Page 1 of 2 1 2 LastLast
Results 1 to 10 of 13
  1. #1
    Join Date
    May 2008
    Posts
    6
    Post Thanks / Like

    Question discourse analysis

    [FONT='Times New Roman','serif']hi everyone...... [/FONT]

    [FONT='Times New Roman','serif']I've a Q for the discourse analysts:[/FONT]

    [FONT='Times New Roman','serif']I'm doing a qualitative-quantitative research on classroom discourse. my question is:[/FONT]
    [FONT='Times New Roman','serif']what's the best statistics to determine [/FONT][FONT='Times New Roman','serif']if participation is equally balanced among students in two types of discussions; written & spoken? & to determin if there are any significant differences in the linguistic features in both types of invirnmets?? [/FONT]
    [FONT='Times New Roman','serif']1- gini coefficients?[/FONT]
    [FONT='Times New Roman','serif']2-t test on the totals?[/FONT]
    [FONT='Times New Roman','serif']3-[/FONT][FONT='Times New Roman','serif']Chi-Square analyses for each measure?[/FONT]
    [FONT='Times New Roman','serif']4-ANOVA?[/FONT]

  2. #2
    Join Date
    Nov 2007
    Posts
    5,425
    Post Thanks / Like

    Re: discourse analysis

    The statistic, parametric or non-parametric, that you choose to analyze the data will depend on the type of data you collect eg , continuous (normally distributed) versus nominal...and this will depend on the form in which your data is gathered.
    Firstly, then, how will you be scoring the qualitative data? This will need a non-parametric test , but which depends on whether the data is nominal or ordinal.
    What test/instrument/whatever will you be using for the quantitative data? Does it come with a manual that gives the test statistics of the instrument (mean, SD, standard error of measurement etc, or is the just some kind of raw 'count'?
    Last edited by David L.; 02-May-2008 at 15:34.

  3. #3
    Join Date
    May 2008
    Posts
    6
    Post Thanks / Like

    Question Re: discourse analysis

    thanks for ur instant reply,......I really do appreciate it.

    well,.....u r right! it dpends heavily on the type of data I collected.

    quantitative data were collected by means of survey questionnaires, where as the qualitative one were collected through the use of observations. the oral classroom discourse was obtained from recordings & was transcribed , while the written discourse was gathered by printing out electronic discussions. the two bodies of data will be analyzed, coded and then compared and examined to look for possible qualitative and quantitative diffeences.

    I was stuck and confused when I reached the statistical part & again i really need ur help!

  4. #4
    Join Date
    Nov 2007
    Posts
    5,425
    Post Thanks / Like

    Re: discourse analysis

    the two bodies of data will be analyzed, coded and then compared and examined to look for possible qualitative and quantitative diffeences.

    After 'two bodies of data', there is a great jump to 'will be analyzed'.
    In what form is the data?
    quantitative data were collected by means of survey questionnaires
    Do you really mean 'surveys'? If so, then I cannot see how you can obtain any combined score on the survey - you would have to use very elementary descriptive statistics looking at, and comparing each individual question, within and across the two groups.
    Start with the written data first: what were the questions and how were the responses scored?

  5. #5
    Join Date
    May 2008
    Posts
    6
    Post Thanks / Like

    Question Re: discourse analysis

    hi....
    ur concern is highly appreciatd.

    I have used a mixed (triangulated) data collection to accomplish the objectives of my study:
    1- pre-test & post-test questionnaire....('the quantitative method' is used to test correlations between amount of participation and students' attitudes in the two invironmets )
    2- observations....(the qualitative one)
    3- oral recordings.....transcribed
    4- electronic discussions.......printed
    Both scripts of oral and written (quantitative data) were categorized according to a scheme, and then counted and analyzed using statistical procedure e.g ANOVA, chi-squre, gini coefficients,..etc ??
    Discriptive statistics is used to show frequencies and percetages of discourse features used by the two groups under investigation.
    my question is:
    Is there a difference between values if I use Chi-square or ANOVA for example to determine significant differences between the two groups? if yes, which of these is more appropriate?

  6. #6
    Join Date
    Nov 2007
    Posts
    5,425
    Post Thanks / Like

    Re: discourse analysis

    How far have you gone with the statistical analysis? I am unsure from your tenses:
    Both scripts of oral and written (quantitative data) were categorized according to a scheme, and then (were) counted and analyzed using statistical procedure e.g ANOVA, chi-squre, gini coefficients,..etc ??

    Have you done this (were) or do you mean you are thinking of doing this?
    Discriptive statistics is used to show frequencies and percetages of discourse features used by the two groups under investigation.
    my question is:
    Is there a difference between values if I use Chi-square or ANOVA for example to determine significant differences between the two groups? if yes, which of these is more appropriate?


    With regard to the bold text, do you mean: is one test more powerful than the other, so that by using, say, ANOVA, the power of that test is more likely to reveal a statistically significant difference, compared to the Chi-Square?
    These two tests require very different kinds of data. I go back to what I was saying: you can look at which tests are appropriate (and then compare the power of each) when you know what TYPE of data you have. I can tell you that correlations are a very very very poor choice in terms of what you are examining. Do you have these scores on computer? How many subjects (people) did you have? Can you pick out and post (or PM me) the scores you are referring to when you write, '(quantitative data) were categorized according to a scheme, and then counted'

  7. #7
    Join Date
    May 2008
    Posts
    6
    Post Thanks / Like

    Exclamation Re: discourse analysis

    to be frank, I'm not done yet with the statistics . I'm still in the process of coding & obtaining the total number(frequency) & caculating percentage of each category used or employed by subjects in both settings(the 1st stage of analysis, as I may call it). in the 2nd stage of analysis (which I haven't reached yet), I have to conduct a statisttical procedure (chi-square, ANOVA ,gini coefficients, etc.) to determine if there are any significant difference between frequencies of responses and features used by students.

    you can look at which tests are appropriate (and then compare the power of each) when you know what TYPE of data you have?? I didn't get any answer!
    that's why i'm posting my question here??? knowing the type of data doesn't imply knowing statistics even the basic ones?

    I can tell you that correlations are a very very very poor choice in terms of what you are examining! Is it possible to tell me why is it a poor choice?

    waiting for ur reply..........


  8. #8
    Join Date
    Nov 2007
    Posts
    5,425
    Post Thanks / Like

    Re: discourse analysis

    that's why i'm posting my question here??? knowing the type of data doesn't imply knowing statistics even the basic ones?

    You are asking me for advice, which statistical test should you use. That depends of the type/the kind of data your gathering methods yield. That was what I was asking for - you seem to be saying that all you will have is frequencies and percentages, so:
    the two types of analysis possible are a simple test of the difference between percentages
    and
    dichotomize the data eg No. above and below 50% for each of the two student groups, and do a 2x2 Chi-Square.
    Do this for me:
    choose 2 students from each of the two groups.
    list the tests, and give the scores for each student eg

    Test : / student A1 A2 student B1 B2

    written xy : /25% 35% 62% 55%

    oral zy : /31% 42% 41% 88%

  9. #9
    Join Date
    May 2008
    Posts
    6
    Post Thanks / Like

    Post Re: discourse analysis

    hi David

    thanx again for ur concern..........

    I've come across a website that illustrates statistical tests more & supports ur previous explanations.
    How to choose a statistical test

    I'm now trying to apply chi-square on a random sample & see how effective is this test going to be!!
    I might return back to u to get some of ur valuable advices & guidance

    I need maybe some time to grasp statistical confusions since I'm a bigginer in this field. u asked me to do 2x2 chi-square & i'm trying now to do it manually.

    I'm still looking forward to ur suggestions on using correlations! why do u look at it as a poor choice?

  10. #10
    Join Date
    Mar 2007
    Posts
    61
    Post Thanks / Like

    Re: discourse analysis

    You need to state what your hypothesis is in the first place. "The presence of ____ factor leads to an increase/decrease of ______ factor." (or something like that) It has to be clearly stated so that you can determine whether one variable has an influence on another variable. (you also need to determine what kind of variables you are dealing with) In addition, your research must have accuracy and reliability that is, does it measure what it purports to measure and can it be repeated by another researcher and achieve the same results?
    Correlations are NOT necessarily a poor choice for statistical analysis because this research is aligned with social science, NOT physical science. Therefore, a lower threshold of percentage of error is allowed in this type of research. Comprenez vous?

Page 1 of 2 1 2 LastLast

Similar Threads

  1. Discourse Analysis
    By Ann nona in forum Ask a Teacher
    Replies: 1
    Last Post: 11-Apr-2008, 20:43
  2. Discourse Analysis
    By bigjohn in forum Ask a Teacher
    Replies: 1
    Last Post: 02-Feb-2008, 12:24
  3. discourse analysis
    By pharmacy in forum Ask a Teacher
    Replies: 1
    Last Post: 09-Dec-2007, 02:40
  4. discourse analysis
    By pharmacy in forum Linguistics
    Replies: 4
    Last Post: 22-Nov-2007, 08:24
  5. Replies: 8
    Last Post: 02-Jan-2007, 14:48

Posting Permissions

  • You may not post new threads
  • You may not post replies
  • You may not post attachments
  • You may not edit your posts
  •