Example of Goodness-of-Fit Test for Poisson

A quality engineer at a consumer electronics company wants to know whether the defects per television set are from a Poisson distribution. The engineer randomly selects 300 televisions and records the number of defects per television.

  1. Open the sample data, TelevisionDefects.MTW.
  2. Choose Stat > Basic Statistics > Goodness-of-Fit Test for Poisson.
  3. In Variable, enter Defects.
  4. In Frequency variable: (optional), enter Observed.
  5. Click OK.

Interpret the results

The null hypothesis states that the data follow a Poisson distribution. Because the p-value is 0.000, which is less than the significance level of 0.05, the engineer rejects the null hypothesis and concludes that the data do not follow a Poisson distribution. The graphs indicate that the difference between the observed and expected values is large for categories 1 and 3, and that category 3 is the highest contributor to the chi-square statistic.

Method

Frequencies in Observed

Descriptive Statistics

NMean
3000.536667

Observed and Expected Counts for Defects

DefectsPoisson
Probability
Observed
Count
Expected
Count
Contribution
to Chi-Square
00.584694213175.4088.056
10.3137864194.13629.993
20.0841991825.2602.086
>=30.017321285.196100.072

Chi-Square Test

Null hypothesisH₀: Data follow a Poisson distribution
Alternative hypothesisH₁: Data do not follow a Poisson distribution
DFChi-SquareP-Value
2140.2080.000