+ Reply to Thread
Results 1 to 6 of 6

Thread: paired samples

  1. #1
    Points: 1,197, Level: 18
    Level completed: 97%, Points required for next Level: 3

    Posts
    14
    Thanks
    0
    Thanked 0 Times in 0 Posts

    paired samples




    Hello!

    I have two separate group. On each group, I could have done paired samples t-test, but I want to compare between the two of them- is there is any option besides mix ANOVA (that doesn't consider the paired sample issue) that I can do?

    Thank you!
    Tal

  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: paired samples

    I don't know if I understand your design right: You have a continuous outcome Y, and you have two groups A and B, and the individuals differ between the two groups. And within each group you have a repeated measurement of the same individuals regarding two different conditions C and D? In this case why not using a mixed ANOVA via the ID-variable? Which software do you use?

  3. #3
    Points: 1,197, Level: 18
    Level completed: 97%, Points required for next Level: 3

    Posts
    14
    Thanks
    0
    Thanked 0 Times in 0 Posts

    Re: paired samples

    Quote Originally Posted by mmercker View Post
    I don't know if I understand your design right: You have a continuous outcome Y, and you have two groups A and B, and the individuals differ between the two groups. And within each group you have a repeated measurement of the same individuals regarding two different conditions C and D? In this case why not using a mixed ANOVA via the ID-variable? Which software do you use?
    Thanks for your answer.
    I Would say that you understood right, but the two conditions C and D are connected among them. I replicate a new paradigm. In the pardigm they only had one group so they did t- test for pair samples. I do the same, but I have to different samples now. I hope it's clear.
    Tal

  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: paired samples

    I am sorry, maybe it's because I am not a native speaker, I still have problems to understand the design.

    Quote Originally Posted by talgef View Post
    I replicate a new paradigm. In the pardigm they only had one group so they did t- test for pair samples.
    So in this study they had N people and two treatment conditions for each individual, am I right?

    Quote Originally Posted by talgef View Post
    I do the same, but I have to different samples now.
    Are you repeating the same experiment but with other people and you want to analyse both datasets together, or do you have a different design with additional groups compared to the above mentioned experiment?

  5. #5
    Points: 1,197, Level: 18
    Level completed: 97%, Points required for next Level: 3

    Posts
    14
    Thanks
    0
    Thanked 0 Times in 0 Posts

    Re: paired samples

    GROUP A
    LEVEL 1 LEVEL 2

    GROUP B
    LEVEL 1 LEVEL 2

    The two levels are connected to each other, if for example I had only one group- I had to do t- test for paired samples (Its the same people who do the two levels). The two groups are separate. And I want to analyze it all together.

  6. #6
    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: paired samples


    OK, thanks, now it's clear. I don't see that there is something against a mixed ANOVA with the two factor variables Group and Level and the participant ID as the random effect. In R for example you can write this as a mixed linear model:

    library(nlme)
    data$group <- factor(data$group)
    data$level <- factor(data$level)
    model <- lme(outcome ~ group + level, data= data, random=~(1|ID))
    summary(model)

    Here, the significance of the variable "group" tells you about differences in the outcome for the different groups A vs. B (pooled over all levels). Is this what you are looking for?
    Last edited by mmercker; 12-21-2015 at 01:42 PM.

+ 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