Dear all,
I have the variable (Surgery duration) represented by values in minutes and the variable (preoperative neuropraxia cases) represented as (Yes/No) values. I want to compare these two variables and find out whether there is an association between presence of preoperative neuropraxia and the increase of surgery time. What test should i use in SPSS?
Thank you for your cooperation!
Kind regards,
Please have a look at this useful link below:
An overview of correlation measures between categorical and continuous variables
https://medium.com/@outside2SDs/an-overview-of-correlation-measures-between-categorical-and-continuous-variables-4c7f85610365
Since you have a 2-category independent variable and a continuous independent variable, you can use a t-Test.
You can simply compare the mean of the surgery duration between the two groups (Yes and No), using an unpaired t-test.
Since your categorical variable has two levels, and can be coded as 0 and 1, you can use standard correlation techniques. The standard approach would be point bi-serial correlation. But this is equivalent to common Pearson correlation, except that there may be differences as to whether the population standard deviation or the sample standard deviation is used (see: https://stats.stackexchange.com/questions/226157/p-value-for-point-biserial-correlation-in-r ). You could also use alternate correlation procedures like Spearman or Kendall if those are more appropriate.
If you want to designate one variable as the dependent variable, you could use e.g. t-test or Mann-Whitney test, or a whole host of different tests.
I assume it wouldn't make sense to think about the Yes/No variable as the dependent variable.
Duration of surgery strikes me as a variable that likely has some extreme outliers that could dramatically affect the group means. With that in mind, quantile regression might be preferable to a t-test. A QUANTILE REGRESSION command was added to SPSS in v27. (If you have an earlier version, you'll have to use the R-programmability feature instead.) The default setting for quantile regression will allow you to compare the median times for the two groups. But you can also compare the groups on any other quantiles of interest (e.g., the conditional 90th percentiles). The links below provide more info. HTH.
https://www.ibm.com/support/knowledgecenter/SSLVMB_sub/statistics_mainhelp_ddita/spss/regression/idh_quantile.html
https://www.ibm.com/support/knowledgecenter/SSLVMB_sub/statistics_reference_project_ddita/spss/advanced/syn_quantile_regression.html
- Badges
- Science topic
- Similar topics
- Mathematical Sciences
- Statistics
- Statistical Software
- SPSS