We're calculating the mean of 29 attributes for 5720 (to be precise) respondents. In SPSS, the overall average comes out to 6.58, but in Excel, it's 6.69. This difference is because SPSS calculates the mean at the respondent level, while in Excel, we average the 29 attribute means to get a grand average. Can SPSS be set up to calculate the mean the same way as Excel?

We're calculating the mean of 29 attributes for 5720 (to be precise) respondents. In SPSS, the overall average comes out to 6.58, but in Excel, it's 6.69. This difference is because SPSS calculates the mean at the respondent level, while in Excel, we average the 29 attribute means to get a grand average. Can SPSS be set up to calculate the mean the same way as Excel?

Hello @Vinod Kumar.  You said you want a "grand average".  I gather you mean that you want to add up all raw scores and divide by the number of raw scores.  If so, and if you want other statistics besides the mean (e.g., the SD of all raw scores), you might find it convenient to restructure the file from WIDE to LONG (assuming it is currently a wide file).  If your variables are called x1 to x29, for example, and assuming you have a numeric ID variable called ID, you could do this:

* Restructure from WIDE to LONG.
VARSTOCASES
  /MAKE X FROM x1 TO x29
  /INDEX=i(29) 
  /KEEP=ID 
  /NULL=KEEP.
* Now the mean of X will be the grand mean of all X-scores.
DESCRIPTIVES X.

I hope this helps.

We're calculating the mean of 29 attributes for 5720 (to be precise) respondents. In SPSS, the overall average comes out to 6.58, but in Excel, it's 6.69. This difference is because SPSS calculates the mean at the respondent level, while in Excel, we average the 29 attribute means to get a grand average. Can SPSS be set up to calculate the mean the same way as Excel?

Thanks for the solution, Bruce.

Yes, stacking the data would give me the same results, the issue is that I have 600 other variables. The mean of the means needs to be included in further calculations, so I can't simply stack my entire dataset. Doing so would make it too long and would require me to revise the entire syntax for other data tables, which have already been set up.

Thanks again!

Hello @Vinod Kumar.  You said you want a "grand average".  I gather you mean that you want to add up all raw scores and divide by the number of raw scores.  If so, and if you want other statistics besides the mean (e.g., the SD of all raw scores), you might find it convenient to restructure the file from WIDE to LONG (assuming it is currently a wide file).  If your variables are called x1 to x29, for example, and assuming you have a numeric ID variable called ID, you could do this:

* Restructure from WIDE to LONG.
VARSTOCASES
  /MAKE X FROM x1 TO x29
  /INDEX=i(29) 
  /KEEP=ID 
  /NULL=KEEP.
* Now the mean of X will be the grand mean of all X-scores.
DESCRIPTIVES X.

I hope this helps.

We're calculating the mean of 29 attributes for 5720 (to be precise) respondents. In SPSS, the overall average comes out to 6.58, but in Excel, it's 6.69. This difference is because SPSS calculates the mean at the respondent level, while in Excel, we average the 29 attribute means to get a grand average. Can SPSS be set up to calculate the mean the same way as Excel?

Thanks for the solution, Bruce.

Yes, stacking the data would give me the same results, the issue is that I have 600 other variables. The mean of the means needs to be included in further calculations, so I can't simply stack my entire dataset. Doing so would make it too long and would require me to revise the entire syntax for other data tables, which have already been set up.

Thanks again!

Hello @Vinod Kumar.  You said you want a "grand average".  I gather you mean that you want to add up all raw scores and divide by the number of raw scores.  If so, and if you want other statistics besides the mean (e.g., the SD of all raw scores), you might find it convenient to restructure the file from WIDE to LONG (assuming it is currently a wide file).  If your variables are called x1 to x29, for example, and assuming you have a numeric ID variable called ID, you could do this:

* Restructure from WIDE to LONG.
VARSTOCASES
  /MAKE X FROM x1 TO x29
  /INDEX=i(29) 
  /KEEP=ID 
  /NULL=KEEP.
* Now the mean of X will be the grand mean of all X-scores.
DESCRIPTIVES X.

I hope this helps.

We're calculating the mean of 29 attributes for 5720 (to be precise) respondents. In SPSS, the overall average comes out to 6.58, but in Excel, it's 6.69. This difference is because SPSS calculates the mean at the respondent level, while in Excel, we average the 29 attribute means to get a grand average. Can SPSS be set up to calculate the mean the same way as Excel?

Hi Jon - That also doesn't gives me correct result.

Thanks for the solution, Bruce.

Yes, stacking the data would give me the same results, the issue is that I have 600 other variables. The mean of the means needs to be included in further calculations, so I can't simply stack my entire dataset. Doing so would make it too long and would require me to revise the entire syntax for other data tables, which have already been set up.

Thanks again!

Hello @Vinod Kumar.  You said you want a "grand average".  I gather you mean that you want to add up all raw scores and divide by the number of raw scores.  If so, and if you want other statistics besides the mean (e.g., the SD of all raw scores), you might find it convenient to restructure the file from WIDE to LONG (assuming it is currently a wide file).  If your variables are called x1 to x29, for example, and assuming you have a numeric ID variable called ID, you could do this:

* Restructure from WIDE to LONG.
VARSTOCASES
  /MAKE X FROM x1 TO x29
  /INDEX=i(29) 
  /KEEP=ID 
  /NULL=KEEP.
* Now the mean of X will be the grand mean of all X-scores.
DESCRIPTIVES X.

I hope this helps.

We're calculating the mean of 29 attributes for 5720 (to be precise) respondents. In SPSS, the overall average comes out to 6.58, but in Excel, it's 6.69. This difference is because SPSS calculates the mean at the respondent level, while in Excel, we average the 29 attribute means to get a grand average. Can SPSS be set up to calculate the mean the same way as Excel?

Hi Jon - That also doesn't gives me correct result.

Thanks for the solution, Bruce.

Yes, stacking the data would give me the same results, the issue is that I have 600 other variables. The mean of the means needs to be included in further calculations, so I can't simply stack my entire dataset. Doing so would make it too long and would require me to revise the entire syntax for other data tables, which have already been set up.

Thanks again!

Hello @Vinod Kumar.  You said you want a "grand average".  I gather you mean that you want to add up all raw scores and divide by the number of raw scores.  If so, and if you want other statistics besides the mean (e.g., the SD of all raw scores), you might find it convenient to restructure the file from WIDE to LONG (assuming it is currently a wide file).  If your variables are called x1 to x29, for example, and assuming you have a numeric ID variable called ID, you could do this:

* Restructure from WIDE to LONG.
VARSTOCASES
  /MAKE X FROM x1 TO x29
  /INDEX=i(29) 
  /KEEP=ID 
  /NULL=KEEP.
* Now the mean of X will be the grand mean of all X-scores.
DESCRIPTIVES X.

I hope this helps.

We're calculating the mean of 29 attributes for 5720 (to be precise) respondents. In SPSS, the overall average comes out to 6.58, but in Excel, it's 6.69. This difference is because SPSS calculates the mean at the respondent level, while in Excel, we average the 29 attribute means to get a grand average. Can SPSS be set up to calculate the mean the same way as Excel?

The main difference seems to lie in how we're calculating the values in Excel versus SPSS. In Excel, we calculate the average for each of the 29 variables individually, and then take the average of those 29 average scores. So, the sum of those 29 average scores divided by 29 gives me a slightly different result. In contrast, in SPSS, we're summing the data for all respondents and dividing it by the total number of respondents. This is the key difference. While stacking the data might work, I can't stack the entire 5270-record database, as doing so would disrupt the syntax I've already set up for tabulation.

Hi Jon - That also doesn't gives me correct result.

Thanks for the solution, Bruce.

Yes, stacking the data would give me the same results, the issue is that I have 600 other variables. The mean of the means needs to be included in further calculations, so I can't simply stack my entire dataset. Doing so would make it too long and would require me to revise the entire syntax for other data tables, which have already been set up.

Thanks again!

Hello @Vinod Kumar.  You said you want a "grand average".  I gather you mean that you want to add up all raw scores and divide by the number of raw scores.  If so, and if you want other statistics besides the mean (e.g., the SD of all raw scores), you might find it convenient to restructure the file from WIDE to LONG (assuming it is currently a wide file).  If your variables are called x1 to x29, for example, and assuming you have a numeric ID variable called ID, you could do this:

* Restructure from WIDE to LONG.
VARSTOCASES
  /MAKE X FROM x1 TO x29
  /INDEX=i(29) 
  /KEEP=ID 
  /NULL=KEEP.
* Now the mean of X will be the grand mean of all X-scores.
DESCRIPTIVES X.

I hope this helps.

We're calculating the mean of 29 attributes for 5720 (to be precise) respondents. In SPSS, the overall average comes out to 6.58, but in Excel, it's 6.69. This difference is because SPSS calculates the mean at the respondent level, while in Excel, we average the 29 attribute means to get a grand average. Can SPSS be set up to calculate the mean the same way as Excel?

The main difference seems to lie in how we're calculating the values in Excel versus SPSS. In Excel, we calculate the average for each of the 29 variables individually, and then take the average of those 29 average scores. So, the sum of those 29 average scores divided by 29 gives me a slightly different result. In contrast, in SPSS, we're summing the data for all respondents and dividing it by the total number of respondents. This is the key difference. While stacking the data might work, I can't stack the entire 5270-record database, as doing so would disrupt the syntax I've already set up for tabulation.

Hi Jon - That also doesn't gives me correct result.

Thanks for the solution, Bruce.

Yes, stacking the data would give me the same results, the issue is that I have 600 other variables. The mean of the means needs to be included in further calculations, so I can't simply stack my entire dataset. Doing so would make it too long and would require me to revise the entire syntax for other data tables, which have already been set up.

Thanks again!

Hello @Vinod Kumar.  You said you want a "grand average".  I gather you mean that you want to add up all raw scores and divide by the number of raw scores.  If so, and if you want other statistics besides the mean (e.g., the SD of all raw scores), you might find it convenient to restructure the file from WIDE to LONG (assuming it is currently a wide file).  If your variables are called x1 to x29, for example, and assuming you have a numeric ID variable called ID, you could do this:

* Restructure from WIDE to LONG.
VARSTOCASES
  /MAKE X FROM x1 TO x29
  /INDEX=i(29) 
  /KEEP=ID 
  /NULL=KEEP.
* Now the mean of X will be the grand mean of all X-scores.
DESCRIPTIVES X.

I hope this helps.

We're calculating the mean of 29 attributes for 5720 (to be precise) respondents. In SPSS, the overall average comes out to 6.58, but in Excel, it's 6.69. This difference is because SPSS calculates the mean at the respondent level, while in Excel, we average the 29 attribute means to get a grand average. Can SPSS be set up to calculate the mean the same way as Excel?