Search results

  1. hlsmith

    Very small sample size

    You are very limited, what is the survival rate for these eight observations?
  2. hlsmith

    two different factors in groups being compared

    Perhaps you should think about quantile regression controlling for both variables.
  3. hlsmith

    Interpreting the COX Model

    You need to think about the ordering of events, can death be from other causes? If so, it may be a competing event.
  4. hlsmith

    When to control for a variable?

    For full disclosure, I read none of the above content. Ideally predictors are independent and you don't have to control for them to find an effect. However, if the predictor and the outcome are both effects of a variable you need to control for it to unconfound the effect. You need to know the...
  5. hlsmith

    Correcting for oversampling

    Well if you control for it in the model is there an issue?
  6. hlsmith

    Statistics Notes

    Thanks for sharing Dik!
  7. hlsmith

    hello this is faisal

    Yeah, but you can get into trouble if the mean in near a bound (e.g., 0 or 100) and variability is large. In those scenarios beta distribution may come into play.
  8. hlsmith

    Clueless in need of Help in experimental design and analysis method

    Please write out your question so we know what variables to think about. If you need a continuous, categorical, and multiplicative interaction for whether a purchase is made - this sounds like multiple logistic regression. We don't know your content area, so we can't tell you what variables to...
  9. hlsmith

    Help with poisson

    Well, with the outcome being 1,2,3 - which are integers where 1.5 may be impossible. I would look toward multinomial logistic regression to get probabilities. For more clarification, Poisson regression is typically used for count data or rates. Where your data may be best represented by groups...
  10. hlsmith

    Multiple data points per subject? Aggregation to mean with or without considering subjects

    Please describe the purpose of the project and primary out come. Without knowing more info it is hard to give recommendations. Also what is sample size and how are variable formatted?
  11. hlsmith

    How to show statistical equivalence between groups of data?

    There is probably some type of exact test, but you still need to state a margin (threshold that they can be similar within. If this value is not your primary objective for the study, perhaps just calculating the differences and 95 percentile bootstrap confidence intervals would be enough. You...
  12. hlsmith

    How to show statistical equivalence between groups of data?

    You should probably find the difference between the two time points and run an equivalence test that the value do not differ from '0' given some margin (e.g., +/- 5%) - probably falls back under something like a TOST. You could also, just plot a histogram for the differences.
  13. hlsmith

    Autoregressive distributed lag model with different duration of lag

    Welcome to the forum @Demographer12 Yes, entering a time variable seems like the best option. Side note, If you had multiple measurements and everyone had obs taken at the approximate same time, which you don't, you can create a power structure weight to account for different lengths of...
  14. hlsmith

    data tranformation

    In your face @Dason , in your face!
  15. hlsmith

    Formatting in proc freq and proc sgplot

    If this is a one time deal, I would just insert the comma via editing. Formatting in SAS is a pain. Can you post your current output, so we have a better idea of what you are referring to. Thx.
  16. hlsmith

    How to get amount to pay in order to be fair

    I would base fair also on the number of tickets sold. If only two tickets are sold and you have a 50% chance of either prize - hmm, what would you pay? But I guess they are giving us the number of tickets sold based on the probabilities.
  17. hlsmith

    How to get amount to pay in order to be fair

    < $20K or < $50k. What do you think? Seems like homework, please show your own due diligence and we are happy to help.
  18. hlsmith


    The +/- value represent how many standard deviations the value is from the mean, 4 is a lot, < then 2 is to be expected for about 95% of observations, so not considered an outlier - traditionally.
  19. hlsmith

    Standards déviation

  20. hlsmith