Our discussion forums are available to anyone to read, but you must be a member to reply or start new topics. Log-in or register to get started.

This topic contains 2 replies, has 2 voices, and was last updated by  Kasey Murphy 8 months, 3 weeks ago.

  • Author
    Posts
  • #331975

    Kasey Murphy
    Participant

    Could you provide further explanation on the confidence interval values on this question:

    The most appropriate interpretation of change score results between the assessment 6 and assessment 7 administrations would be which of the following?

    a. The child made measurable progress with gross motor skills that exceeded the threshold for test error
    b. The child plateaued with gross motor skills, as evidenced by the upper limit confidence interval score
    c. The child regressed with gross motor skills, as evidenced by a decrease in his percentile score
    d. The child made measurable progress with gross motor skills, although he has not exceeded the threshold for test error
    Even though a child’s score may change, when the confidence intervals around the two scores overlap, the change may be due to measurement variability or error as opposed to a true functional change.

  • #332355

    Jessica Lewis
    Participant

    For Assessment 6 the patient had a GMFM-66 score of 53.09 (CI 50.68-55.50) and for Assessment 7 the patient had a GMFM-66 score of 55.92 (CI 53.63-58.21). The confidence interval takes into account the possibility for test error and gives a range for what numbers we can be 95% confident in that the patient’s score falls between. Since the patient’s confidence intervals from Assessment 6 and 7 overlap (between the scores of 53.63-55.50) it is possible that the patient’s scores actually did not really change if we consider that there could have been test error that is causing the score to be higher or lower than it really should be. So, yes, the patient had a positive change score of 2.83 (measurable change), but since the confidence intervals overlap the scores did not exceed the threshold for test error. Does this help?

  • #332360

    Kasey Murphy
    Participant

    Yes, definitely helps! thank you!

You must be logged in to reply to this topic.