+ Reply to Thread
Results 1 to 5 of 5

Thread: ANOVA or Correlational analysis? Which one to use?

  1. #1
    Points: 15, Level: 1
    Level completed: 29%, Points required for next Level: 35

    Posts
    2
    Thanks
    0
    Thanked 0 Times in 0 Posts

    ANOVA or Correlational analysis? Which one to use?




    Hello everyone,

    I am currently working on my dissertation and aim to identify the relationship between parent-child interaction and behavior of children with low average intelligence (IQ= 80-89).

    I intend to use two surveys, the Child Behavioral Checklist (CBCL)- parent form and the Parenting Stress Index- Short form (PSI-SF-4). Both surveys generate raw scores and T scores for the data (which is collected on a 3 and 5 point likert scale respectively).

    I also hypothesize that negative parent-child interaction, may negatively impact behavior in children.

    Given this information, I am confused as to the type of statistical analysis to conduct on the information obtained. Any help regarding if I should conduct an ANOVA or a Correlation (if not another test), and how to do so would be very helpful. I will be using the SPSS software to analyze my data.

    Kindly help! I apologize if my questions sound too trivial!

  2. #2
    Points: 2,462, Level: 30
    Level completed: 8%, Points required for next Level: 138

    Posts
    200
    Thanks
    20
    Thanked 48 Times in 43 Posts

    Re: ANOVA or Correlational analysis? Which one to use?

    Hi,

    as far as I know, likert scales are ordinal. That means, you have two ordinal variables and you want to look if there is a relationship between them? A first approach would be to use a Spearman's rank correlation, which is also appropriate for ordinal data.

  3. #3
    Points: 15, Level: 1
    Level completed: 29%, Points required for next Level: 35

    Posts
    2
    Thanks
    0
    Thanked 0 Times in 0 Posts

    Re: ANOVA or Correlational analysis? Which one to use?

    Hi,

    Thank you for the reply. I have suggested the use of Spearman's rank correlation in my proposal. But my advisor's suggest the use of some parametric test as well to meet the standards of a masters level dissertation. Any suggestion on what I could use in addition to Spearman's ran correlation?

    Thank you!


    Quote Originally Posted by mmercker View Post
    Hi,

    as far as I know, likert scales are ordinal. That means, you have two ordinal variables and you want to look if there is a relationship between them? A first approach would be to use a Spearman's rank correlation, which is also appropriate for ordinal data.

  4. #4
    Points: 2,462, Level: 30
    Level completed: 8%, Points required for next Level: 138

    Posts
    200
    Thanks
    20
    Thanked 48 Times in 43 Posts

    Re: ANOVA or Correlational analysis? Which one to use?

    Quote Originally Posted by Andy26 View Post
    But my advisor's suggest the use of some parametric test as well to meet the standards of a masters level dissertation.
    Thank you!
    As far as I know, the use of other tests (such as parametric ANOVA or the non-parametric Kruskal-Wallis Test) with likert data is controversial. You find some useful information here, here, and here.

  5. #5
    TS Contributor
    Points: 17,779, Level: 84
    Level completed: 86%, Points required for next Level: 71
    Karabiner's Avatar
    Location
    FC Schalke 04, Germany
    Posts
    2,542
    Thanks
    56
    Thanked 640 Times in 602 Posts

    Re: ANOVA or Correlational analysis? Which one to use?


    Quote Originally Posted by mmercker View Post
    As far as I know, the use of other tests (such as parametric ANOVA or the non-parametric Kruskal-Wallis Test) with likert data is controversial. You find some useful information here, here, and here.
    This seemingly is confusing Likert scales with Likert-type items (items using the Likert response format)? This is exactly what is critizied by your third reference. And the references critisize treatment of Likert-type items as interval scaled , but AFAICS they do not claim that Likert scales could not be treated as interval. The original poster seemingly uses 2 multi-item scales, which could supposedly be treated as interval. By the way, why shouldn't an ordinal scaled respone variable not be analysed using Kruskal-Wallis, which is a test for ordinal/ranked data?

    As to the original question, seemingly only Spearman's rho was used here, the straightforward alternative would be Pearson's r.

    With kind regards

    K.

+ Reply to Thread

           




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