When collecting and comparing data, is the variation of the "expected" data, compared to the "observed" data due to chance or is it due to one of the variables you’re actually using in the test?
The ChiSquare QlikView App is based on a "coin flipping" experiment shown in the YouTube video, https://www.youtube.com/watch?v=WXPBoFDqNVk
Coin Flipping Experiment:
After flipping a coin 50 times, there was an observed result of 28 heads and 22 tails. Was the result due to chance, is there something wrong with the coin, or a possible an error in the way the coin is being flipped. The Chi-Square test allows you to answer those questions.
Null Hypothesis – There is no difference between the observed and expected frequency results
The YouTube video goes on to explain the null hypothesis, degrees of freedom, Chi-Square value (0.72) and critical values (3.841)
ChiSquare QlikView App
Uses the INPUT DATA, and computes the
LOAD * Inline [
Gender, Coin, Observed, Expected
Male, Heads, 28, 25
Male, Tails, 22, 25
The whole point of the Chi-Square test is to accept or reject the null hypothesis. You have to either exceed or not exceed the critical value. Because the Chi-Square value (0.72) is lower than the critical value (3.841) you will accept the null hypothesis.
The ChiSquare QlikView App will provide an example of how to create the Chi2-Test using the Statistics Chart Wizard.