+ Reply to Thread
Results 1 to 5 of 5

Thread: Reverse coding in SPSS

  1. #1
    Points: 1,586, Level: 22
    Level completed: 86%, Points required for next Level: 14

    Thanked 0 Times in 0 Posts

    Reverse coding in SPSS

    I am inputting my data into SPSS, and I am really confused about how to input the reverse codes.

    I have a few likert scales, from strongly disagree (1) to strongly agree (5). I havent started putting in any data yet. Can someone tell me what to do with the responses that are reverse coded?

    Do I just label the values the opposite way in variable view? As in '1-strongly agree, 5-strongly disagree' etc, and then just put the responses in as they are?

    Or do I have to leave all the value labels the same, and just change the responses as I put them in? So if someone responds 2-disagree, do I put it in as 4-agree instead?

    Its something so silly, but I'm so lost!

  2. #2
    Super Moderator
    Points: 31,766, Level: 100
    Level completed: 0%, Points required for next Level: 0
    bugman's Avatar
    Thanked 324 Times in 265 Posts

    Re: Reverse coding in SPSS

    As long as they represent the same response and are both on the same scale (i.e. 1-5) this should be fine.
    The earth is round: P<0.05

  3. #3
    Points: 1,253, Level: 19
    Level completed: 53%, Points required for next Level: 47

    Thanked 0 Times in 0 Posts

    Re: Reverse coding in SPSS

    Label your variables and enter your scores as normal, i.e., as if no items were reverse coded.

    This is safer and more efficient than you mentally reversing the scores for certain items and entering them as reversed - you might make a mistake along the way.

    Then, click on Transform => Recode into Same Variable OR Recode into Different Variables (whichever you like - I use the latter in case I've forgotten if I recoded something or not. The former works for datasets you are coding in one take).

    The first reverses the scores for you. The second reverses the scores and includes them as an extra variable, so you have 2 columns per reversed item, one raw score, one reversed score.

    After this, select the variable you need to recode. Enter the new values for each response (so, if Strongly Agree is "5", on a 5-point Likert scale, it is now "1", and so on). Do this for EACH possible value, 1 2 3 4 5. You have to do it for EACH value in one take, or it recodes all your 5s into 1s, and when you go back later to recode your 1s into 5s, you have a value of "5" for both your original 5s and original 1s.

    If using "Recode into Different Variables" option, name your new variable item1_reversed, or similar.

    Hope that helps.

  4. #4
    Points: 2, Level: 1
    Level completed: 3%, Points required for next Level: 48

    Thanked 0 Times in 0 Posts

    Re: Reverse coding in SPSS

    Bugman--what do you mean by "this"?

  5. #5
    Points: 802, Level: 15
    Level completed: 2%, Points required for next Level: 98

    Rhode Island
    Thanked 1 Time in 1 Post

    Re: Reverse coding in SPSS

    The best practice is to enter the data as they appear in the survey, without regard for which items are reverse coded. If you try to reverse code mentally, mistakes will occur.
    If missing data have values in your data coding, make sure that they are coded as missing.
    Then check the data entry by running FREQUENCIES (or DESCRIPTIVES with Min and Max) to make sure that your Likert items are in range.
    Then use SYNTAX to recode the variables. When transforming data, it is useful to have a syntax file to point to later so you can show how you did the recoding. Imagine that you wanted to tell your grandchildren about it.
    A simple way to recode a 5 point Likert scale is with this command, if you want to create a new variable RQ01 to contain the reverse coded Q01:

    COMPUTE RQ01 = ( 6 - Q01) .

    or if you do not want to create a new variable:

    COMPUTE Q01 = ( 6 - Q01 ) .

    This assumes that you have 5 levels and you have defined your missing values. You can also run

    RECODE Q01 ( 1 = 5 ) ( 2 = 4 ) ( 4 = 2 ) (5 = 1 ) or the RECODE ..... INTO ......

    but the COMPUTE command involves less typing.
    "Statistics is the Grammar of Science" Karl Pearson

+ Reply to Thread


Similar Threads

  1. What are reverse confidence intervals?
    By c2q in forum Statistics
    Replies: 6
    Last Post: 04-08-2011, 03:00 PM
  2. Reverse Transpose
    By Emilyab in forum SAS
    Replies: 4
    Last Post: 03-04-2011, 10:38 AM
  3. Sunfish NMDS/PERMANOVA in reverse
    By sms16 in forum Statistical Research
    Replies: 3
    Last Post: 02-08-2011, 12:41 AM
  4. SPSS Coding question
    By finderquest in forum SPSS
    Replies: 1
    Last Post: 03-05-2008, 07:46 AM
  5. Reverse Psychology
    By psychgradstudent in forum Statistics
    Replies: 7
    Last Post: 02-21-2008, 09:43 AM

Posting Permissions

  • You may not post new threads
  • You may not post replies
  • You may not post attachments
  • You may not edit your posts

Advertise on Talk Stats