Re: Which Statistical test to use?!

Posted by Saeed Rezaei-2 on Dec 19, 2012; 1:36pm
URL: http://spssx-discussion.165.s1.nabble.com/Which-Statistical-test-to-use-tp5717005p5717051.html

The age group were defined a priori and the respondents ticked the related boxes.

On Thu, Dec 20, 2012 at 12:33 AM, Art Kendall <[hidden email]> wrote:

did the respondents have to endorse one of those categories, or were they asked their age, and the data coarsened into categories later?

Art Kendall
Social Research Consultants

On 12/19/2012 7:36 AM, Saeed Rezaei wrote:

Age is categorized into
11-15, 16-20, 21-25, and 26+

On Wed, Dec 19, 2012 at 11:34 PM, Art Kendall <[hidden email]> wrote:

Do you have age as a continuous variable?

Gender is a dichotomy so it is okay to treat it as interval variable (like a dummy variable).  Since there is only one interval, all intervals are perfectly equal to each other.

Also education is at least ordinal. 

It is usually good to try to avoid coarsening your data.

Art Kendall
Social Research Consultants

On 12/19/2012 12:05 AM, Saeed Rezaei wrote:

Thanks Kumar for your comments but I dont think Chi-square is appropriate cuz the data from the questionnaire is score (interval data) and only the age, gender, and education are transformed into categories. I can use Chi-square if the data from the questionnaire is also categorical like low, medium, and high i language identity. I dont wanna transform my already interval data to ordinal or nominal data cuz it would culminate in missing information.
What do you think?

On Wed, Dec 19, 2012 at 2:13 AM, Alok Kumar Jha <[hidden email]> wrote:

Saeed
Use Chi-Square Test of association... it allows you to test if there is a
significant relationship between two or more categorical variables....

Thanks & Regards
Alok Kumar
Business Intelligence & Data Warehousing,
Lab Services & Solutions-India,
India Software Lab, EGL Business Park,
Bangalore – 560 071 India
====================================
Lotus Notes: Alok Kumar Jha/India/IBM
Internet address: [hidden email]
Office: 91 80 41777089
Mobile: [phone]

I have no special talents. I am only passionately curious.
(Embedded image moved to file: pic19194.jpg)Silver Redbook Author
(Embedded image moved to file: pic04976.jpg)



From:   Saeed Rezaei <[hidden email]>
To:     [hidden email]
Date:   12/18/2012 08:13 PM

Subject:        Re: Which Statistical test to use?!

Sent by:        "SPSSX(r) Discussion" <[hidden email]>

Thanks for your informative comments everybody. I have 1805 participants in
my survey.


On Wed, Dec 19, 2012 at 12:49 AM, Art Kendall <[hidden email]> wrote:
  first crosstab age, gender, and language proficiency level.
  do you have sufficient numbers in all of the cells to examine the three
  way interaction?

  then use glm to do a repeated measures anova using as many interaction
  terms as your cells sizes allow.  The subscales are the repeats.
  Art Kendall
  Social Research Consultants
  On 12/18/2012 2:01 AM, Saeed Rezaei wrote:
        Hi all,
        I am dubious about which statistical test to use for my research
        questions. I would be glad to receive your comments. My research
        question is:

        Is there any relationship between the Iranian English language
        learners' language identity and their age, gender, and language
        proficiency level.
        The language identity is measured through a validated questionnaire
        with 19 items. And the language identity included six subscales in
        the questionnaire namely:
        1. attachment to persian language
        2. pronunciation attitude
        3. language and social status
        4. L1 use and exposure in the society
        5. language knowledge
        6. script

        Each of the above subscales had 3 to 4 items.

        Now the first research question will be transformed into multiple
        research questions including:
              1.        There  is  no  relationship between Iranian English
              language learners’ attachment to Persian and their
                 ·         gender,
                 ·          age, and
                 ·         proficiency level
              2.        There  is  no  relationship between Iranian English
              language learners’ pronunciation attitude and their
                 ·         gender,
                 ·          age, and
                 ·         proficiency level
              3.        There  is  no  relationship between Iranian English
              language learners’ language and social status and their
                 ·         gender,
                 ·          age, and
                 ·         proficiency level
              4.        There  is  no  relationship between Iranian English
              language learners’ L1 use/exposure in the society and their
                 ·         gender,
                 ·          age, and
                 ·         proficiency level
              5.        There  is  no  relationship between Iranian English
              language learners’ language knowledge and their
                 ·         gender,
                 ·          age, and
                 ·         proficiency level
              6.        There  is  no  relationship between Iranian English
              language learners’ script/alphabet and their
                 ·         gender,
                 ·          age, and
                 ·         proficiency level





        Now....what do you think would be the best statistical test? Do you
        think I should use MANOVA? Do not forget that I am after "relation"
        not "effect". How should I enter the data for each subscale.
        Thanks in advance to all of you for your kind attention.

        Saeed