Search results

  1. hlsmith

    Logistic regression for prediction of outcome?

    I would guess they have a pool of patients with binary defined plural effusion and want model predictors. Per @ondansetron how is outcome defined and formatted.
  2. hlsmith

    [meta] Forest plot does not give subgroup total N

    For a follow-up on my end, I will not likely get around to playing with the code until next Tuesday, pretty busy right now. If I figure anything out I will post at that time. Thanks.
  3. hlsmith

    Effect size in regression

    You can also post a link to the paper. That may help others as well.
  4. hlsmith

    Bayesian inference project

    Well we can't create something for you. It would be like me saying hey I want to get into building things with a hammer. I have supplies. You have no idea what supplies I have, so making a suggestion would be fruitless. I would say, come up with your research question which can be supported by...
  5. hlsmith

    Effect size in regression

    I could not tell from that snippet, but it does seem as though they are trying to get at a standardized coefficient, which seem popular in the social sciences since you can try to rank the coefficients by it. Not well versed in the methods.
  6. hlsmith

    Predictors insignificant & F-stat is highly significant!!

    How about you just post your output so we can see all the details. Thanks.
  7. hlsmith

    sample size and type 2 error

    Type II error, failing to reject the null when it is false. Perhaps your sample is not large enough and has a lot of variability. As the sample size increases it gets closer to the target population. Thus excess variability due to sampling error gets smaller. More over the SE gets small given it...
  8. hlsmith

    How to compare differences between groups in predictors of an outcome variable?

    You have to understand the relationship between the covariates, you can't just use the results of the table since the variables may be multicollinear or interact with each other. You have to use your content expertise when selecting variable for the model.
  9. hlsmith

    How to interpret the intercept in a regression model which has a categorical covariate?

    Represents female, the reference group. You also have to rationalize in your mind such a person does not exist, since it is impossible to have a wt, age, ht of '0'. Many people will standardized the continuous variables before entering them into the model, which would represent the mean of the...
  10. hlsmith

    How to compare differences between groups in predictors of an outcome variable?

    I would start by just making a table with groups as the columns and rows as the rows. I would then fill in the table with descriptive statistics. That may be good enough. since you will be reporting the differences between the groups in a table of the model results. Some times people run too...
  11. hlsmith

    [meta] Forest plot does not give subgroup total N

    Can you post your code please. I can't remember if I have use 'meta' before or just 'metafor' - I think.
  12. hlsmith

    Bayesian inference project

    You have to be more specific than that, do you have data?
  13. hlsmith

    Cohen d greater than one

    Is there correlation between the other genes? I wonder if naïve Bayes may work with your data? A simple approach may be to just calculate the probability of the genes with the outcome and provide 99% confidence intervals. Bigger picture you could use logistic regression, but you would want...
  14. hlsmith

    Binary Logistic Regression Questions

    No. What do you mean by "fixed factor"? Also, if the covariates you are controlling for are continuous, don't let SPSS call them categorical. Figure out what your issue is, it might be that they are categorical. Also, provide a snippet of the dataframe, so we can see what you are working with...
  15. hlsmith

    Binary Logistic Regression Questions

    Please define what you mean by "fixed factor". Do you mean fixed effect, meaning it is not a random effect?
  16. hlsmith

    is regression analysis possible for this dataset?

    Yes, 'all' models are wrong unless you are doing a data simulation or fixed experiment and know the true underlying data generating function. This is further affected given most things are probabilistic not deterministic, due a lack of precision in data collection, etc. Furthermore, if a...
  17. hlsmith

    Help with Regression analysis

    Hey @mrlonuk79 The presentation of your inquiries/question is very poor. It is difficult to understand your concerns and the context of the project. Please provide more information in a clear manner and we would be happy to help you. Tell use about your variables and the outcome as well as...
  18. hlsmith

    Cohen d greater than one

    Well if you tell us more about the background context we can be of more help to you. One of the better effect estimates is to use risk differences, but you do not provide enough data for use to make particular recommendations given models will have assumptions.
  19. hlsmith

    Large sample size but lots of missing data for ONE co-variate

    Do you know why data is missing? Can you classify missingness as MCAR, MAR, NMAR?
  20. hlsmith

    How does p-values influence prediction result?

    Well you have already tried one approach, so your results are now conditional on that one failing your expectations. I am just mentioning this since given any purpose, fitting multiple models can contribute to false discovery. You should run a least absolute shrinkage selection operator...