Discussion: View Thread

  • 1.  Dealing with categorical variables analysis

    Posted 05-02-2012 00:30
    Dear Colleagues,
            I am analysing a data set in which all the variables are tapped in categorical answers of 1=yes and 2= No. One variable has 10 items and others same like this. All item are categorical. Can any one help me if i want determine causality among variables, how should i deal with them. Should i take the mean for ten items, or make dummy variables or how i use them when IV and DV both are categorical and in multiple items. I want apply regression on it. If there is any reading which may help in determining the suitable technique and procedure to handle this type of data or any advice is welcomed. Thanks in anticipation.
     
    Regards
     
    Farooq


  • 2.  Dealing with categorical variables analysis

    Posted 05-02-2012 10:04
    Dear Farooq,
    when you convert the variables in binary 0-1 variables you can use Boolean Algebra and Bayesian network analysis.
    http://en.wikipedia.org/wiki/Bayesian_network
    http://en.wikipedia.org/wiki/Boolean_algebra
    best wishes
    Gerhard Fink

    jam farooq schrieb:
    1335933008.69734.YahooMailClassic@web29703.mail.ird.yahoo.com" type="cite">
    Dear Colleagues,
            I am analysing a data set in which all the variables are tapped in categorical answers of 1=yes and 2= No. One variable has 10 items and others same like this. All item are categorical. Can any one help me if i want determine causality among variables, how should i deal with them. Should i take the mean for ten items, or make dummy variables or how i use them when IV and DV both are categorical and in multiple items. I want apply regression on it. If there is any reading which may help in determining the suitable technique and procedure to handle this type of data or any advice is welcomed. Thanks in anticipation.
     
    Regards
     
    Farooq

    --

    Gerhard Fink
    Jean Monnet Professor

    E-Mail: gerhard.fink@wu.ac.at

    View my research on my Author pages:

    http://ssrn.com/author=92836

    http://wuvienna.academia.edu/GerhardFink

     

    Outstanding Paper Award Winner

    Emerald Literati Network Awards for Excellence 2012:

    Maurice Yolles, Gerhard Fink and Daniel Dauber "Organisations as emergent normative personalities: part 1, the concepts"

    published in Kybernetes, Vol. 40, Nos. 5-6, pp. 635-669, 2011.

    See also: http://papers.ssrn.com/sol3/papers.cfm?abstract_id=1924056

    View my research on my SSRN Author page: http://ssrn.com/author=92836

     

    IACCM 2012

    Political change, cultural dynamics and competitiveness of firms

    20 – 22 June, 2012, Università Parthenope di Napoli

    http://www.wu.ac.at/iaccm/iaccm_2012

     



  • 3.  Dealing with categorical variables analysis

    Posted 05-02-2012 11:56

    This is a good resource for conducting regressions with categorical variables using STATA:

    http://www.ats.ucla.edu/stat/stata/webbooks/reg/chapter3/statareg3.htm

     

    From: Organizational Behavior Division Listserv [mailto:OB@AOMLISTS.PACE.EDU] On Behalf Of jam farooq
    Sent: Tuesday, May 01, 2012 9:30 PM
    To: OB@AOMLISTS.PACE.EDU
    Subject: [OB-LIST] Dealing with categorical variables analysis

     

    Dear Colleagues,

            I am analysing a data set in which all the variables are tapped in categorical answers of 1=yes and 2= No. One variable has 10 items and others same like this. All item are categorical. Can any one help me if i want determine causality among variables, how should i deal with them. Should i take the mean for ten items, or make dummy variables or how i use them when IV and DV both are categorical and in multiple items. I want apply regression on it. If there is any reading which may help in determining the suitable technique and procedure to handle this type of data or any advice is welcomed. Thanks in anticipation.

     

    Regards

     

    Farooq

     



  • 4.  Dealing with categorical variables analysis

    Posted 05-02-2012 14:14
    Dear Farooq,
     
    If you have only two variables that are tapped on nominal scales, then first the "arthimetic mean" can not be determined, however, you can use "Mode" for this pupose. Secondly, causality can only be measured if you have longitudinal data or doing experimental type of research. It seems that you have cross sectional data so i'll suggest u in using association/co-relation between the variables instated of causality.
     
    In answering to your question a statistical test namely X2(chi square) is used for measuring association when IV and DV both are tapped on nominal scale (yes, no). Further, regression is used when you have DV as continuous (interval/ratio measured), also the dummy coding is performed when you have three categories for an IV. Whereas, for two categories of an IV, poin-biserial correlation is used. Thanks with regards 
     
    Faqir Sajjad Ul Hassan (PhD Scholar)
    Department of Public Administration,
    Gomal University, D.I.Khan
    00923467998990
     

    From: jam farooq <jam_farooq@YAHOO.CO.UK>
    To: OB@AOMLISTS.PACE.EDU
    Sent: Wednesday, May 2, 2012 9:30 AM
    Subject: [OB-LIST] Dealing with categorical variables analysis
    Dear Colleagues,
            I am analysing a data set in which all the variables are tapped in categorical answers of 1=yes and 2= No. One variable has 10 items and others same like this. All item are categorical. Can any one help me if i want determine causality among variables, how should i deal with them. Should i take the mean for ten items, or make dummy variables or how i use them when IV and DV both are categorical and in multiple items. I want apply regression on it. If there is any reading which may help in determining the suitable technique and procedure to handle this type of data or any advice is welcomed. Thanks in anticipation.
     
    Regards
     
    Farooq