# Reporting a chi squared with 2 different sample sizes

#### Puriwai

##### New Member
Hi - odes anyone know how to report a chi square result with two different sample sizes. Is it N=27,29)= or N=27,N=29 ???
Thanks a bunch

#### katxt

##### Active Member
APA uses the total number as the sample size. N = 56

#### obh

##### Well-Known Member
PS what chi square test did you run?

#### hlsmith

##### Less is more. Stay pure. Stay poor.
Can you provide more details? Are those two n-values for the two dependent variable groups or independent? You should report both the overall and group values! Provide sufficient information so others know what you did and can reproduce your results.

#### Puriwai

##### New Member
Thanks for that - 2 independent groups.
This is a simple audit cycle - two groups of 30
27 adults in the Audit 1 group and 29 in Audit 2
When I am reporting the Chi-Square result - the adult result will br N=56, and overall results N=60 as katxt has pointed out ?
Is that correct ? It makes perfect sense
I am, as you no doubt have picked up, very new to biostats

#### katxt

##### Active Member
It's not sounding as much like a normal Chi square as I had imagined. Can you post the actual data?

#### Puriwai

##### New Member
Sure thing, for example:

Parameter : Waist circumference
Audit 1. Audit 2
N(%) N(%)
Children 3(100) 1(100)
Totals 14(47) 22(73)

Audit 1 has 27 adults and 3 children
Audit 2 has 29 adults and 1 child

chi-square (1,N=56)=6.70,p=.01, and overall by 26%, chi-square(1,N=60)=4.44,p=0.035.

Its the N=56 and N=60 I need to make sure is correct.

#### katxt

##### Active Member
I'm not entirely sure what is going on but this is what I have pieced together. Let me know if I have misinterpreted.
You are looking at two separate tests, Adults and Totals. (We will ignore the fact that these two tests use much the same data and so are not independent)
Let's look at the Totals first. We have two audits of 30 total each, a grand total of 60.
14 and 22 adhere, a total of 36.
If there was no difference you would expect 30/60x36 = 18 in each group.
Is a 14 22 split significantly different from a 18 18 split? This is your chi square test. X2 turns out to be 1.78 with a p value of 0.18, so X2(1, N=36) = 1.78, p = 0.18 not sig.

For the adults, 27 and 29 per group 56 total
32 adults adhered so expect 27/56x32 and 29/56x32 = 15.4 and 16.6 for the audits
We actually see 11 and 21. Is this significantly different from 15.4 and 16.6? Chi square test.
This one turns out X2(1, N=32) = 2.45, p = 0.12 not sig.

#### Puriwai

##### New Member
I must admit I am a little lost here - I am seeing a statistician tomorrow and will get back to you katxt - with thanks

#### katxt

##### Active Member
Great. Show them what I've written. kat