Statistics in a Modern World 800
Solutions to Practice Exam 2

1.  
   1. (iv) multiple boxplots (for values of one measurement variable compared for two categorical groups)
   2. (ii) piechart (for one categorical variable)
   3. (iii) scatterplot (for two measurement variables)
2.  
   1. (iii) skewed right/high outliers
   2. (iv) 23: mean greater than median (outliers pull mean up) but 32 would be way too high
   3. (iii) 6 is the typical distance of those values from their mean; 0.06 and 0.6 would be way too low and 60 would be way too high
3. (a) +1 because it is a strong positive relationship
4. (d) [(a) not appropriate because heights level off before 20, and the relationship is not at all linear; (b) not appropriate because living situation is categorical]
5.  
   1. 0.60(230)/82=$1.68
   2. (ii)$2.50 is too high because $1.68 would have kept pace with inflation
   3. -92+.0468(2013)=2.21
   4. extrapolation
6.  
   1. z=(7-12)/2=-2.5; the proportion below is 0.005, according to Table
   2. z=(14.8-12)/2=1.4; the proportion below is 0.92, so proportion above is 0.08.
   3. shortest 1% have z=-2.33, observed value = 12-2.33(2)=7.34
   4. longest 10% have z=+1.28, observed value = 12+1.28(2)=14.56
   5. (ii) Mean equals median because normal curves are symmetric.
   6. The curve is marked off with the following seven points: 6, 8, 10, 12, 14, 16, 18.
7.  
   1. (ii) because sales overall are increasing
   2. (iii)
   3. (iii)
8.  
   1. (7)
   2. 5 (slower metaolism is a common cause)
   3. 1 (TV is a direct cause of obesity
   4. 3 (TV is a contributing cause to obesity)
   5. 2 (causation in the opposite direction, with obesity causing children to watch TV)
   6. 6 (TV and weight both increasing over time)
   7. 4 (socio-economic status is a confounding variable)
9.  
   1. (ii)males: 61/120=0.51 as opposed to 89/240=0.37 for females
   2. calculate (column total * row total)/table total: 100, 50, 140, 70
   3. 1.2, 2.4, 0.9, 1.7
   4. 1.2+2.4+0.9+1.7=6.2
   5. (ii) because 6.2 > 3.84
   6. (ii): we have convincing statistical evidence of a relationship, but we can't draw definite conclusions about the reason for the relationship.

[ Home | Calendar | Assignments | Handouts ]