In: Statistics and Probability
An advertising agency is trying to predict the percent of people who want certain colors of nerds. They are claiming that 35% of the people like blue nerds, 45% would like red nerds and 20% would like mixed nerds. A random sample of 500 people revealed that 250 wanted blue, 200 wanted red, and 50 wanted mixed. Does this data suggest that the percentages on which the staffing is correct? Use a level of significance of .05 to test the claim.
null hypothesis:Ho: Sample distribution of colors is similar to as stated by agency .
Alternate hypothesis:Ha: Sample distribution of colors is different then as stated by agency .
degree of freedom =categories-1= | 2 |
for 2 df and 0.05 level of signifcance critical region χ2= | 5.991 |
Applying chi square goodness of fit test:
relative | observed | Expected | residual | Chi square | |
category | frequency | Oi | Ei=total*p | R2i=(Oi-Ei)/√Ei | R2i=(Oi-Ei)2/Ei |
Blue | 0.35 | 250 | 175.00 | 5.67 | 32.143 |
red | 0.45 | 200 | 225.00 | -1.67 | 2.778 |
Mixed | 0.20 | 50 | 100.00 | -5.00 | 25.000 |
total | 1.000 | 500 | 500 | 59.921 |
as test statistic X2 =59.921 is in critical region we reject null hypothesis
we have sufficient evidence at 0.05 level to conclude that Sample distribution of colors is different then as stated by agency .