Geog 301a

Fall 2005

Home

Quiz 9
Perform a bivariate regression on the following data (17 marks)
 

Calculate the regression variance (3 marks)
 

Quiz 8

Determine the correlation coefficient between a river’s length and its discharge. (15) marks
Test the significance of r. (5 marks)

Extra Credit (2 marks)

What is the Spearman’s correlation coefficient for the above data?

Quiz Seven

There are 4 neighbourhoods sampled with housing prices recorded in each. Determine if there is a difference in the mean housing price of homes in the neighbourhoods.
Provide the hypotheses and all relevant calculations. Use α=.05.
 

Quiz Six

Geography 301a, 2005

1. Complete the following table if it is known that there are 3 samples with sizes 5, 7 and 7 respectively. (10 points)
 

What is your conclusion about differences between the means?
Critical value is 3.63 so we conclude there in no differences between the means

2. Analysis of variance results are not affected in each of the following cases, please explain why this is true in each case. (2 points each)

a) The same constant is added to every sample score.
Variance is a measure of spread of the data so adding a constant to each value will not change that
b) The order of the samples is changed.
The order of the samples isn’t used in the calculations so its irrelevant

3. What was the case study about in the significance test video? (4 points)
Shakespeare

4. Circle the letter of the correct formula for straight line distance. (2 points)
equation b is correct

Quiz 5
Geography 301a, 2005

1. What is Simpson’s paradox? (4 marks)
The direction of an association can be reversed by a lurking variable

2. What product was tested in the video on t tests with the pairwise t test?(2 marks)
Nutrasweet in pop

3.Determine whether the point pattern in the map below is random or not (14 points).
Use a t-test.

conclusion not random

Quiz 4
Geography 301a, 2005


1. A pair of cases is concordant if the value of each variable is larger (or smaller) for one case than for the other case.
Variable 1 Variable 2
 

For each of the pairs of cases identify which are concordant if any , which are disconcordant if any and which are tied, if any. (3 marks)

cases 1,2 ____________
cases 1,3 ____________
cases 2,3____________
Case 1,2 is concordant
case 1,3 is concordant
case 2,3 is tied

2. Determine whether the 3 colored dice are fair using chi square (6 marks)
Using theory o-e sq o-e sq/e
 

3. If there are 9 concordant pairs and 5 disconcordant pairs what is the value of Goodman & Kruskal’s Gamma? (4 marks)

4. Why is the use of a paired t-test preferable to a regular t-test when comparing the means of paired comparisons? (4 marks)
Since the same objects or persons are used in before and after much of the error is reduced since the characteristics of the items are help constant. The key was to talk about a reduction of error

5. What are the 3 types of t-tests?
1 sample, 2 sample, and paired

Quiz 3 Name_____________________________

1. Using the data from the last year’s class quizzes determine lambda. What can you conclude about the relationship between the 2 quizzes (14 marks)

E1 pick row 3 with value of 16, so error is 50-16=34
E2 col 1 pick row 4 with value of 4 so error is 9-4=5
E2 col 2 pick row 3 with value of 7 so error is 19-7=12
E2 col 3 pick row 2 with value of 7 so error is 17-7=10
E2 col 4 pick row 3 or row 4 with value of 2 so error is 5-2=3
sum of errors for E2 = 5+12+10+3=30

so lambda = (E1-E2)/E2=(34-30)/34=4/34=.12
 

2. Of the 3 migration models presented in class, which of the distance, population and simplified gravity model predictions were not statistically different than the actual migration data?
All of them were statistically different so the answer is none.  (2 marks)

3. You throw a 6 sided die 120 times and you record the number of times you get each number.

The table of observed values is:
 

If you wanted to do a chi square test of the fairness of the die what would be your expected values? (4 marks)
20 for each value, so a uniform distribution
 

Quiz 2 Name_____________________________

1. What is the likely effect of increasing the number of categories in a chi-square test on the probability of finding a significant difference?
As you increase the number of categories it will become harder to find a difference. (3 marks)

2. If you have a mark that is exactly plus one standard deviation from the mean and the class has 150 students in it, how many students received a lower mark than you?(4 marks)
using the rule for a normal distribution we have 68% of students within 1 std dev. So 150*.68=102 students with scores within 1 std dev
since you are exactly plus one std dev you are the 51st student out from the mean
since the question asks only about lower there are all the students left of the mean or 75 students plus the 50 below you that are above the mean so there are 125 students who have a grade lower than you

3. Given the following data, which of the 2 samples has the greatest dispersion. You must support your decision by using the coefficient of variation. (3 marks)
 

4. If we assume that sample 1 is our observed distribution and sample 2 is our expected distribution, is sample one significantly different that sample 2? (6 marks)

 

5. What is the effect of indeterminacy in classification into categories when using chi-square? (3 marks)
It can create a situation where when forced to place all items into a set number of classes you can end up with a table that is not truly reflective of the data. The video gave the case of Mendel and how his results were thought by some to be too good to be true


6. Why is knowledge of the mean so important in defining a normal distribution?(1 marks)
Its one of the 2 parameters required to define it

Quiz 1: Geography 301a, 2005

1. What are the 3 steps in analysis that the video suggested?
1. Producing data
2. Describing data
3. Conclusion from data (3 marks)

2. Name 6 of the case studies provided in the video.
Lightning strikes
human growth hormone
Chesapeake Bay pollution
baseball players salaries
potato chips
manatees
heart attacks
space shuttle
gambling
Hispanic FBI agents
batteries
Salem witchcraft trials
Shakespeare’s poem
welfare mothers
children’s creativity (6 marks)

3. Provide 2 of the 5 preconditions for statistics to be valuable.
1. Can only give answers if the data collection and the data collected allow such answers
2. User is aware the statistics is just another strategy for finding, patterns in the data
3. Statistics are based on certain assumptions If those assumptions are not true the technique can still be applied but significance tests must treated with caution
4. User is aware that techniques are mathematical models. Reality in all its complexity cannot be modeled in a useful way. Complex models may imitate reality but they will be equally complex and therefore not useful. Summarizing data in a complex way is not a step forward.
5. Data exploration needs to be done before any higher level modeling

(2 marks)

4. What is the preferred method for dealing with missing data?
Listwise deletion (1 mark)

5. What is the weighted mean of following numbers: (please remember false precision!).

datum weight
1.5         50
50.333     1
6.02         3
22.1         2
 

weighted mean=3.35 (8 marks)