Applying the Chi-Square Goodness of Fit Test for Frequency Data Analysis

11-2 Goodness of Fit Test In This section we consider sample data consisting of observed frequency counts arranged in a single row or column (called a one-way frequency table). We will use a hypothesis test for the claim that the observed frequency counts agree with some claimed distribution, so that there is a good fit of the observed data with the claimed distribution. A goodness-of-fit test is used to test the hypothesis that an observed frequency distribution fits (or conforms to) some claimed distribution. H0: The random variable follows a particular distribution. H1: The random variable does not follow the distribution specified in H0. Ex 1) Consider the observed frequencies and relative frequencies of browser preference form a survey of 200 Internet users. Browser Observed frequency Relative frequency Microsoft Internet Explorer 140 0.785 Firefox 40 0.15 Safari/other 20 0.065 The following model shows how the market shares are distributed in the null hypothesis: H0: PMs IE= 0.785 , Pfirefox=0.15, PSafari/other= 0.065 H1: The random variable does not follow the distribution specified in H0. How a Goodness- of -Fit Test Works The goodness -of -fit test is based on a comparison of the observed frequencies (actual data from the field) with the expected frequencies when H0 is true. That is, we compare what we actually see with what would expect to see if H0 were true. If the difference between the observed and expected frequencies is large, we reject H0. As usual, it comes down to how large a difference is large. The hypothesis we conduct to answer this question relies on χ2 distribution. Performing the χ2 Goodness of Fit Test The following conditions must be met:  The data have been randomly selected.  The sample data consist of frequency counts for each of the different categories.  None of the expected frequencies is less than 1.  For each category, the expected frequency is at least 5. Finding Expected Frequencies The expected frequency for a category is the frequency that would occur if the data actually have the distribution that is being claimed. For the ith category, the expected frequency is Ei = n . pi, where n represent the number of trials and pi represents the population proportion for the ith category. If we assume that all expected frequencies are equal, then each expected frequency is E = n/k, where n is the total number of observations and k is the number of categories. The χ2 goodness of fit test may be performed using (a) the critical value, and (b) the p-value method. (a) χ2 goodness of fit test. (Critical value method) Step 1: State the hypotheses and check the conditions. The null hypothesis is states that the qualitative random variable follows a particular distribution. The alternative hypothesis states that the random variable does not follow that distribution. Step 2: Find the χ2 critical value, χ2 critical, from table A-4 by using k -1 degrees of freedom, where k is the number of categories. Note, Goodness-of-fit hypothesis are always right tailed. And state the rejection rule. Reject if χ2data > χ2 critical. Step 3: Find the test statistic χ2data. χ2data= 𝑂𝑖 −𝐸𝑖 2 𝐸𝑖 Where Oi = observed frequency, and Ei = expected frequency. Step 4: State the conclusion and the interpretation. Ex 2) Perform the hypothesis test shown in example 1, use 0.05 as a significance level. H0: PMs IE= 0.785 , Pfirefox=0.15, PSafari/other= 0.065 H1: The random variable does not follow the distribution specified in H0. Browser Observed freq Category Oi Relative freq pi __________________________________________________________________ MS IE 140 0.785 Firefox 40 0.15 Safari/other 20 0.065 Interpretation: There is evidence that the random variable browser does not follow the distribution in H0. In other words, there is evidence that the market shares for internet browsers have changed. Note carefully what this conclusion says and what it doesn’t say. The χ2goodness of fit test shows that there is evidence that the random variable does not follow the distribution specified in H0. In particular, the conclusion does not state, for example, that Firefox’s proportion is significantly greater. (b) χ2 goodness of fit test: (p-value Method) Step 1: State the hypotheses and check the conditions. Step 2: Find the test statistic χ2data. χ2data= 𝑂𝑖 −𝐸𝑖 2 𝐸𝑖 Where Oi = observed frequency, and Ei = expected frequency Step 3: Find the p-value. p-value = P (χ2 > χ2 data) Step 4: State the conclusion and the interpretation. Ex (3) The following tables show figures on the market share of cable modem, DSL, and wireless broadband from a 2002 survey and a 2006 survey which was based on a random sample of 1000 home broadband users. Test whether the population proportions have changed since 2002, using the p-value method, and level of significance is 0.05. 2002 broadband adoption survey Cable modem 67% DSL 28% Wireless/other 5% 2006 broadband adoption survey Cable modem 410 ` DSL Wireless/other 500 90