+ Reply to Thread
Results 1 to 3 of 3

Thread: Replacing missing values in SPSS

  1. #1
    Points: 2,684, Level: 31
    Level completed: 56%, Points required for next Level: 66

    Posts
    3
    Thanks
    0
    Thanked 0 Times in 0 Posts

    Replacing missing values in SPSS




    I have a set of five questions that I plan on putting together as a scale. Possible scores on each item is 1-5, and the scale score will be the mean of the scores of all items.

    I want to exclude everyone that have two or more missing items. If the subject has only one missing item I want to give them the sample mean for that item/question.

    I think this should be done by choosing "series mean" in "replace missing values" for each of the five items. I'd appreciate if anyone could let me know if I am on the right track.

    However, if I do this before excluding the ones that have two or more missing values I don't think I would know who to exclude (there will be no missing items because they will all have the sample mean as their score). Is this correct?

    If it is then I don't know where I should start. If I exclude the cases from the dataset I would loose them "forever" and will not be able to keep them for other analyses later (which I may want to). Should I use "select cases" to include only the ones that have 3 or more valid responses, and then do the "replace missing values"?

    Sorry if this was poorly explained. It's a bit hard for me to write in English and especially about something that confuses me as much as this Please let me know if you need more information in order to answer my question

  2. #2
    Points: 4,760, Level: 44
    Level completed: 5%, Points required for next Level: 190

    Location
    Ithaca, NY
    Posts
    266
    Thanks
    0
    Thanked 2 Times in 2 Posts
    Hi Apple,

    I did a blog post on this exact topic. http://www.analysisfactor.com/statchat/?p=35.

    You don't need to impute the mean of the other four items for a missing value if you're using it as part of a mean. The mean of the four items with one missing will be the same as five items in which one of them has an imputed mean.

    Example:

    The mean of (1,2,4,5) is 3.

    The mean of (1,2,3,4,5) is also 3.

    And you really only want to do mean imputation if your % missing is small. Say, less than 5%. More than that and you'll start to bias your results. I did a free webinar on missing data recently. You can get the recording at: http://www.analysisfactor.com/learni...training6.html.

    Karen
    The Analysis Factor
    http://TheAnalysisFactor.com

  3. #3
    Points: 666, Level: 13
    Level completed: 32%, Points required for next Level: 34

    Posts
    11
    Thanks
    0
    Thanked 0 Times in 0 Posts

    Re: Replacing missing values in SPSS


    Thanks Karen for your webinar and Apple for the post

+ Reply to Thread

           




Similar Threads

  1. Question regarding missing values
    By Rachie in forum SPSS
    Replies: 9
    Last Post: 09-30-2010, 10:55 AM
  2. Missing values
    By SadieKhan in forum Statistics
    Replies: 11
    Last Post: 09-05-2010, 04:08 PM
  3. How have you dealt with missing values?
    By Freaked_out in forum Statistical Research
    Replies: 0
    Last Post: 06-21-2010, 11:54 PM
  4. need help on missing values
    By mohsenn in forum Psychology Statistics
    Replies: 2
    Last Post: 02-18-2009, 01:16 PM
  5. Missing Values
    By marble in forum Statistics
    Replies: 0
    Last Post: 04-24-2007, 07:55 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