This preview has intentionally blurred sections. Sign up to view the full version.
View Full DocumentThis preview has intentionally blurred sections. Sign up to view the full version.
View Full Document
Unformatted text preview: 1. A correlation of —1 implies a perfect linear relationship between the two variables. T F 2. In adding a constant to each value in a bivariate data set, the correlation would be unchanged. T F 3. I draw 100 tickets with replacement from a box whose tickets are num bered. The average of the numbers on the tickets in the box is 0 and the standard deviation is 10. There is approximately a probability of 0.68 that the average of the 100 tickets will be between —1 and + 1. T F 4. Observational studies usually do not permit us to attribute cause. If the number of subjects is large enough, however, we can be 95 per cent certain (or more) of causality. T F 5. The major advantage of controlled experiments is that we randomize subjects to treatments. This randomization means that in the long run, the in¡uence of confounding variables tends to be balanced equally between the treatment groups. T F 6. A Type I error is committed when a null hypothesis is rejected when it is false. T F 7. To test H : μ = 100 against H 1 : μ < 100 one would use a two—tailed test. T F 1 8. Suppose you can choose to toss a fair coin either 10 or 100 times. Various plans are described below. Circle your preferred number of tosses in each of the following cases. (i) You win $1 each time a head occurs (and lose nothing if a tails occurs). 10 tosses 100 tosses (ii) You win $1 if the proportion of heads is larger than 0.8 (and lose nothing otherwise). 10 tosses 100 tosses (iii) You win $1 if the proportion of heads is between 0.4 and 0.6 (and lose nothing otherwise). 10 tosses 100 tosses 9. Increasing the signiFcance level of a hypothesis test (circle one of the following) (a) does not change the Type I error probability. (b) does not change the Type II error probability. (c) increases the Type I error probability. (d) decreases the Type I error probability. (e) can either increase or decrease the Type I error probability. 2 10. The point estimate of the mean of a distribution which has lowest variance is the sample (a) mean (b) standard deviation (c) mode (d) interquartile range (e) median 11. Based on a sample of 30 from a Normal distribution, a 95% conFdence interval for the mean was computed. Using the same sample, a 99% conFdence interval would be (a) the same width (b) possibly narrower, possibly wider (c) impossible to determine from such a sample (d) narrower (e) wider 12. A professor gives his class a quiz consisting of 10 true—false questions. Students get one point for a correct answer and 0 for an incorrect answer. There are eight students in the class. Their scores are: 5, 6, 6, 7, 7, 7, 8, 10. The average of these is 7. The standard deviation is 1.51....
View
Full Document
 Spring '08
 KARIM
 Correlation, Standard Deviation, Null hypothesis, graduate student

Click to edit the document details