+ Reply to Thread
Results 1 to 9 of 9

Thread: Help on Factor analysis and Extension

  1. #1
    Points: 2,622, Level: 31
    Level completed: 15%, Points required for next Level: 128

    Posts
    4
    Thanks
    0
    Thanked 0 Times in 0 Posts

    Help on Factor analysis and Extension




    Hello frnds ,

    Recently we had done a project on Factor analysis approach to understand the indian consumer behavior

    Our survey questioner contains 140 variables on that 40 are independent type (Yes or No ,age , income group ,place ,gender ,price is a factor ..yes or no etc....) and 100 are dependent variables on a scale of 1 to 5 .

    Those 100 variables divided in to 4 factor analysis parts based up on their respective categories like Psychology , Service provider services etc...

    After doing factor analysis on each part we got 6 or 7 factors an average .

    For independent variables we just applied Cross Tabulation analysis that's all.

    Can any body tell me what to do further development and how to relate these independent variables with extracted factors of factor analysis .

    Plzz Help me ............. Im confused in this part ,im unable to do any further analysis .

    One of internet frnd had given some replay just go for Multivariate analysis.But i don't know what to with that plzz give me some clear idea please .

  2. #2
    TS Contributor
    Points: 8,362, Level: 61
    Level completed: 71%, Points required for next Level: 88

    Location
    Crete, Greece
    Posts
    717
    Thanks
    0
    Thanked 35 Times in 34 Posts
    first of all the FA method requires continuous variables, so Yes or No type variables must not use FA with them. You can do clustering.

  3. #3
    Points: 2,622, Level: 31
    Level completed: 15%, Points required for next Level: 128

    Posts
    4
    Thanks
    0
    Thanked 0 Times in 0 Posts
    those 100 variables are continuous variables , 40 are Yes are No type ..........

  4. #4
    TS Contributor
    Points: 8,362, Level: 61
    Level completed: 71%, Points required for next Level: 88

    Location
    Crete, Greece
    Posts
    717
    Thanks
    0
    Thanked 35 Times in 34 Posts
    if the 100 variables are also strongly correlated then you are ok with FA. I assume you use SPSS. It has a demonstration of how to do FA in help. or else go http://statlink.tripod.com and then statistical packages and then SPSS. The first link can be very useful to you.

  5. #5
    Points: 3,448, Level: 36
    Level completed: 66%, Points required for next Level: 52

    Location
    Indiana
    Posts
    27
    Thanks
    0
    Thanked 1 Time in 1 Post

    SPSS default with FA

    I love, love, love SPSS and have used it extensively for EFA.

    But please do remember that the default setting for extraction is Principal Components. So be sure to change it (a common one is "Principal Axis Factoring") if you are wanting to do a factor analysis.

  6. #6
    TS Contributor
    Points: 8,362, Level: 61
    Level completed: 71%, Points required for next Level: 88

    Location
    Crete, Greece
    Posts
    717
    Thanks
    0
    Thanked 35 Times in 34 Posts
    You change it (rotate the axes actually) for other reason, to clarify the data view.

  7. #7
    Points: 2,798, Level: 32
    Level completed: 32%, Points required for next Level: 102

    Posts
    45
    Thanks
    0
    Thanked 0 Times in 0 Posts
    aah, regarding your initial question of what to do after factor analysis, i think u can look back to ur study (or research) question:
    your topic is :Factor analysis approach to understand the indian consumer behavior:
    that is, i guess, u want to say that which age group is more psychological, and all that kind of thing.
    and for that u can use ANOVA in spss. use this for age, for gender, or for any other group u have and analyse in detail.
    i think this will be enough for ur study. (if ur a beginner in statistics/this research course)

  8. #8
    Points: 2,622, Level: 31
    Level completed: 15%, Points required for next Level: 128

    Posts
    4
    Thanks
    0
    Thanked 0 Times in 0 Posts
    Sir by using One way ANOVA i have to perform nearly 40x4 operations ,Because my survey contains 40 independent variables and 100 Dependent variables(in 4 categories )

    So by selecting each categorie i have to apply 40 operations.

    My theoretical model contains follwing factors

    Dependent variabls:
    Mobile technical questions( 20 questions , WAP ,camera,..etc )
    Service provider services ( 34 qus ,like ring tones ,wallpapers)
    Influencing factors for mobile purchase ( 20qus , like frnds,family ,Tv ads ,online ads ..)
    Influencing factors for service provider( 20 qus ,like frnds ,family ,..etc)

    Independent variables total 40 questions)
    1)Age
    2)Income group
    3)Gender
    4)camera ( VGA, MP)
    5)FM radio (yes , No)
    6)MP3 player (Yes , No)
    7)memory upgrade( 100 MB, 512 MB ,1Gb)

    in this way we prepared 40 questions related to independent variables .

    Our final Goal is to see customer behavior of Mobile phone.

    for that 100 dependent variables we done factor analysis .On each catagorie we got 6 or 7 factors each .

    Now tell me what to do ????

    1)one-way ANOVA
    im getting a long list for each independent variable

    ( using ANOVA - "Income group" with "Mobile technical questions" im getting these errors too.
    1)Groups with only one case are ignored in computing the test of homogeneity of variance for "Battery"
    2)Post hoc tests are not performed for Battery because at least one group has fewer than two cases.)

  9. #9
    Points: 2,622, Level: 31
    Level completed: 15%, Points required for next Level: 128

    Posts
    4
    Thanks
    0
    Thanked 0 Times in 0 Posts

    My aim is to find the behavior of customer in this way

    Customer behavior-------->Male , age (20-25 group) ,income group( 1-2 lakh) , camera(VGA), Browser enabled (Yes ) ,data transfer (GPRS) ----------------> Rating scale questions ( Tech specifications , Influence factors etc....)

    In this way with my own specification of independent variables i have to finf customer behavior.

    Can you please help me out .

+ Reply to Thread

           




Similar Threads

  1. [JMP] Scheirer-Ray-Hare extension of Kruskal-Wallis test
    By adrienbaudouin in forum Other Software
    Replies: 6
    Last Post: 04-25-2013, 07:24 PM
  2. Factor Analysis vs Correspondence Analysis
    By bsaikrishna in forum Statistical Research
    Replies: 1
    Last Post: 12-27-2010, 05:00 AM
  3. Factor Analysis - flipping a factor
    By dave25 in forum Statistics
    Replies: 3
    Last Post: 03-14-2009, 07:26 PM
  4. Replies: 1
    Last Post: 02-12-2008, 03:16 PM
  5. Replies: 3
    Last Post: 09-08-2006, 11:35 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