# Add Variables together in SPSS using the Compute Procedure (Using the Sum Function) – Part 1

On December 27, 2019 by Raul Dinwiddie

In this tutorial I want to look at and

compare the Sum function in SPSS with just manually adding the variables both of these approaches using the

Compute procedure. So here in this data set we have 5

questions which is very similar to the first tutorial I showed on how to add

variables together the only difference is now I have notice this period here for

person 9 on question five, Q5. This indicates a missing value. So in

other words for person 9 when they completed the five questions for some

reason they didn’t complete question number 5. Now we don’t know if they didn’t

complete it because they didn’t like something about it or if it was simply a

mistake and they just missed it. Let’s go ahead and look at how to

compute how to add these variables together using two approaches and we’ll

see if there’s any difference between the two approaches in how they treat

this missing value. So first we’ll go to Transform, Compute

Variable and I showed this in a previous tutorial so I’m not going to go through

it again except just briefly summarize it I have Total=Q1 + Q2 + Q3

+ Q4 + Q5. That is going to add together all the variables. But let’s

actually add one thing to our total here. Let’s call it Total_Sum indicating that

we are just summing the variables together manually. We’ll click OK and then we get some

output there which will disregard this amount of output from previously which

is irrelevant to what we’re doing here. OK notice the total sum it added

together the five questions but notice what happened for person number 9 now it didn’t add together the values

using this approach where the questions are just manually added together using

the Compute procedure if there’s a missing value anywhere among the five

questions then SPSS simply puts a missing value in the place of the total. OK now

I want to show you another approach using the Compute procedure. So we’ll go

back to Transform and then Compute Variable and now what I want to do is

use a function. OK so now we’re going to call this total

sum function. I’ll just call it Tot_SumF

for function and here I need to delete this area so

highlight it and press Delete and then we’re going to go to the Function group.

Notice if you scroll down you will see Statistical. Double-click on that and

then it brings up – down below – a number of different functions that are available

and we want to use the Sum function that’s going to add up all the values

and there’s a description in the box here that you can read. OK so we want the Sum function so I’m

going to double-click that and notice what it does it puts in here SUM(?,?)

and we need to fill or replace the

question mark replace the question marks with our

variables. So we have Q1 here and then we have Q2 and the format is you put the variable

and then a comma and then the next variable and so on. So Q3, Q4, Q5. OK so we

have SUM Q1 comma, etc., all the way through Q5.

Well.. .I did -20 lbs past two months.More here hawght.so#oWyLHYt

Seems interesting, but .I did -35 lbs in 1 week.Go to hawght.so#7YlEdcg

Cb8GeLJCFq7ocIgQ1

hi, how do i make sense of the computed variable figures?

Hello

how I can add multiple group of variables ?

Your tutorial was very helpful, thank you!

hi, thanks for the assistance. i have a question, when computing the variables, do I only use the rescored variables or just the original values? or both?

Hi,

I've several likert scale questions measuring the same dependent variable. In order to analyze my data, I would like to create a new variable that would combine these scores into one new variable. Looking at this video, I don't think the sum function would be the solution to my question? I've tried to do some research online, but I can't seem to find what I'm looking for.

Thanks Sir

Thank you for your helpful video. Please can you answer my question? I use two-way ANOVA and I have conducted a survey with nine questions grouped into 3 variables (3 questions for each variable). Now I would like to have the score my variable which results from the 3 questions in each variable. So now I'm wondering how can I get my scores for the each single variable? Is it the total sum like your video suggested or should I use the means?

Thanks!

Hello Quantitative Specialists, I have a question hope you can help me. I did a survey to measure the level of materialism and it´s correlation with the preference for big logos. My problem is that I had to many variables that measured materialism so I compute them into one as "Total Materialism" but instead of the sum I did the mean because my answer were all likert scale type, is it ok? I´m doing a CHI Square to analyze the results. Thanks in advance!

hello, thank u for this. Question about correlations betwee scales.. what if the scales are different? e.g 6 point likert scale versus 7 point likert scale? Do i need to standardize first before running a correlation test? many thanks!

Hi QS,

I was wondering if I have 10 questions measuring one construct, say for example Need to belong, is there a way to convert or merge all the 10 questions in one variable?

As I intend to find significance between the hypotheses, it's a construct measured against another construct.

Thank you

Is adding the variable together is the same as grouping the likert scale?

Very helpful tutorial even 3 years later haha, thank you!

Hi! Thank you for this video!

I was wondering if you could help me 🙂 I have a scale (14 items) and I want to create 1 variable, so do I compute variable using the sum function or the mean function. I have to create 1 variable and present it in a table including the mean, SD and alpha coefficient.However, I get different means with both functions and I don't know which one is correct.

hello, thanx,it's really helpful. I hav 1 question,do we need to compute variables after we have entered all the data?i have entered half of my data and then computed it and later when i enter new data its not computed. so can you plz help me with this.

Why would you need to add variables together of a Likert scale? What and where does this help on the analysis? Thanks

Hello, How can I recode the total sum and conduct the ordinal regression? I have four questions for one work family conflict variable, it is ordinal and I added them but I don't know what to do next.

Hey, thanks for the guide, I just have one important question. I had computed variables for 104 respondents. Now I've gathered 30 more respondents. How can I make the computed variables automatically calculate the score for the new respondents? Or do I have to start all over again?

Thank you so much, this will spare me a lot of time in the future 🙂

how to compute a variable which will show only agree and strongly agree ad s collective variable

Dear sir, thanks for you easy explanation. Would you please tell me can I analyze the attitude/influence of celebrity endorsement to product consumers' purchase using Correlation/regression in SPSS? Please Answer and help me how?

Hello i am conducting thesis….on employee job satisfaction on public and private sector banks it is comparative study and i collect the five point likret scale questionnaire from both sector . my question is how can i enter and analysis questionnaire response the two sector banks and made comparison by spss

Thanks alot man, u saved me a lot of time, i had to add 50 variables

Thanks immensely

Thanks for your tutorial. if the number of independent variable measured in likert scale and and dependent variable is from secondary sources. How we find the relationship between independent variable and dependent variable?