In: Statistics and Probability
1)Background: Morris Saldov conducted a study
in Eastern and Central Newfoundland in 1988 to examine public
attitudes towards social spending. In particular, the study tried
to determine if knowing someone on public assistance (yes, no)
affected one's views on social spending (too little, about right,
too much). The data from the study is summarized in the table
below.
Yes | No | Total | |
Too little | 38 | 5 | 43 |
About right | 15 | 15 | 30 |
Too much | 10 | 5 | 15 |
Total | 63 | 25 | 88 |
Source: Morris Saldov, Public Attitudes to Social Spending in Newfoundland," Canadian Review of Social Policy, 26, November 1990, pages 10-14.
Directions: Conduct a chi-square test for independence to determine if the association between knowing someone on public assistance and views on social spending is statistically significant.
Yes | No | |
Too little | ||
About right | ||
Too much |
2) A financial analyst claims that 19% make all purchases with cash, 17% make most purchases with cash, 20% make half of purchases with cash, 33% make some purchases with cash and 11% make no purchases with cash. You take a random selection to see if you can conclude that the distribution is different than what the financial analyst claims. Use a 5% significance to decide and round to the fourth.
Categories | Observed Frequency |
Expected Frequency |
---|---|---|
All Cash | 28 | |
Most Cash | 84 | |
Half Cash | 103 | |
Some Cash | 159 | |
No Cash | 94 |
Test Statistic:
Degrees of Freedom:
p-val:
Decision Rule: Select an answer Reject the Null Accept the Null
Fail to Reject the Null
Did something significant happen? Select an answer Nothing
Significant Happened Significance Happened
There Select an answer is not is enough evidence to
conclude Select an answer that the distribution is what the
financial analyst claims that the distribution is different than
what the financial analyst claims
select one - that the is what the financial analyst is claims or is different
H0: There is no association between knowing someone on public assistance and views on social spending.
Ha There is an association between knowing someone on public assistance and views on social spending.
Applying chi square test of independence: |
Expected | Ei=row total*column total/grand total | Yes | No | Total |
too little | 30.7841 | 12.2159 | 43.00 | |
About right | 21.4773 | 8.5227 | 30.00 | |
too much | 10.7386 | 4.2614 | 15.00 | |
total | 63.00 | 25.00 | 88.00 | |
chi square χ2 | =(Oi-Ei)2/Ei | Yes | No | Total |
too little | 1.691 | 4.262 | 5.9539 | |
About right | 1.953 | 4.923 | 6.8762 | |
too much | 0.051 | 0.128 | 0.1788 | |
total | 3.6957 | 9.3132 | 13.009 | |
test statistic X2 = | 13.009 |
degree of freedom(df) =(rows-1)*(columns-1)= | 2 |
p value = | 0.0015 |
The p-value provides strong evidence against the null hypothesis. The association between knowing someone on public assistance and attitudes towards social spending is statistically significant.
2)
applying chi square goodness of fit test: |
relative | observed | Expected | residual | Chi square | |
category | frequency(p) | Oi | Ei=total*p | R2i=(Oi-Ei)/√Ei | R2i=(Oi-Ei)2/Ei |
1 | 0.19 | 28.00 | 88.92 | -6.46 | 41.737 |
2 | 0.17 | 84.00 | 79.56 | 0.50 | 0.248 |
3 | 0.20 | 103.00 | 93.60 | 0.97 | 0.944 |
4 | 0.33 | 159.00 | 154.44 | 0.37 | 0.135 |
5 | 0.11 | 94.00 | 51.48 | 5.93 | 35.119 |
total | 1.000 | 468 | 468 | 78.1828 | |
test statistic X2 = | 78.183 |
degree of freedom =categories-1= | 4 |
p value = | 0.0000 |
Significance Happened
There is enough evidence to conclude is different than what the financial analyst claims