3 Introduction Students learning Spanish at the Defense Language Institute Foreign Language Center (DLIFLC) take five standardized tests, including a final test in their target language. DLIFLC follows the Interagency Language Roundtable (ILR) standards, which range from level 0 to level 5, 0 being no proficiency and 5 being functionally native proficiency (See Appendix A). ILR is ...an unfunded Federal interagency organization established for the coordination and sharing of information about language-related activities at the Federal level (www.govtilr.org/IRL%20History.htm). According to the DLIFLC webpage (2014), DLIFLC ...develops standardized in-course language tests and provides related reporting, test administration support, analysis, and consultation to promote continuous improvement in teaching and learning in DLIFLCs resident Basic Course language programs (Test Management Division section). The In-Course Proficiency Tests (ICPTs), are developed by the ...Test Development Division of the Directorate of Evaluation and Standardization, DLIFLC is the largest standardized foreign language test development organization in the United States (Test Management Division section). ICPT is a prochievement test which includes features of both proficiency and achievement. ICPT was introduced to bridge the gap between unit tests and the final test, the Defense Language Proficiency Test (DLPT). Unit tests are achievement tests designed to be taken after completing each unit, consisting of three to four chapters of the students textbook. The purpose of achievement testing is to determine an examinees acquisition of various aspects of course content, usually after a specific part of the curriculum has been covered. However, DLPT is a proficiency test that does not cover a specified curriculum. The purpose of language proficiency testing is to assess how well the students use the target language DATA ANALYSIS DLIFLC LC 1+ ICPT 101 & ICPT 102
4 in real-life situations. In contrast to achievement testing, proficiency testing focuses on overall language competence without regard to the place, length of time, or manner in which that competence has been acquired. For this reason, the unit test results do not indicate student language proficiency level sufficiently. According to DLIFLCs website (2014): The content domains of these ICPTs are representative samples of the domains of expected student cumulative language mastery in Basic Course instruction. Performance standards of the individual ICPTs are keyed to language mastery targets of specific subsets of the Basic Course that are identified by course titles and numbers within the Basic Course Transcript. Timing for administration of these tests is likewise keyed to the scheduled completion of each course subset. ICPTs test either global language listening and reading proficiency development or Final Learning Objective (FLO) job-related adjunct language performance skills. The ICPT is used to measure the students achievement in reading and listening comprehension prior to completing the DLPT, which is the final standardized test that students have to pass to successfully graduate from DLIFLC. The ICPT mimics the DLPT in all areas and aims to scaffold students from the lowest level of achievement, ILR level 0, to the highest possible level believed to be attainable during their stay in DLIFLC, ILR level 3. Prior to taking the DLPT, the ICPT is conducted five times and tests are identified as 101, 102, 201, 202, and 301. Those who achieve ILR levels 2, 2, 1+ in listening, reading, speaking respectively on the DLPT, can graduate and become a military linguist. It is therefore imperative to monitor students achievement through the ICPT series in order to ensure they are making adequate progress toward their learning goals. DATA ANALYSIS DLIFLC LC 1+ ICPT 101 & ICPT 102
5 Purpose The purpose of this study is to analyze a group of fourteen students listening comprehension performance at the ILR Level 1+ for ICPT 101 and 102. By comparing two succeeding ICPT results, teaching teams can determine whether students language proficiency is improving or not. They can then evaluate their instruction and curriculum according to the ICPT results. In addition, students can find out their language proficiency level and determine the areas they need to improve in. A comparison of the percent correct answers on the level 1+ questions between ICPT 101 and 102 listening tests will show whether or not those students improved their target language listening proficiency over the month. Those students who receive high scores on the ICPT can become more confident they will be successful on the DLPT. Nine out of ten cases among those who fail to pass DLPT is because they fail to achieve a level 2 in DLPT listening, although they pass level 2 and 1+ in DLPT reading and speaking. In other words, it is more difficult to achieve a level 2 in listening than achieving a level 2 in reading and 1+ in speaking. Students who are strong in listening but weak in reading and speaking will have a better chance to pass DLPT than those who are strong in reading and speaking but weak in listening. Considering level 2 in listening is the passing DLPT score, measuring the number of the correct answers of level 1+ questions is important because it will allow instructors to intervene early on if further support and study is needed. Compared to level 0 and level 1 questions, level 1+ consists of high proficiency related questions. Students who correctly answer more level 1+ questions indicate higher language proficiency. Level 1+ is not level 1.5. Level 1+ describes language characteristic of the next higher level but not sustained at that level. In other words, level 1+ is closer to level 1.9 which is very close to level 2. For this reason, it is important to compare the number of the correct DATA ANALYSIS DLIFLC LC 1+ ICPT 101 & ICPT 102
6 answers of level 1+ questions between ICPT 101 and 102 listening tests to determine of if student proficiency is adequately improving.
Methods ICPTs typically consist of 30 to 45 test items which are administered within a two hour time period. Due to confidentiality, the actual test items are classified and can not be disclosed. However, it can be revealed that the questions consist of multiple-choice, short answers, translation/transcription, and paragraph summaries. ICPT testing material topics cover broad areas such as economy, art, medical, politics, and so on. The questions are designed to measure student language proficiency in real life situations such as understanding conversations, answering questions, and finding directions. The tests are produced in paper-and-pencil and computer-delivered versions, with computer-delivered versions superseding the paper-and-pencil versions when test lab scheduling permits. This way, the answers are electronically recorded to ensure accuracy. The study group of fourteen students took the Spanish ICPT 101 for listening comprehension on February 28, 2013. Of the 40 items on this test, 17 assessed level 0+ proficiency, 18 assessed level 1 proficiency, and 5 questions assessed the targeted level 1+ proficiency. After a month of training and instruction, the same students took ICPT 102 for listening comprehension on March 28, 2013. This test was composed of 30 total test items, 5 assessing level 0+ proficiency, 16 assessing level 1 proficiency, 8 assessing the target level 1+ proficiency, and 1 assessing level 2 proficiency. Therefore, the target level 1+ for listening comprehension consisted of 5 test items on ICPT 101 and 8 test items on ICPT 102. DATA ANALYSIS DLIFLC LC 1+ ICPT 101 & ICPT 102
7 The tests are scored and computed by the DLIFLC testing department. The results are then sent back to the teaching team leader electronically within a week of the original test date. The score is calculated as both a numerical number and a percentage of the items that were correctly completed by the examinees. Due to the differing number of test items for level 1+ between ICPT 101 and 102, only the percent correct scores are used for this analysis to ensure equity in comparison. For this group of students, the mean for level 1+ corrected questions is 56% on ICPT 101 listening comprehension and 61% on ICPT 102 listening comprehension.
t Stat - 0.755112673 P(T<=t) one-tail 0.231821552 t Critical one-tail 1.770933396 P(T<=t) two-tail 0.463643104
t Critical two-tail 2.160368656
Table 2: ICPT Spanish Listening t-Test Results
The hypotheses are as follows; H 0 : 1 >= 2 and H 1 : 1 < 2 , where 1 and 2 denote respectively the population means for the 101 and 102 test score percentages for ILR level 1+ in listening comprehension. The test subjects are paired; therefore the t-Test for paired two sample means is used. The null hypothesis is directional, so the one-tailed t-Test value calculation is observed. The conventional alpha level is set at 0.05 and the degree of freedom (df) is 13.
Analysis In comparing the absolute value of the t-stat with the one-tail t-critical value, it is apparent that the t-Stat result of |-0.76| is much smaller than the one tailed t-critical value of 1.77 (0.76 < 1.77). Therefore, the null hypothesis cannot be rejected. Furthermore, the calculated one-tailed p-value of 0.23 is much larger than the conventional alpha level of 0.05 (0.23>0.05). This observation provides further evidence that the null hypothesis cannot be rejected. DATA ANALYSIS DLIFLC LC 1+ ICPT 101 & ICPT 102
9 After review of the statistical analysis, it can be concluded that 2 is not statistically greater than 1 ; therefore, the data does not support the research hypothesis of 1 < 2 . Based on this result, it is concluded that students did not achieve higher listening proficiency on the level 1+ questions between ICPT 102 than ICPT 101. In other words, after one month language instruction, students Spanish listening proficiency has not improved significantly.
Discussion Measuring the language proficiency of a group of individuals is a challenging process, even though ICPT is designed to produce two specific data sets, which are achievement and proficiency. Considering the number of questions (30 to 45 questions depending on which test) within an ICPT listening test, the number of questions may hold limitations that can possibly produce inaccurate data regarding learning results for real world situations. In other words, the reliability of ICPT may not be consistent and may not accurately reflect the actual student language proficiency. For this study, only level 1+ proficiency was analyzed due to the fact that this particular level is the precursor of level 2, which is the minimum needed to pass the DLPT. IPCT 101 only had five test items related to level 1+ and ICPT 102 only had eight. This is a very small number of questions from which to base an assessment on whether or not someone is proficient in Spanish listening comprehension. However, because of the rigor and standardization used in the development of these tests by DLIFLC, it may be assumed that the number of test items used to assess level 1+ are an accurate measure of proficiency. At the end of the program, students are required to enhance their skill level and acquire the target language proficiency through new instructional materials that have been proven to reflect these expectations. DATA ANALYSIS DLIFLC LC 1+ ICPT 101 & ICPT 102
10 Another issue with the assessment items used is that DLIFLC has strict confidentiality measures and the actual test items can not be disclosed. Ultimately, there is uncertainty as to which types of questions were measured according to the data reviewed. Certainly, it is easier to guess a correct answer on a multiple-choice question than to produce a correct answer on a short answer or paragraph summary when ones comprehension is not at the level of the subject matter. It is unknown what type of questions were on the test, which could make it difficult to measure the results of one test to the following test. For example, if ICPT 101 had five multiple- choice and ICPT 102 had five multiple choice, two short answer, and a paragraph summary, it may be more difficult to determine the level of correlation between ICPT 102 and ICPT 101. This concept could potentially affect the ability for students to show improvement from ICPT 101 to ICPT 102 in level 1+. However, DLIFLCs tests are standardized and have been created by the Test Review and Education Division of the Directorate of Evaluation and Standardization. It can therefore be assumed that the testing items have been evaluated thoroughly and accurately reflect the students listening comprehension proficiency.
Recommendations After careful review and consideration, it is recommended to use a holistic approach when evaluating the early stages of language proficiency related to classroom learning. The teaching team should use caution in their interpretation of ICPT 101 and 102 results because student language proficiency may have not been accurately measured by the limited test questions or the data analysis tool (measurement). In regards to the target audience and the early stages of their learning, language test results such as ICPT 101 and 102 should not be used to predict the final outcomes of DLPT. In other words, teachers need to remember that DATA ANALYSIS DLIFLC LC 1+ ICPT 101 & ICPT 102
11 students will increase their skill level and language proficiency with time and their ability to complete assigned tasks such as study guides etc. However, when it is noted that students have not improved in the early stages, it may be necessary to provide interventions that will enhance retention and transfer for later tests. ICPT results can be used to inform the content and materials used in the teaching teams daily lesson plans, especially in the early stages. For example, teaching teams may incorporate an increased amount of authentic materials such as news, drama, and radio in their classroom instructions. Also, the test results can be used to prepare and provide tailored instructions for those who received lower scores on ICPTs. Low proficiency students will be focused by one-on- one instruction in order to increase their language proficiency. Results can also be used to reshuffle students between classes in order to create a balance in the student proficiency level ratio. Teaching teams should avoid placing low proficiency students together in same classroom because peer role modeling is important in language learning. In addition to these strategies, Voluntary Special Assistance is a form of extra help that is provided at the last hour hour of classes every day between 3:00pm and 3:50pm. Those students can benefit from that extra help but the teaching team would have to recommend it by an official counseling. Although, the mean scores showed a small increase from ICPT 101 to ICPT 102, the concluding evidence shows that there is no statistical increase between these tests. It is therefore recommended that the data from this analysis is used to inform the teaching team that this group of students will require further support and intervention strategies before they can achieve level 2 proficiency in listening comprehension. Early intervention will help to ensure success for these students so that they may graduate from DLIFLC as military Spanish linguists.
DATA ANALYSIS DLIFLC LC 1+ ICPT 101 & ICPT 102
12 REFERENCES
Defense Language Institute Foreign Language Center (2014). In-Course Proficiency Test Development Division. Retrieved on June 21, 2014 from: http://www.dliflc.edu/in-courseprofici.html
Interagency Language Roundtable (2014). Interagency Language Roundtable Language Skill Level Descriptions Listening. Retrieved on June 21, 2014 from: http://www.govtilr.org/Skills/ILRscale3.htm
DATA ANALYSIS DLIFLC LC 1+ ICPT 101 & ICPT 102
13 Appendix A
The following are the definitions of the Listening Comprehension levels according to the ILR webpage (http://www.govtilr.org/skills/ILRscale3.htm, 2014). Listening 0 (No Proficiency) No practical understanding of the spoken language. Understanding is limited to occasional isolated words with essentially no ability to comprehend communication. Listening 0+ (Memorized Proficiency) Sufficient comprehension to understand a number of memorized utterances in areas of immediate needs. Slight increase in utterance length understood but requires frequent long pauses between understood phrases and repeated requests on the listener's part for repetition. Understands with reasonable accuracy only when this involves short memorized utterances or formulae. Utterances understood are relatively short in length. Misunderstandings arise due to ignoring or inaccurately hearing sounds or word endings (both inflectional and non-inflectional), distorting the original meaning. Can understand only with difficulty even such people as teachers who are used to speaking with non-native speakers. Can understand best those statements where context strongly supports the utterance's meaning. Gets some main ideas. Listening 1 (Elementary Proficiency) Sufficient comprehension to understand utterances about basic survival needs and minimum courtesy and travel requirements in areas of immediate need or on very familiar topics, can understand simple questions and answers, simple statements and very simple face-to-face conversations in a standard dialect. These must often be delivered more clearly than normal at a rate slower than normal with frequent repetitions or paraphrase (that is, by a native used to dealing with foreigners). Once learned, these sentences can be varied for similar level vocabulary and grammar and still be understood. In the majority of utterances, misunderstandings arise due to overlooked or misunderstood syntax and other grammatical clues. DATA ANALYSIS DLIFLC LC 1+ ICPT 101 & ICPT 102
14 Comprehension vocabulary inadequate to understand anything but the most elementary needs. Strong interference from the candidate's native language occurs. Little precision in the information understood owing to the tentative state of passive grammar and lack of vocabulary. Comprehension areas include basic needs such as: meals, lodging, transportation, time and simple directions (including both route instructions and orders from customs officials, policemen, etc.). Understands main ideas. Listening 1+ (Elementary Proficiency, Plus) Sufficient comprehension to understand short conversations about all survival needs and limited social demands. Developing flexibility evident in understanding a range of circumstances beyond immediate survival needs. Shows spontaneity in understanding by speed, although consistency of understanding is uneven. Limited vocabulary range necessitates repetition for understanding. Understands more common time forms and most question forms, some word order patterns, but miscommunication still occurs with more complex patterns. Cannot sustain understanding of coherent structures in longer utterances or in unfamiliar situations. Understanding of descriptions and the giving of precise information is limited. Aware of basic cohesive features (e.g., pronouns, verb inflections) but many are unreliably understood, especially if less immediate in reference. Understanding is largely limited to a series of short, discrete utterances. Still has to ask for utterances to be repeated. Some ability to understand facts. Listening 2 (Limited Working Proficiency) Sufficient comprehension to understand conversations on routine social demands and limited job requirements. Able to understand face- to-face speech in a standard dialect, delivered at a normal rate with some repetition and rewording, by a native speaker not used to dealing with foreigners, about everyday topics, common personal and family news, well-known current events and routine office matters DATA ANALYSIS DLIFLC LC 1+ ICPT 101 & ICPT 102
15 through descriptions and narration about current, past and future events; can follow essential points of discussion or speech at an elementary level on topics in his/her special professional field. Only understands occasional words and phrases of statements made in unfavorable conditions, for example through loudspeakers outdoors. Understands factual content. Native language causes less interference in listening comprehension. Able to understand facts; i.e., the lines but not between or beyond the lines. Listening 2+ (Limited Working Proficiency, Plus) Sufficient comprehension to understand most routine social demands and most conversations on work requirements as well as some discussions on concrete topics related to particular interests and special fields of competence. Often shows remarkable ability and ease of understanding, but under tension or pressure may break down. Candidate may display weakness or deficiency due to inadequate vocabulary base or less than secure knowledge of grammar and syntax. Normally understands general vocabulary with some hesitant understanding of everyday vocabulary still evident. Can sometimes detect emotional overtones. Some ability to understand implications. Listening 3 (General Professional Proficiency) Able to understand the essentials of all speech in a standard dialect including technical discussions within a special field. Has effective understanding of face-to-face speech, delivered with normal clarity and speed in a standard dialect on general topics and areas of special interest; understands hypothesizing and supported opinions. Has broad enough vocabulary that rarely has to ask for paraphrasing or explanation. Can follow accurately the essentials of conversations between educated native speakers, reasonably clear telephone calls, radio broadcasts, news stories similar to wire service reports, oral reports, some oral technical reports and public addresses on non-technical subjects; can understand without difficulty all forms of standard speech concerning a special professional DATA ANALYSIS DLIFLC LC 1+ ICPT 101 & ICPT 102
16 field. Does not understand native speakers if they speak very quickly or use some slang or dialect. Can often detect emotional overtones. Can understand implications. (Has been coded L-3 in some nonautomated applications.) [Data Code 30] Listening 3+ (General Professional Proficiency, Plus) Comprehends most of the content and intent of a variety of forms and styles of speech pertinent to professional needs, as well as general topics and social conversation. Ability to comprehend many sociolinguistic and cultural references. However, may miss some subtleties and nuances. Increased ability to comprehend unusually complex structures in lengthy utterances and to comprehend many distinctions in language tailored for different audiences. Increased ability to understand native speakers talking quickly, using nonstandard dialect or slang; however, comprehension is not complete. Can discern some relationships among sophisticated listening materials in the context of broad experience. Can follow some unpredictable turns of thought readily, for example, in informal and formal speeches covering editorial, conjectural and literary material in subject matter areas directed to the general listener. Listening 4 (Advanced Professional Proficiency) Able to understand all forms and styles of speech pertinent to professional needs. Able to understand fully all speech with extensive and precise vocabulary, subtleties and nuances in all standard dialects on any subject relevant to professional needs within the range of his/her experience, including social conversations; all intelligible broadcasts and telephone calls; and many kinds of technical discussions and discourse. Understands language specifically tailored (including persuasion, representation, counseling and negotiating) to different audiences. Able to understand the essentials of speech in some non-standard dialects. Has difficulty in understanding extreme dialect and slang, also in understanding speech in unfavorable conditions, for example through bad loudspeakers outdoors. DATA ANALYSIS DLIFLC LC 1+ ICPT 101 & ICPT 102
17 Can discern relationships among sophisticated listening materials in the context of broad experience. Can follow unpredictable turns of thought readily, for example, in informal and formal speeches covering editorial, conjectural and literary material in any subject matter directed to the general listener. Listening 4+ (Advanced Professional Proficiency, Plus) Increased ability to understand extremely difficult and abstract speech as well as ability to understand all forms and styles of speech pertinent to professional needs, including social conversations. Increased ability to comprehend native speakers using extreme nonstandard dialects and slang, as well as to understand speech in unfavorable conditions. Strong sensitivity to sociolinguistic and cultural references. Accuracy is close to that of the well-educated native listener but still not equivalent. (Has been coded L-4+ in some nonautomated applications.) [Data Code 46] Listening 5 (Functionally Native Proficiency) Comprehension equivalent to that of the well- educated native listener. Able to understand fully all forms and styles of speech intelligible to the well-educated native listener, including a number of regional and illiterate dialects, highly colloquial speech and conversations and discourse distorted by marked interference from other noise. Able to understand how natives think as they create discourse. Able to understand extremely difficult and abstract speech.