Recent content by hlsmith

  1. hlsmith

    Simple anova?

    I suppose, but that is a lot comparisons. What are you hypothesizing, linear increase/decrease?
  2. hlsmith

    Statistical Validation of a Labelling Method

    Sounds like you need to start off by just doing a straight validation study. I know you have limited adjudication resources, but you need to have all adjudicators review the same observations. This means all experts and non-experts. Doing this will let you understand inter-reliability between...
  3. hlsmith

    Correlation between two set of variables

    What is the purpose of the project and what do the two datasets represent (e.g., two subsets)?
  4. hlsmith

    Advice please: statistical testing a number of proportions from two samples

    I am thinking this is just some thing like a contingency table, columns say are males and females while rows are favorite ninja turtles. Would this be correct? It is hard to process without context.
  5. hlsmith

    Can you use correlation for percentages?

    Depends on what you are using this for. If it is just for a general in-house analysis (that you are not publishing) - yeah a spearman correlation may give you a general idea of the relationship. I would also strongly recommend creating a scatterplot with these data. If you want to do something...
  6. hlsmith

    Which competing risks survival analysis to choose

    Another issue you may run into, it seems, could be a "Well defined Treatment". It gets difficult to establish causality/association when you have multiple version of treatments (dosing) and incomplete compliance. I am slightly familiar with the types of models you are writing about, but not...
  7. hlsmith

    More predictors than observations?

    Look up "sparsity" in data or "curse of dimensionality. The issue is that you have wider than long data, so possibly more descriptors than you can have adequately represented within samples. Extreme example, say you have three observations and five predictors (age, gender, race, employment...
  8. hlsmith

    which type of regression analysis to use

    It is all about writing out your study questions and defining how variables are formatted. Everything else falls into place. State those and we can better help.
  9. hlsmith

    Logistic Regression Models Without Main Effects?

    I havent really figured it out but I have heard some experts in conferences saying in their life they haven't ever seen a real interaction. I don't know if they mean the interaction is just a proxy of something else.
  10. hlsmith

    Hi everyone.

    Welcome aboard Kostis. We look forward to your questions and contributions! :)
  11. hlsmith

    Logistic Regression Models Without Main Effects?

    Standard practice is to include main effects even if you don't think you care about them. The relationship between the variables is conditional. Side note, Andrew Gelman posted yesterday on his forum that you need, I think, 16 times more data if you want to include an interaction term in the...
  12. hlsmith

    Sample Size Affecting my Estimates

    Do you just have an over-powered analysis? Meaning you can find statistical significance, even when results aren't really contextually significant (e.g., small effect sizes). If you run a model on one subsample and then the same model on the other subsample are the estimates pretty much...
  13. hlsmith

    Sample Size Affecting my Estimates

    Can you give us some scope? Number of variables, full sample size, What is the outcome and is it balanced?
  14. hlsmith

    convert odds ratio to logistic coefficient

    It is traditional to present stratified estimates for the two groups. You should also plug numbers into the model and plot the slopes of years by category, you are looking for non-parallel lines (disordinal).
  15. hlsmith

    convert odds ratio to logistic coefficient

    Well, it is pretty straightforward - there is a bigger impact on the odds of the outcome for the continuous variable given which category you are in for the categorical variable. So if you have a gene and an environmental exposure, people with the gene have a greater odds of the outcome when...