Search results

  1. hlsmith

    Uncertainty about normality (Q-Q Plot)

    @Karabiner - funny enough I read recently from someone of influence to just always use robust SEs as well. I don't think I will unless there is an obvious threat. And yeah, I have also heard it is just safe to always use pooled SEs in ttests. @tt13 - you should look for heterogeneity. Even if...
  2. hlsmith

    If one can represent the whole group?

    Can you provide a snippet of your data (fictitious or not), so we can better understand what you are writing about. Thank you.
  3. hlsmith

    Interaction-Test

    You could, but if you think about the number of Obs, things will get fairly sparse. Meaning I believe with ordinal logistic reg you will be comparing multiple groups to a reference, plus the other main effect term and the interaction. So you would be comparing 1 to 5,..,1 to 2.
  4. hlsmith

    Don't know what statistical test to run for my data

    What is your sample size, how many school types are there, and how many options are in the Likert scale and what do the responses look like (near the top, bottom, or middle of response options)?
  5. hlsmith

    Interaction-Test

    Not traditional practice yet, but I would fit a multiple quantile regression model. The process is just like linear regression, but you wont have to worry about that hideous DV. I haven't done one with a binary*continuous interaction, but it seems like it should be fine.
  6. hlsmith

    Interaction-Test

    What do the distribution of response to the DV look like? Are they centered, or a whole bunch right at 5?
  7. hlsmith

    Interaction-Test

    Please define your sample size and provide more information about the DV.
  8. hlsmith

    Interpretation of logistic regression results

    Well you also have your intercept, which would be when both predictors are set to zero, then purchage type when custom set to zero, vice versa, and then the when both set to 1. It is all about making sure you know which variable group is set as the reference, we can't tell from your post. Can...
  9. hlsmith

    Comparing group means with drastically different sample sizes

    What program do you think you will be using? I just grabbed this image off the internet, but this is what you would expect as the output: You sent them up similar to a regression and get output like this for each variable. So the y-axis is for the outcome and the x-axis is the comparison of...
  10. hlsmith

    Compare regression coefficients

    I would look at the partial R-squared values in a multiple regression. So regress true height on X1 and X2 and calculate the partial R-square values with confidence intervals. This isn't a direct comparison between them, but it will quantify the amount of variability in Y explained by each...
  11. hlsmith

    regression issues

    You need to think about the relationships between the variables. If it is exogenous, it won't be a big deal. If it is a mediator, you just won't know its effect but its cause should help explain the outcome. It all depends on the purpose. If you are using ORs, RRs, or RDs as your outcome...
  12. hlsmith

    regression issues

    I have told you this before, run the model with and without the extreme outliers, if it doesn't substantially influence the estimates you are fine.
  13. hlsmith

    Why depend on numbers?

    Ditto!
  14. hlsmith

    Duel Links Card Pull Probability

    Cold people are not the only ones that shake.
  15. hlsmith

    Post hoc power analysis

    I wouldn't call it post hoc, since you haven't ran the study analyses yet. Post hoc power analyses are a major faux pas, since most people do them after finding significant results.
  16. hlsmith

    Using racial distribution for a population data as a reference

    No problem. Welcome to the forum!
  17. hlsmith

    Using racial distribution for a population data as a reference

    If you are just desiring to report on historic data and are not generalizing or predicting, then you are correct. So your conclusions will be your conclusions. Though you can still use stats (means, correlations, rates), you just don't have to add precision values (confidence intervals or...
  18. hlsmith

    Using racial distribution for a population data as a reference

    Please better describe this, it currently isn't clear what you plan to do. Also, since this is time to event data, a proportional hazards model may be more appropriate.
  19. hlsmith

    COVID Probability

    T cells, yes. I'll update duration next week, but someone could have 95% lower antibody response hypothetically, though what is the true threshold for not being able to combat exposure. UNKNOWN.
  20. hlsmith

    COVID Probability

    Aerosalized transmission is the treat. Idiots at my gym have never worn masks. I have been fully vaccinated for months and always wear my mask, part of the 3% that do. Though these idiots all of a sudden wipe the machines down, which wouldNt be such an issue if you wore a mask and balked...