In: Statistics and Probability
An organization believes that the thought of adults ages 55 and over on which industry has the most trustworthy advertising is uniformly distributed. To test this claim, you randomly select 800 adults age 55 and over and ask each which industry has the most untrustworthy advertising. The results are shown in the table. At a = 0.05, can you reject the claim that the distribution is uniform ( ie, that each frequency should be the same - in this case, 160, since there are 800 total adults sampled?) Explain.
Auto Companies: 128
Fast Food Companies 192
Finanical Services Companies 112
Pharmaceutical Companies 152
Soft Drink Companies 216
Solution
The solution is based on the Chi-square Test for Goodness of Fit.
Goodness of Fit – Uniform distribution
Let Oi and Ei be respectively the observed and expected frequencies of the ith class, i = 1 to k, k being the number of classes given.
Claim: The distribution is uniform.
Hypotheses:
Null: H0: Observed frequencies are in accordance with the Uniform Distribution Vs
Alternative HA: H0 is false
Test Statistic:
χ2 = ∑[i = 1,k]{(Oi - Ei)2/Ei},
Under H0, Ei follow Uniform distribution and hence Ei’s are the same for all i. So, let Ei = E for all i.
So, χ2 = (1/E)∑[i = 1,k]Oi2 - N, where N = total frequency.
Details of calculations
i |
1 |
2 |
3 |
4 |
5 |
Total |
Oi |
128 |
192 |
112 |
152 |
216 |
800 |
Oi2 |
16384 |
36864 |
12544 |
23104 |
46656 |
135552 |
N |
800 |
|||||
k |
5 |
|||||
E |
160 |
|||||
χ2 |
47.2 |
|||||
s |
1 |
|||||
k - s |
4 |
|||||
α |
0.05 |
|||||
χ2crit |
9.487729037 |
|||||
p-value |
1.38543E-09 |
|||||
Decision |
Reject Ho |
Distribution, Significance Level, α, Critical Value, p-value
Under H0, χ2 ~ χ2k – s, Chi-square distribution with degrees of freedom = k – s, where k = number of classes and
s =number of parameters estimated.
p-value = P(χ2k – s > χ2cal)
Given significance level = α , critical value = χ2crit = upper α% of χ2k - s,
Critical value and p-value obtained using Excel Function: Statistical CHIINV and CHIDIST are as shown in the above table.
Decision
Since, χ2cal > χ2crit, or equivalently, since p-value < α, H0 is rejected.
Conclusion
There is not sufficient statistical evidence to conclude that the claim is valid and hence we conclude that
the thought of adults ages 55 and over on which industry has the most trustworthy advertising depends on the industry. Answer
DONE