Announcement

Collapse
No announcement yet.
X
  • Filter
  • Time
  • Show
Clear All
new posts

  • Interpreting Output from Fisher Exact Test

    Hello!

    I hope all is well. I conducted a survey of over 200 Black women between ages 18 and 75 to assess potential factors that might explain Black women's contemporary birth control choices. Because the majority of my data is nominal/ categorical, I have been employing Chi-square and Fisher tests. Unfortunately, these tests do not illuminate the direction of the relationship. As such, I have been trying to interpret the column percentages to make sense of things. However, in doing so, I am finding that, though there is statistical significance, the relationships have no/ minimal practical value. I have included an example below:

    In this example, I ran a fisher exact test to determine whether there was a statistically significant association between knowledge of past
    historical incidents of Black women’s negative experiences with the medical establishment (knowev) and current types of birth control methods (classbc).

    In looking at the column percentages, although STATA determined that there was a statistically significant association, it appears that it has no practical significance. I thought that one particular reason for this might be due to the fact that my
    sample is skewed toward people between ages 18 and 30 (this group makes up more than 74% of my sample).

    Does anyone know whether demographic skew impacts statistical significance? If so, can someone provide insight into how I can zero in on this segment of my sample and explore trends within this subset of respondents?




  • #2
    Maya:
    your chances of getting helpful replies are conditional on posting what you typed and what Stata gave you back (As per FAQ).
    That said, if you have several independent variables and one (I guess) multinomial dependent variable (ie, birth control choices), why not considering -mlogit-?
    Kind regards,
    Carlo
    (StataNow 19.0)

    Comment


    • #3
      Hi Carlos,

      Thank you for your reply! I attached a screenshot of what I typed into STATA and what it gave me. Did the picture come through on your end? I have attached it here in case the picture did not post.
      Click image for larger version

Name:	Example Output.png
Views:	2
Size:	47.3 KB
ID:	1539010


      Comment


      • #4
        Example of what I typed in and what STATA gave me.
        Attached Files

        Comment


        • #5
          Maya:
          Fisher's exact test outcome tells you that there's a statistically significant association between -knowev- and -classbc-.
          Kind regards,
          Carlo
          (StataNow 19.0)

          Comment

          Working...
          X