A researcher wants to examine the reliability of a new test of knowledge about diseases. The researcher administers the test to 50 randomly selected medical students and then lists the number of correct responses to odd questions and the number of correct responses to even questions. The researcher prepares for a Pearson r correlation. 

Reread Brown on checking reliability for single-test administrations.

Data file: corstudy4.sav
Variables: STU, ODD, EVEN
(Data are in dual columns in order to save space.)

STU    ODD    EVEN              STU     ODD    EVEN
1       23      11              26       21      17
2       21      22              27       25      18
3       13      23              28       10      19
4       17      24              29       19      21
5       18      25              30       20      24
6       21      11              31       19       7
7       22       9              32       17      11
8       10      22              33       16      10
9        8      23              34        4       7
10      12      25              35       12      15
11       7      10              36       11       4
12      22       4              37       16      17
13      25      12              38       15      21
14      24      14              39       13      22
15      11      18              40       19       1
16      14       8              41       20      25
17      13      12              42       21      24
18      18      19              43       22      20
19      21      20              44       25      19
20      20      21              45        4      18
21      19       6              46       12       6
22       3       1              47        9       9
23      11      19              48       10      10
24      16      20              49        7      17
25      17      22              50       25      18
 

Answer the following questions:

1. How can the researcher determine the reliability of the test?
    
    
    
2. The researcher wants to run a Pearson r statistic on these data. Do the data meet the required assumptions of the Pearson r? Given the purpose of the analysis, is the Pearson r enough?
    
    
3.  Is this problem one-tailed or two-tailed? Explain.
    
    
    
    
4. If permissible, calculate the Pearson r. If necessary, apply any additional statistic before you write the result below. Reread Brown's discussion of split-half reliability before you write.
    
    
    
5. Is this result significant: and, if so, what is the probability that it occurred by accident? Check table in Brown, p. 140.
    
    
    
6. Is the test reliable? This will relate to the strength of the correlation.