Avant STAMP 4S User Guides
Test takers receive a score (Benchmark Level) for each section of the test. The Reading and Listening items are scored automatically by the computer. The Writing and Speaking items are scored by Avant’s Valid-certified raters who use a Scoring Rubric (a scoring tool that lists the criteria for meeting Benchmark Levels). The rubric outlines in detail what expectations are placed upon the test takers for performance at any given level. Test takers who hope to score at a specific level know the characteristics of the work they must produce to achieve that level.
Scoring is done using Benchmark Levels 1-9 for Reading and Listening and Benchmark Levels 1-8 for Writing and Speaking. The levels are associated with Benchmark Categories of Novice, Intermediate and Advanced as shown in the Level Keys below. Similar to ACTFL's low, mid and high designations, these designations will assist in seeing a further breakdown of the test taker’s ability. Because it takes a great deal of time and practice for students to acquire the skills necessary to move from Novice to Intermediate, student growth can be tracked using the numeric levels within the categories.
This Benchmark Scale aligns to the ACTFL scale as shown below.
Each Reading and Listening test item has an associated level. Because the Reading and Listening sections are computer-adaptive, the test taker experiences questions at various levels. The algorithm used to determine the test taker’s scores for Reading and Listening looks at the number of questions that were answered correctly at each test item level.
The chart below shows topics typically associated with the Benchmark Levels for Reading and Listening test items. It also shows the Reading and Listening characteristics associated with the levels.
Keep in mind that topic is not the only criteria for determining the level for STAMP 4S test items. A Novice-level test taker is likely to know learned phrases and basic vocabulary about topics that may be listed under Intermediate or Advanced.
Also, test items may have elements of multiple topics. For example, a Novice-level question about Calendar/Time might reference Holidays/Celebrations. The topics may be used as a guide for instruction, but are not a complete scope and sequence.
Scoring for the Avant STAMP for Hebrew Writing and Speaking responses is done using the following three-step process and associated rubric.
Overall evaluation of the response to assess for Ratability — is the response on topic and appropriate for the given prompt/ task? If the response is off topic, contains any profanity or includes any menacing or threatening language, the response will be treated as unresponsive or non-ratable and the test taker will receive a “0” score for that task.
Evaluation of each ratable response to determine a Benchmark Level based upon the Text-Type characteristics, followed by rating of Language Control factors (see next page).
The system will then post ratings for each response, which are used to calculate the final score for the domain.
Note: 20% of all test taker scores are rated by a second rater to assist in tracking rater accuracy and reliability.
The same rubric is used for all Speaking and Writing items. Once rated, the average ratings across all Speaking and Writing items will appear on the summary report page and scores for each sample will appear on the detailed score page.
|Benchmark Level||Text Type Characteristics||Language Control|
|Produces words in target language with no connection. Does not have enough vocabulary or the ability necessary to formulate even simple phrases to address the prompt.||Limited language control. Inability to create more than individual words. L1 (first language) influence may be strong. Errors are expected at this level, but the student must be able to produce at least two comprehensible words.|
|Language production is beyond individual words but clearly shows the lack of ability to construct more than phrases. May include one simple sentence, but incapable of showing more.||May make frequent errors, but usually comprehensible to a sympathetic reader/listener. L1 (first language) influence may be present.|
|Short, common expressions or memorized statements that may be combine together. Able to create at least 2 different simple sentences.||Good accuracy for high frequency expressions. Usually comprehensible to a sympathetic reader/listener. Grammatical (syntax, spelling, conjugation) errors are expected at this level but sentences must make sense to be acceptable.|
|Variety of sentences that utilize different verbs to create independent thoughts, mostly composed of a recombination of learned simple sentences with added detail in the form or prepositional phrases and verbal phrases.||Good accuracy with formulaic sentences with some added detail. Errors may occur as student attempts higher level skills. Good Language Control is expected with the majority of the response.|
|Logical organization of ideas and flow of sentences or statements. Contains at least 2 unique and non-formulaic sentences intermixed with a variety of sentences creating ‘groupings of sentences showing connectedness.||Good accuracy evident with possible errors that don’t affect the overall meaning. Delivery may be somewhat choppy. May have repetitive use of concrete vocabulary with occasional use of expanding terms. Accuracy for complex sentences is emerging.|
|Able to demonstrate an Intermediate-High control of the language. Complexity is evident. Transition words and connectors are used correctly and efficiently a majority of the time to create a flow within the response. Groups of sentences focus on different aspects of the prompt and include transition words or phrases to introduce next concept. Does not have Advanced vocabulary or language to move into the Advanced level.||Response is well written and constructed. Intermediate-High sentence structures can be found throughout response. Demonstrates beginning ability to create a coherent response with increased use of complexity as well as transition words and phrases. No evidence of advanced vocabulary.|
|A paragraph response with advanced language with complexity, syntactical and grammatical control transitional words and phrases found within the text creating a natural flow. Demonstrates a connection of thoughts that create a coherent and extended discourse.||Language is error-free a majority of the time with familiar topics. If errors exist, they are patterned and do not hinder overall meaning. Delivery is mostly fluent with only occasional hesitancy. Some abstract and precise use of vocabulary and terms with familiar topics is evident.|
|Variety of cohesive devices and organizational patterns are evident throughout response. Vocabulary is clear, specific and natural. Language is smooth and native-like in delivery and without noticeable errors.||Language is presented with limited errors, if any. Ability to create complex language using precise and extensive vocabulary. Control of the abstract as well as ease of use in idiomatic phrases and concepts. Clear sequential ordering evident (if required) and accurately follows target language conventions.|
Updated July 2016