Saturday, October 20, 2018

Analyzing Chi-Square Output (SPSS)

In prior articles, we discussed how to generate output within the SPSS platform as it pertains to the chi-squared methodology. The purpose of this entry, is to answer inquires which I have received related to the chi-squared analysis. Specifically, how to properly assess Risk Estimate and Cross-Tabulation tables.

To aid in the assessment of these output types, I have created the following charts and corresponding keys. Though the data which was utilized to create these tables is fictional, the fundamental aspects of the charts remain un-impacted.

Cross-Tabulation Chart and Key

a = Individuals who smoked and received a cancer diagnosis.

b = Individuals who smoked and did not receive a cancer diagnosis.

c = Total number of individuals who were smokers.

d = Percentage of individual who were smokers and received a cancer diagnosis.

e = Percentage of individuals who were smokers and did not received a cancer diagnosis.

f = Total percentage of individual smokers.

g = Individuals who did not smoke and received a cancer diagnosis.

h = Individuals who did not smoke and did not receive a cancer diagnosis.

I = Total number of individuals who were not smokers.

j = Percentage of individuals who were not smokers and received a cancer diagnosis.

k = Percentage of individuals who were not smokers and did not receive a cancer diagnosis.

l = Total percentage of individual non-smokers.

m = Total number of individuals diagnosed with cancer.

n = Total number of individuals not diagnosed with cancer.

o = Total number of individuals surveyed.

p = Percentage of total surveyed individuals who were diagnosed with cancer.

q = Percentage of total surveyed individuals who were not diagnosed with cancer.

r = Total percentage of surveyed individuals.

Risk Estimate Chart and Key

a = The odds ratio indicates that the odds of finding cancer within an individual who smokes, as compared to an individual who does not smoke, is 9.333.

b = The outcome of this event (Cancer Diagnosis) was 2.667 times more likely to occur within the smoker group.

c = The number of total individuals surveyed.

Calculating Relative Outcome: | 1 – risk estimate value | * 100

| 1 – 2.667 | * 100 = 167

Risk ratios indicated that the risk of the outcome variable (cancer), within the category of smokers, increased by 167 % relative to the group of non-smokers.