测试学论文spass

Evaluation of a Final Achievement T estI. IntroductionHowever well designed a test may be, and however carefully it has been edited, it is impossible to know how it will work until it has been tired out on students. Therefore, it is indispensable to analyze the test data by both quantitative and quantitative means after a particular test is administered in order to determine whether an appropriate balance among the qualities has been achieved on which the overall usefulness of the test depends.II. Evaluation on the Total LevelBefore examining the overall picture of the reading test, we would like to display the data obtained from this particular test in table.总分N Valid 28Missing 0Mean 65.7500Median 64.5000Mode 73.00Std. Deviation 15.20873Variance 231.30556Skewness .029Std. Error of Skewness .441Kurtosis -.634Std. Error of Kurtosis .858Range 55.00Minimum 36.00Maximum 91.00The data presented in table visually present the information about how the students have performed on the test. Generally speaking, on the condition that the pass mark was set at 25% of the students could pass the test, whereas 75% of them failed it. This proves that the test is quite difficult for them. To examine the details of the performance of the students, we must survey their group behaviors.Frequency Percent ValidPercent Cumulative PercentValid 2.00 3 10.7 10.7 10.73.00 2 7.1 7.1 17.94.00 2 7.1 7.1 25.05.00 4 14.3 14.3 39.36.00 9 32.1 32.1 71.47.00 1 3.6 3.6 75.08.00 1 3.6 3.6 78.69.00 4 14.3 14.3 92.910.00 2 7.1 7.1 100.0Total 28 100.0 100.0According to Brown, central tendency refers to the typical behavior of a given group, while dispersion refers to how the performance of those individualsmay vary from that typical behavior. To investigate central tendency, we mustobserve three different indicators: mean, mode, and median. To investigatedispersion, we must observe two indicators: range and standard deviation (SD).By means of the statistical methods provides by Brown (ibid), we have computedthe central tendency and dispersion of the reading test. The data are displayed inTable.III. Evaluation an the Component Level阅读完形英译汉汉译英词汇N Valid 28 28 28 28 28Missing 0 0 0 0 0 Mean 21.3571 5.8929 13.4286 14.0000 10.6786 Median 20.0000 6.0000 14.0000 15.5000 10.5000 Mode 20.00 6.00 8.00(a) 16.00 13.00a Multiple modes exist. The smallest value is shownCentral Tendency and Dispersion of Total ScoresCentral Tendency DispersionMean Median Mode Low High Range SD54.8519 54.0000 71.50 21.50 76.30 54.00 15.7825**Total Score = 100 N = 28we hold the belief that in the light of the scientific testing and reading theories. On the basis of the chart, Our investigation of the above items indicates that the Grammar’s Skewness is positive number. so the Grammar submit positively skewed. Reading skill, Close, Chinese-English, English-Chinese submit negatively skewed. According to the Variance, On the one hand, as for Chinese-English and English-Chinese, the students have a big differ each others. As for Grammar, Reading skill, close, they have small differ mutual.IV. Evaluation on the Item LevelDistribution of Responses to Improper ItemsItem Group Option D P36 H A* B C D Omit8 0 0 0 0 0.67 0.43Std. Deviation 3.44496 2.36235 4.99947 4.33760 3.39993Variance11.86772 5.58069 24.9947118.8148111.55952Skewness .025 .083 -.530 -1.044 .261 Std. Error of Skewness .441 .441 .441 .441 .441 Kurtosis -.583 -.660 -.556 .340 -.316 Std. Error of Kurtosis .858 .858 .858 .858 .858 Range 14.00 8.00 18.00 16.00 14.00 Minimum 14.00 2.00 2.00 3.00 4.00 Maximum 28.00 10.00 20.00 19.00 18.00L 4 2 1 1 037 HL A B* C D Omit0 8 0 0 01 42 1 00.67 0.4338 HL A B C* D Omit0 0 8 0 01 4 3 0 00.73 0.3939 HL A* B C D Omit8 0 0 0 02 3 1 2 00.8 0.3640 HL A B* C D Omit0 8 0 0 01 2 3 2 00.8 0.3641 HL A B C D* Omit0 1 1 6 00 6 0 2 00.75 0.2942 HL A* B C D Omit8 0 0 0 03 3 1 1 00.73 0.3943 HL A B* C D Omit5 3 0 0 07 0 0 1 01 0.1144 HL A* B C D Omit8 0 0 0 04 1 2 1 00.67 0.4345 HL A B* C D Omit0 8 0 0 01 2 3 2 00.8 0.36H = high score group L = low score group D = discriminability P =difficulty*= correct response N = 28Our investigation of the above items indicates that item 43, have reached a high level of discriminability. However, their difficulty level is too high, they are too difficult for the students. Therefore, some adjustment ought to be implemented to keep the items in the proper difficulty range.Our investigation also indicates that items 36 37 38 41 have achieved a low level of difficult they are too easy for the students. Meanwhile their discriminability levels are quite low, beyond the range of the lowest acceptable discriminability. Distractor analysis shows the following results.Reliability****** Method 1 (space saver) will be used for this analysis ******R E L I A B I L I T Y A N A L Y S I S - S C A L E (S P L I T)Reliability CoefficientsN of Cases = 28.0 N of Items = 6Correlation between forms = .8502 Equal-length Spearman-Brown = .9190Guttman Split-half = .7202 Unequal-length Spearman-Brown = .91903 Items in part 1 3 Items in part 2Alpha for part 1 = .6738 Alpha for part 2 = .5638V. ConclusionIn communicative language testing, the optimal result is, naturally, which contribute to test usefulness, come up to a high level simulataneously. Although it is difficult to achieve them all in practice, compromise can be reached to obtain an appropriate balance among them according to the test purpose.Based on the low level of the internal consistency of the test scores, we can conclude that the test is not quite reliable. Meanwhile, the investigation of the item difficulty and discriminability confirms our conclusion. Reliability is the consistency of test scores.On the basis of the analyses of content relevance and content coverage of the test. Validity refers to a set of tests or test content reached its expected test the degree of intent.References:1.Brown, J. D. 1988. Understanding Research in Second Language Learning.Cambridge: Cambridge University Press.2.王晓军. 西部英语测试实证研究&SPSS操作. 宁夏：宁夏人名出版社，2006.3. 苏金明.统计软件SPPS系列应用实战篇.北京：电子工业出版社，20024. 杨惠中，Weir, C.J..大学英语四,六级考试效度研究.上海：上海外语教育出版社，1998。