Reverse Scoring in SPSS

ExpExchHelp
ExpExchHelp used Ask the Experts™
on
Hello:

Not sure if this forum has many SPSS supporters... yet, I'll give it a try.

For a survey (to be released in the near future), I plan on conducting the analysis via SPSS.

Questions will use 7-point Likert scales (Strongly Disagree / Strongly Agree) via drop-down menus.  

Now, some questions are formed around a "benefit" factor while others address "cost scenarios".

As you know, a "Strongly Agree" for both a benefit-related question and cost-related question will receive a score of 7.   That's fine... however, I need to somehow reverse the scoring for the cost-related question(s) so that a "7" will be converted into a "1/7" (and a "1" should be multiplied by 7).   This is required in order to measure "good things" on an equal scale.

Now, maybe I could copy/paste an SPSS column in Excel and do the conversion there.   Then copy/paste the updated values back into SPSS.

In the end, however, there's room for error in this approach and I'm wondering if SPSS has a build-in function/setting that can account for the reverse scoring.

If yes, please let me know where/how to set it up?

Thanks,
EEH
Comment
Watch Question

Do more with

Expert Office
EXPERT OFFICE® is a registered trademark of EXPERTS EXCHANGE®
For the reverse shouldn't you be doing 8-x?
So 1 comes out as 7(8-1) and 7 comes out as 1(8-7).
ExpExchHelpAnalyst

Author

Commented:
RobinD:

Come to think about it... you're absolutely right.  
x = 1     = 8 - x   = 8 - 1   = 7
x = 2     = 8 - x   = 8 - 2   = 6
x = 3     = 8 - x   = 8 - 3   = 5
x = 4     = 8 - x   = 8 - 4   = 4
x = 5     = 8 - x   = 8 - 5   = 3
x = 6     = 8 - x   = 8 - 6   = 2
x = 7     = 8 - x   = 8 - 7   = 1

Thus 1 becomes 7, ..., ..., and 7 becomes 1... great!

Now, do you know how this can be done within SPSS (vs. exporting/re-importing in/from another program)?

Thanks,
EEH
I'm sorry I have no knowledge of that, but I hope I helped with the algorithm.
Ensure you’re charging the right price for your IT

Do you wonder if your IT business is truly profitable or if you should raise your prices? Learn how to calculate your overhead burden using our free interactive tool and use it to determine the right price for your IT services. Start calculating Now!

First of all, I'd suggest the SPSS area of Experts Exchange.

Second, SPSS has two built-in functions for this, but under the Transform menu.  One is Recode into Same Variables, the other Recode into Different Variables.  In some versions of SPSS, both are in a submenu under a general Recode menu item.  They differ as you'd expect from the names.  You can specify a list of what you'd like to change into what within each function.  

Both go through the conversion list sequentially but only modify each data row once.  So if you, for example, listed "all values below 9" to be converted into 3, then below that listed 3 to be converted into 1, a dataset full of numbers below 9 would be converted into all 3s.  No numbers that started as 3 would change to 1 because they would have all already been changed to 3 (from 3).  This allows you very fine control over what changes to what.  You'll just want a basic list of each:

1>7
2>6
3>5
4>4
5>3
6>2
7>1
All other values>System-Missing

I recommend using Recode into Different Variables and indicating reversed items with names... for example if Q2 and Q4 are reversed, recoding Q2 into Q2_REV and Q4_REV.  Easier to keep track of your original data - if you just use Recode into Same Variables, it's easy to forget if you've reversed already or not when leaving and returning to the dataset later.

Here is a bit more detail: http://www.theanalysisfactor.com/recoding-variables-spss/
ExpExchHelpAnalyst

Author

Commented:
richdiesal:

Thanks... I searched for the SPSS group... 'didn't see it.  I'll look again for potential next posts.

And thank you for your answer in response to my question.  I haven't completed the process yet but thanks to your guidance should have no trouble (I believe) to complete it.

Very much appreciated.

EEH

P.S.  Given that both posts are valuable, I'll accept yours and Robin's solution.   I hope you're ok w/ that.
It's not very active, so it's easy to miss: http://www.experts-exchange.com/Database/Data_Manipulation/SPSS/

Most people post SPSS questions elsewhere, which pretty much prevents it from becoming active too!  :)
ExpExchHelpAnalyst

Author

Commented:
richdiesal:

Thanks... I'll the forum in mind for next questions.

Btw, I've the the transformation data values in a test data set.   It worked like a charm.  Thanks for sharing the process w/ me.   ;)

EEH
ExpExchHelpAnalyst

Author

Commented:
Oh... may I ask one more question?

I'm sure you're familiar w/ Cronbach's Alpha.  If so, do you recommend computing it BEFORE or AFTER data transformation?    Does it make a difference?

Thanks,
EEH

Do more with

Expert Office
Submit tech questions to Ask the Experts™ at any time to receive solutions, advice, and new ideas from leading industry professionals.

Start 7-Day Free Trial