In: Statistics and Probability
Annual inflation in Canada has been low for several years.
Suppose a provincial survey is conducted by a polling organization,
and Saskatchewanians answered the question: “are you not at all
concerned, somewhat concerned, or very concerned that inflation
will climb?”. The results are given below.
Not at all concerned |
Somewhat concerned |
Very concerned |
120 |
360 |
230 |
The same survey was conducted in Ontario, and it was found that 20%
of Ontarians were very concerned, 45% were somewhat concerned, and
35% were not at all concerned. The polling organization is
interested in determining if the proportions of people who are very
concerned, somewhat concerned, or not at all concerned about
inflation rising are the same in Saskatchewan as they are in
Ontario. Conduct an appropriate hypothesis test by filling in the
blanks below.
1) State the null and alternative hypotheses [1 mark]:
2) Calculate the test statistic. Enter your value below. Note:
you may use Excel to calculate the test statistic, and it doesn’t
matter if you round for each individual calculation or just at the
end, you’ll get the same value. No need to show your work.
[1 mark]
3) The critical value is from a chi-squared distribution on 2 degrees of freedom, which is 5.99147.
4) Based on this critical value, do you reject the null
hypothesis (Yes or No)? [1 mark]
Observed (Oi ) | Expected ( Ei) | Oi-Ei | (Oi-Ei)^2 | (Oi-Ei)^2/Ei |
0.169 | 0.35 | -0.18 | 0.0328 | 0.0937 |
0.507 | 0.45 | 0.057 | 0.0032 | 0.0071 |
0.3239 | 0.2 | 0.124 | 0.0154 | 0.077 |
observed freq are Oi
not at all concerned = 120/((120+360+230)) = 0.169
somewhat concerned = 360/((120+360+230)) = 0.507
very concerned= 230/((120+360+230)) = 0.324
------------------------------------------------------------------
expected frequencies are Ei
20% were very concerned
45% were somewhat concerned
35% were not at all concerned
------------------------------------------------------------------
and claiming hypothesis is
null, Ho: no association exists b/w the proportions of people who
are very concerned
alternative, H1: association exists b/w the proportions of people
who are very concerned
level of significance, alpha = 0.05
from standard normal table, chi square value at right tailed, ᴪ^2
alpha/2 =5.9915
since our test is right tailed,reject Ho when ᴪ^2 o >
5.9915
we use test statistic ᴪ^2 o = Σ(Oi-Ei)^2/Ei
from the table take sum of (Oi-Ei)^2/Ei, we get ᴪ^2 o =
0.1778
critical value
the value of |ᴪ^2 alpha| at los 0.05 with d.f, n - 1 = 3 - 1 = 2 is
5.9915
we got | ᴪ^2| =0.1778 & | ᴪ^2 alpha | =5.9915
make decision
hence value of | ᴪ^2 o | < | ᴪ^2 alpha | and here we do not
reject Ho
----------------------------------------------------------------------------------------
null, Ho: no association exists b/w the proportions of people who
are very concerned
alternative, H1: exists association exists b/w the proportions of
people who are very concerned
test statistic: 0.1778
critical value: 5.9915
decision: do not reject Ho