Should I use an ANCOVA? Please help!

Home Forums Methodspace discussion Should I use an ANCOVA? Please help!

Viewing 3 posts - 1 through 3 (of 3 total)
  • Author
    Posts
  • #1536

    Hi all,

    I am doing a prospective observational study where I have two groups of subjects (drug and no drug) and they are being assessed on a continuous variable outcome measure.

    It is also known that age affects this outcome. Unfortunately, due to the observational nature of the study, there is a significant difference in age between the two groups.

    At first I thought I could use an ANCOVA approach, adjusting for age. However, after doing some research, it appears to me that an ANCOVA would only be appropriate if there were wide age ranges within both my groups, but not between groups (and thus the ANCOVA would help clean up the variability).

    Can anyone please point me in the right direction I should be going to adjust/correct for this difference in age? I really appreciate it.

    -Mike

    #1538
    Dave Collingridge
    Participant

    Mike,

    You say that there is a significant difference between ages in the two groups. Have you tested it with a t-test yet? Is this difference random or is it due to some other reason? If the difference is random then you just proceed with the analysis and acknowledge the problem in your study limitations. If the age difference is due to some other predictable effect, such as that older people typically are assigned to one treatment level and younger people to another treatment level, then you might consider leaving age out of the analysis and state in your write up that your results apply to situations where older people are characteristically in one tmt level and younger people are characteristically in the other tmt level.

    #1537

    yes you can  use ANCOVA. to see  the effect of age on your response run the analysis with age as covariate and with out age and compare whether the age has an effect on your response variable or not. 

Viewing 3 posts - 1 through 3 (of 3 total)
  • You must be logged in to reply to this topic.