percent change - please help!

Discussion in 'SPSS' started by BindiChai Tea, Oct 21, 2011.

  1. Hi!

    I'm trying to calculate the percent change between bullying from year
    1 to year 2 and percent change in harassment from year 1 to year 2. My
    data is set up as follows:


    Year Incident
    1 Bullying
    1 Bullying
    1 Bullying
    1 Harassment
    2 Harassment
    2 Bullying
    2 Harassment

    Anyone have any ideas on how to do this in spss? Any help would be
    very appreciated! Thanks.
     
    BindiChai Tea, Oct 21, 2011
    #1
    1. Advertisements

  2. BindiChai Tea

    David Guest

    What constitutes the base for the calculation of these percentages?
    I don't believe you have provided sufficient information.
     
    David, Oct 21, 2011
    #2
    1. Advertisements

  3. BindiChai Tea

    Rich Ulrich Guest


    To approach the question that you describe, you would want to
    be able to describe what fraction of people suffered from bullying/
    harassment in each year, and how many incidents that each
    person had.

    I don't see how you begin to have that data, even if your
    sample for each year starts out as a narrowly circumscribed
    set of individuals.

    - If you are just trying to learn how to run FREQUENCIES and
    other routines in SPSS, there are tutorials available in SPSS
    and online.
     
    Rich Ulrich, Oct 21, 2011
    #3
  4. Thanks Rich for the response! Sorry, I should have mentioned that this
    is school level data - School A reported 3 incidences of bullying
    among students in year 1 and 1 incidence of harassment among students.
    In year 2 it was 2 incidences of harassment and 1 incidence of
    bullying among students. Please help clarify how I would calculate a
    percent change with this data as it looks like this in the database:

    Year Incident
    1 Bullying
    1 Bullying
    1 Bullying
    1 Harassment
    2 Harassment
    2 Bullying
    2 Harassment
    Thanks so much!
     
    BindiChai Tea, Oct 21, 2011
    #4
  5. BindiChai Tea

    Rich Ulrich Guest

    [snip, previous exchange]
    That still does not look like an adequate data representation.
    Is one line supposed to represent one incident, for one school?
    Is there no School_ID, even?


    Here is a recent report on some school bullying, which
    is even a bit higher than what I expected -
    http://www.foxbororeporter.com/articles/2011/06/23/news/9781823.txt

    .... where over half of their students report occasional bullying, with
    more than 10% at "almost daily."

    So your survey must be some attempt to report adminstrative
    inadequacy in regards to "reported bullying/ harassment."

    I think you need to give the raw counts when they are so
    low as 1 and 3, because any attempt at percentages would
    be misleading.

    There are tutorials.

    To get a percent-change, a convenient method would be
    to aggregate by year, and use vars-to-cases to put the totals
    into one line, then compute the fractions.
     
    Rich Ulrich, Oct 21, 2011
    #5
    1. Advertisements

Ask a Question

Want to reply to this thread or ask your own question?

You'll need to choose a username for the site, which only take a couple of moments (here). After that, you can post your question and our members will help you out.