Part 1: Terminology 
Part 2: Procedures 
Part 3: Test Interpretation 
Part 4: Assessment Ethics 
INSTRUCTIONS: Specify the mean and standard deviation for the following types of normreferenced tests assuming that the curve has a normal distribution:
36.  quartile score  mean =  standard deviation=  
37.  percentile score  mean =  standard deviation=  
38.  stanine score  mean =  standard deviation=  
39.  T score  mean =  standard deviation=  
40.  z score  mean =  standard deviation= 
INSTRUCTIONS: Look at the data from the test below, then answer Questions 4145 using any electronic device or software program that you know how to operate.
Raw score sections of four sections of a normreferenced language test of general English ability. (Correct number of items for each section of the test appears below) 

Section 1  Section 2  Section 3  Section 4  Total  
k (# of items)  10  30  20  30  90 
1. Diana  8  28  15  14  65 
2. Cindy  7  22  10  15  54 
3. Marilyn  4  11  9  8  32 
4. Jack  10  26  19  26  81 
5. Chris  5  15  10  16  46 
6. Faith  7  18  15  22  62 
7. Doug  9  10  12  21  52 
8. James  3  10  5  11  29 
9. Emiko  8  23  16  25  72 
10. Eric  6  19  12  18  55 
etc. . . 
41. What is the mean of the total test? 

42. What is the standard deviation? 

43. Which student(s) is/are more than one standard deviation from the mean? 

44. Do any sections of this test correlate closely in a way that's statistically significant at a p<.05 level (If so, mention which) 

45. What sort of distribution curve does this test have so far? 
[ p. 64 ]
(C) Exercise 3INSTRUCTIONS: The table below indicates the hypothetical data for a 50item test that were given to two different population samples. Look at that data then calculate the statistics mentioned in Questions 4650:
Population A  Population B  
sample size:  20  80 
mean score:  32  25 
standard deviation:  7.5  6 
lowhigh:  14  48  12  50 
alpha reliability estimate:  .7  .8 
46. ANOVA: 

47. Fratio: 

48. Chisquare distribution: 

49. effect size: 

50. standard error of measurement: 
INSTRUCTIONS: Compare the oral interview ratings below by two raters of the same student, then calculate the statistics mentioned in Questions 5155. Note that all ratings are in terms of 5point bands, with 5 representing the highest possible rating.
Category  Rater A  Rater B 
Grammar  3.5  3 
Fluency  4  4 
Pronunciation  4  3.5 
Cohesion  4  3.5 
Vocabulary  4.5  4 
Total  20  18 
51. The interrater reliability coefficient for A and B is . 

52. The Pearson correlation index for the two raters is . 

53. The index of concordance among the two raters is . 

54. The chisquare test of independence for these two raters is . 

55. The kappa coefficient of the combined rating is . 
INSTRUCTIONS: Read this hypothetical data comparing a 60item classroom pretest/posttest, then complete the sentences below. Note that following the pretest, the top onethird students were classified into an "upper group" and the lower onethird were classified into a "bottom group":
Category  Pretest  Posttest 
sample size:  48  42 
total mean:  30  33 
total range:  744  1252 
total standard deviation:  3.6  4.3 
upper group mean score:  45  50 
upper group standard deviation:  4.0  3.9 
bottom group mean:  20  20 
bottom group standard deviation:  4.2  5.8 
56. How did the upper group perform differently from the bottom group? . 

57. What sort of distribution curve would this posttest likely have? . 

58. Which type of ANOVA, if any, would be suitable for measuring the pretest/posttest gains made by this sample group? . 

59. What sort claims could validly be made about the "progress" of this class? . 
Main Article  Appendix A: I II III IV  Appendix B  Appendix C: I II III IV 