Steven "No Vibe Coding" Wong

Research Data Specialist

Board of Registered Nursing

 

 

Phone: (916) 574-7681

Email: steven.wong@dca.ca.gov

 

Thanks Steve.  There is a lot of missing data in that file.  Nevertheless, the following code...

NEW FILE.
DATASET CLOSE ALL.
GET FILE='C:\Temp\spss_2yrs_Sample_07102025.sav'.
DATASET NAME raw.

FREQUENCIES Year.

*Nonparametric Tests: Independent Samples. 
NPTESTS 
  /INDEPENDENT TEST (SMB1 SMB2 SMB3 SMB4 SMB5 SMB6 SMB7 W1 W2 W3 W4 W5 W6)
   GROUP (Year) MANN_WHITNEY 
  /MISSING SCOPE=ANALYSIS USERMISSING=EXCLUDE
  /CRITERIA ALPHA=0.05  CILEVEL=95.

... generated these results:

As footnote d says, no test could be computed for variable W3 because at least one of the groups has no data.  As the following output shows, W3 has no data for 2023.

MEANS SMB1 TO W6 BY Year.

What results do you get using the dataset you uploaded?  

Sure Bruce attached is the sample dataset

 

Steven "No Vibe Coding" Wong

Research Data Specialist

Board of Registered Nursing

 

 

Phone: (916) 574-7681

Email: steven.wong@dca.ca.gov

 

Can you provide a small dataset that allows others to reproduce the result you are getting?  

Hello Steve.  I suspect the problem is that your Year variable has more than 2 levels.  In support of that hypothesis,  I believe you are using the NPTESTS command like this:

*Nonparametric Tests: Independent Samples. 
NPTESTS 
  /INDEPENDENT TEST (Y1 Y2 Y3 Y4 Y5 Y6) GROUP (Year) MANN_WHITNEY 
  /MISSING SCOPE=ANALYSIS USERMISSING=EXCLUDE
  /CRITERIA ALPHA=0.05  CILEVEL=95.

The Mann-Whitney test requires that Year have only 2 values.  If there are more than 2 values, the Decision will be "Unable to compute".  Changing MANN_WHITNEY to KRUSKAL_WALLIS will very likely make your code run and produce results (that would ignore the ordinal nature of Year).  But as I write that, I am reminded of the so-called XY problem.  In other words, I think you should back up and tell us about the nature of your data, and the research questions you are trying to address.  Here are some initial questions that come to mind for me:

1. How many years are there?
2. Does each year represent an independent sample?  Or do some participants have data across multiple years?
3. Can you show us basic descriptive stats for each year?
4. What is the nature of the 6 dependent variables?  Is each one a Likert-type item?  Or are they "scales" for which means & SDs could be legitimately used for description? 

Thanks for clarifying.

Cheers,
Bruce

I cant figure out why the Mann Whitney U test cant be computed.

 

N sizes are different but not extreme

There are frequency distributions...but they look doable

 

Anyone have a reason why?

 

 

 

 

Steven "No Vibe Coding" Wong

Research Data Specialist

Board of Registered Nursing

 

 

Phone: (916) 574-7681

Email: steven.wong@dca.ca.gov

 

Ok...I have tried...pretty much everything, so Im sending the dataset

 

I still cant do a Mann Whitney U with 2 defined groups

 

Can someone help me understand why?

 

Steven "No Vibe Coding" Wong

Research Data Specialist

Board of Registered Nursing

 

 

Phone: (916) 574-7681

Email: steven.wong@dca.ca.gov

 

Thanks Steve.  There is a lot of missing data in that file.  Nevertheless, the following code...

NEW FILE.
DATASET CLOSE ALL.
GET FILE='C:\Temp\spss_2yrs_Sample_07102025.sav'.
DATASET NAME raw.

FREQUENCIES Year.

*Nonparametric Tests: Independent Samples. 
NPTESTS 
  /INDEPENDENT TEST (SMB1 SMB2 SMB3 SMB4 SMB5 SMB6 SMB7 W1 W2 W3 W4 W5 W6)
   GROUP (Year) MANN_WHITNEY 
  /MISSING SCOPE=ANALYSIS USERMISSING=EXCLUDE
  /CRITERIA ALPHA=0.05  CILEVEL=95.

... generated these results:

As footnote d says, no test could be computed for variable W3 because at least one of the groups has no data.  As the following output shows, W3 has no data for 2023.

MEANS SMB1 TO W6 BY Year.

What results do you get using the dataset you uploaded?  

Sure Bruce attached is the sample dataset

 

Steven "No Vibe Coding" Wong

Research Data Specialist

Board of Registered Nursing

 

 

Phone: (916) 574-7681

Email: steven.wong@dca.ca.gov

 

Can you provide a small dataset that allows others to reproduce the result you are getting?  

Hello Steve.  I suspect the problem is that your Year variable has more than 2 levels.  In support of that hypothesis,  I believe you are using the NPTESTS command like this:

*Nonparametric Tests: Independent Samples. 
NPTESTS 
  /INDEPENDENT TEST (Y1 Y2 Y3 Y4 Y5 Y6) GROUP (Year) MANN_WHITNEY 
  /MISSING SCOPE=ANALYSIS USERMISSING=EXCLUDE
  /CRITERIA ALPHA=0.05  CILEVEL=95.

The Mann-Whitney test requires that Year have only 2 values.  If there are more than 2 values, the Decision will be "Unable to compute".  Changing MANN_WHITNEY to KRUSKAL_WALLIS will very likely make your code run and produce results (that would ignore the ordinal nature of Year).  But as I write that, I am reminded of the so-called XY problem.  In other words, I think you should back up and tell us about the nature of your data, and the research questions you are trying to address.  Here are some initial questions that come to mind for me:

1. How many years are there?
2. Does each year represent an independent sample?  Or do some participants have data across multiple years?
3. Can you show us basic descriptive stats for each year?
4. What is the nature of the 6 dependent variables?  Is each one a Likert-type item?  Or are they "scales" for which means & SDs could be legitimately used for description? 

Thanks for clarifying.

Cheers,
Bruce

I cant figure out why the Mann Whitney U test cant be computed.

 

N sizes are different but not extreme

There are frequency distributions...but they look doable

 

Anyone have a reason why?

 

 

 

 

Steven "No Vibe Coding" Wong

Research Data Specialist

Board of Registered Nursing

 

 

Phone: (916) 574-7681

Email: steven.wong@dca.ca.gov

 

There is only one value of year in that file. 

GET FILE = "C:\Temp\SPSS_NO_MannWhitneyU_071782025.sav".
DATASET NAME full.
FREQUENCIES Year.

Ok...I have tried...pretty much everything, so Im sending the dataset

 

I still cant do a Mann Whitney U with 2 defined groups

 

Can someone help me understand why?

 

Steven "No Vibe Coding" Wong

Research Data Specialist

Board of Registered Nursing

 

 

Phone: (916) 574-7681

Email: steven.wong@dca.ca.gov

 

Thanks Steve.  There is a lot of missing data in that file.  Nevertheless, the following code...

NEW FILE.
DATASET CLOSE ALL.
GET FILE='C:\Temp\spss_2yrs_Sample_07102025.sav'.
DATASET NAME raw.

FREQUENCIES Year.

*Nonparametric Tests: Independent Samples. 
NPTESTS 
  /INDEPENDENT TEST (SMB1 SMB2 SMB3 SMB4 SMB5 SMB6 SMB7 W1 W2 W3 W4 W5 W6)
   GROUP (Year) MANN_WHITNEY 
  /MISSING SCOPE=ANALYSIS USERMISSING=EXCLUDE
  /CRITERIA ALPHA=0.05  CILEVEL=95.

... generated these results:

As footnote d says, no test could be computed for variable W3 because at least one of the groups has no data.  As the following output shows, W3 has no data for 2023.

MEANS SMB1 TO W6 BY Year.

What results do you get using the dataset you uploaded?  

Sure Bruce attached is the sample dataset

 

Steven "No Vibe Coding" Wong

Research Data Specialist

Board of Registered Nursing

 

 

Phone: (916) 574-7681

Email: steven.wong@dca.ca.gov

 

Can you provide a small dataset that allows others to reproduce the result you are getting?  

Hello Steve.  I suspect the problem is that your Year variable has more than 2 levels.  In support of that hypothesis,  I believe you are using the NPTESTS command like this:

*Nonparametric Tests: Independent Samples. 
NPTESTS 
  /INDEPENDENT TEST (Y1 Y2 Y3 Y4 Y5 Y6) GROUP (Year) MANN_WHITNEY 
  /MISSING SCOPE=ANALYSIS USERMISSING=EXCLUDE
  /CRITERIA ALPHA=0.05  CILEVEL=95.

The Mann-Whitney test requires that Year have only 2 values.  If there are more than 2 values, the Decision will be "Unable to compute".  Changing MANN_WHITNEY to KRUSKAL_WALLIS will very likely make your code run and produce results (that would ignore the ordinal nature of Year).  But as I write that, I am reminded of the so-called XY problem.  In other words, I think you should back up and tell us about the nature of your data, and the research questions you are trying to address.  Here are some initial questions that come to mind for me:

1. How many years are there?
2. Does each year represent an independent sample?  Or do some participants have data across multiple years?
3. Can you show us basic descriptive stats for each year?
4. What is the nature of the 6 dependent variables?  Is each one a Likert-type item?  Or are they "scales" for which means & SDs could be legitimately used for description? 

Thanks for clarifying.

Cheers,
Bruce

I cant figure out why the Mann Whitney U test cant be computed.

 

N sizes are different but not extreme

There are frequency distributions...but they look doable

 

Anyone have a reason why?

 

 

 

 

Steven "No Vibe Coding" Wong

Research Data Specialist

Board of Registered Nursing

 

 

Phone: (916) 574-7681

Email: steven.wong@dca.ca.gov