statistics - Subsetting data for a chi-squared or z test -


variables involved in analysis:

bct

  1. yes
  2. no

insurance status

  1. uninsured
  2. insured

er status

  1. positive
  2. negative

null hypothesis: there no difference in proportions between 2 groups (p1 = p2)

what have defined 2 groups?

group 1: bct = yes, er = positive, insurance status = uninsured

group 2: bct = yes, er = positive, insurance status = insured

if you'll notice in image included analysis, did chi squared , z test on same categorical variables, in z test, further subset variable, bct, include yes values, resulted in 2x1 table. whereas in chi squared, included bct yes , no values resulted in 2x2 table. of course, ended getting different p-values these 2 tests.

so question is: better further subset data as possible, did in z test, or 2x2 table created in chi squared test preferred? 1 more accurate?

and guess should ask, should doing subsetting @ all?? reason being because have made table this, without excluding of categories variables:

                      insurance status bct & er status           uninsured              insured  bct yes & er negative        n                      n bct yes & er positive        n                      n bct no & er negative         n                      n bct no & er positive         n                      n 

i totally confused on way right way evaluate hypothesis. appreciated.


Comments

Popular posts from this blog

inversion of control - Autofac named registration constructor injection -

verilog - Systemverilog dynamic casting issues -

ios - Change Storyboard View using Seague -