JP2009229606A - Method and system for analyzing and processing examination answer - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To extract educational subjects by measuring, from variety of perspectives, a series of comprehension degree of the points included in a specific learning item and obtaining the tendency of comprehension of the population through collecting the results of measurement. <P>SOLUTION: A check box is created by extracting, from a typical example solution which is one example answer to a question statement, the parts corresponding to a plurality of sub-items obtained by subdividing the learning items and by listing check elements. Example answer patterns composed of combinations of the presence and absence of the check elements are obtained for each check element of the check box, and the check elements are corrected so that none of the example answer patterns are the same with one another. Then, individual patterns with respect to the description of answering procedure of each examinee composed of combinations of the presence and absence of the check elements are obtained for each check element of the check box, and the individual patterns are recorded together with identification codes for identifying the examinees. The number of examinees of the same individual pattern is summed up. <P>COPYRIGHT: (C)2010,JPO&INPIT

Description

  The present invention provides an answer analysis processing method and an answer analysis for giving a question sentence including a plurality of information elements to a plurality of examinees and analyzing an understanding tendency of a predetermined learning item of the entire examinees from description of a solution procedure for the information. It relates to a processing system.

  In entrance examinations at various schools, students are given a question sentence containing multiple information elements and asked for a description of the solution procedure or opinion, and the ability to process the question for each candidate is measured (scored). Judgment of admission is made. In such a test, since the measurement result affects the profit / disadvantage of the examinee, the fairness of the measurement, that is, the uniformity that excludes the variation for each measurement or measurer (scorer) is required. Therefore, for example, there is a case where a “multiple-choice type test” that allows the examinee to select a final answer from several options may be employed. The candidate's answers are aggregated into choices, so the measurer can handle this mechanically and eliminate variations from measurement to measurement.

  By the way, when the educational guidelines of the national and local governments are established, it is necessary to measure the understanding tendency of the population with respect to a specific learning item from a multifaceted test and extract educational problems. Also for such a test, the uniformity (excluding the variation for each measurement (scoring) or measurer (scorer) is required, so the “multiple-choice test” can be adopted. In other words, specific learning items are subdivided into points in accordance with the predicted understanding tendency, and problems and options corresponding to these points are created. If the answers of examinees are traced in reverse, the tendency of the population to understand the specific learning items can be measured. In such a method, in order to accurately extract educational problems, the points of learning items must reflect the understanding tendency of the population in detail. However, the understanding tendency of the population is not always predictable in detail. In addition, even if some candidates select an option without going through a predetermined procedure, this is indistinguishable from a candidate who has selected through a predetermined procedure, and the measurement tends to be inaccurate.

  Therefore, a problem sentence including multiple information elements is given to all or a part of the population, and a fixed amount of free description such as solution procedures or thoughts and opinions is sought, and it is included in a specific learning item for each examinee. A “descriptive test” may be employed in which a series of comprehension of a given point is measured in a multifaceted manner and collected to measure an understanding tendency of a population to a specific learning item.

  By the way, in the “descriptive test”, an attempt has been made to improve the uniformity of measurement by eliminating the variation for each measurement or measurement person.

  Patent Document 1 discloses an automatic scoring system for mechanically processing an examinee's free description in a “descriptive test”. An answer model is created by linguistic analysis of the free description of the examinee's problem, and the score is compared with the model answer stored in advance. Since the answer model and the model answer are provided with a certain range by language analysis, it is possible to easily determine whether or not they match each other. However, here, only the comparison between the model answer prepared in advance and the free description of the examinee is performed, and a series of comprehension for a specific learning item is not measured.

Patent Document 2 discloses a system for scoring (measuring) the free description of the examinee in a “descriptive test” in which the answer is not uniquely determined. If the scoring results of one grader can be corrected on the system, and the scoring results do not match among multiple scoring personnel, and there is doubt about the objectivity of the scoring results, the measurement basis of other measurers This can be corrected by re-evaluating the measurement results with reference. By sharing measurement results and measurement grounds among multiple measurers, it is possible to re-evaluate answers from various viewpoints without being constrained by subjective measurement results. In this case as well, only a comparison between the prepared model answer and the free description of the examinee is performed, and a series of comprehension levels for a specific learning item is not measured.
JP 2001-56634 A JP 2006-277086 A

  By the way, when you need to conduct a “descriptive test” on a population of tens of thousands or hundreds of thousands of people, measure a series of understanding trends for specific learning items in the population, and extract educational issues There is. On the other hand, as described above, in the conventional descriptive test, the person who processes the description of the examinee's solution procedure, that is, the judgment / discretion of the grader is intervened in the process. There was a problem that it differs depending on each grader. This tendency becomes a bigger problem as the number of examinees increases.

  The present invention measures even a large number of such candidates and measures the multiplicity of understanding of the points included in a specific learning item, and collects them to obtain an understanding tendency of the population. There is provided an answer analysis processing method and system capable of extracting an educational problem.

  The method of the present invention provides an answer analysis processing method in which a question sentence including a plurality of information elements is given to a plurality of examinees, and a description of a solution procedure for the information is requested, and an understanding tendency of a predetermined learning item of the entire examinees is analyzed. A portion corresponding to a plurality of small items obtained by subdividing the learning item from a typical answer example which is one answer example of the question sentence, and a set of numerical values, formulas, A plurality of other answer examples consisting of a check box creation step for creating check boxes by listing check elements consisting of words or sentences and a description of a solution procedure that gives a difference in the degree of understanding of the predetermined learning item Create and obtain an answer example pattern that is a combination of the presence or absence of each check element of the check box for each of the typical answer example and the answer example. A correction step for correcting the check elements so as not to be the same, and a description of the solution procedure for each examinee for a personal pattern that is a combination of the presence or absence of each check element of the check box, and the examination A personal recording step of recording the personal pattern together with an identification code for identifying a person, and a counting step of counting the number of examinees who have the same personal pattern.

  According to the present invention, the grader can check whether or not there is a check element (term) consisting of a set of numerical values, an expression, a word, or a sentence in the description (answer) of the candidate's solution. Since it is only checked, the judgment and discretion of the grader is not involved in the processing. Therefore, even for examinees having a scale of tens of thousands or hundreds of thousands, the answer can be quickly processed by a plurality of graders or machines, and the processing result is always constant. In addition, since such check elements correspond to multiple sub-items that subdivide the learning items, the combination of presence / absence for each check element reflects a series of comprehension of points included in a specific learning item. Is. Therefore, if the processing data is aggregated for the combination of presence / absence of each check element, an understanding tendency about a specific learning item of the population can be obtained, so that an educational problem can be extracted.

  Further, the system of the present invention provides a question analysis that gives a plurality of examinees a question sentence including a plurality of information elements and requests a description of a procedure for solving the problem, thereby analyzing an understanding tendency of a predetermined learning item of the whole examinee. A processing system, which extracts a portion corresponding to a plurality of small items obtained by subdividing the learning item from a typical answer example which is one answer example of the question sentence, and a set of several numerical values constituting the same, Check box display means for allowing and displaying a check element consisting of an expression, a word, or a sentence, and an input for specifying the test taker for the description of the solution procedure for each test taker In response to this, an identification code is generated, and the presence / absence of each check element in the check box displayed by the check box display means is allowed to be input. A check means for generating a personal pattern and recording the personal pattern together with the identification code as a personal record, and a counting means for counting the number of examinees who have the same personal pattern. And

  According to the present invention, as described above, even for a large-scale examinee, it is possible to obtain an understanding tendency about a specific learning item of the population, and thus it is possible to extract an educational problem. .

  In order to explain the principle of the present invention, a method for analyzing the understanding tendency of a predetermined learning item for a population of n examinees will be described with reference to FIGS.

  As shown in FIG. 1, one learning item is an aggregate made up of many small items (A, B, C, D, E, F...). In order to analyze the understanding tendency of the population for this predetermined learning item, several sub-items (A, B, C, D, E) are selected (S1), and information elements corresponding to these are selected. A problem is created so as to include (a, b, c, d, e) (S2). Note that some information elements (e) may be inevitably derived from the outside in solving the problem even if not specified in the problem sentence. (For example, the value of the pi may be used as needed even if it is not specified in the problem text.) Such an information element (e) may not be specified in the problem text.

  Next, a typical answer example which is one answer example including a solution process for solving a problem is created (S3). Typical answer examples are expressed by “terms” (p1 to p13) such as a set of numerical values, expressions, sentences, and words.

  On the other hand, analysis items (A, B, C) for analyzing the understanding tendency are determined from the selected small items (A, B, C, D, E) (S4). In addition, the more the analysis items, the more detailed understanding tendency of the population with respect to the predetermined learning items can be known. On the other hand, when there are too many, the difference between analysis items will become indistinct, and the complexity of a process will increase. Therefore, it is preferable to adjust the number of small items as analysis items as appropriate.

  “Terms” corresponding to each of the analysis items (A, B, C) are extracted from “terms” (p1 to p13) in the typical answer example (S5). For example, for analysis item A (p1, p2, p3, p4), for analysis item B (p2, p5, p6, p7, p8) and for analysis item C (p3, p9, p10) It is assumed that it is expressed in the typical answer example by the body. Each aggregate may include a “term” common to the analysis items (A, B, C). In addition to the typical answer examples, other “terms” may be given corresponding to the analysis items.

As shown in FIG. 2, at least one or more “terms” are selected and listed from a collection of “terms” for each analysis item (A, B, C). Check box that lists the check box elements to check whether the selected "term" (p1, p2, p5, p7, p9, p10) is included or not included in the description of the solution procedure of the answer create. Here, the check pattern, which is the result of checking with two alternatives for all the check box elements, has 2 ⁿ , that is, 2 ⁿ combinations here, where ⁿ is the number of check box elements. The series of "term" is also possible to make a linked this (p1 + p2, p5, p7 , p9 + p10), In this case, the check pattern is a combination of ^two ways ^4.

  An answer example α, β, γ... Reflecting a specific understanding tendency and degree of understanding is created (S7). The answer examples α, β, γ... May include typical answer examples. The answer examples α, β, γ... Are also composed of a collection of “terms”. For example, for the answer example α (p1 to 13), for the answer example β (p1 to p6, p7 ′, p8 ′, p9 ′), for the answer example γ (p1, p12, p13) Let it be expressed. For each of the answer examples α, β, γ..., The presence / absence of the “term” (p1, p2, p5, p7, p9, p10) selected as the check box element is checked to obtain a check pattern. Here, “Yes” for each term is indicated by ○ and “No” is indicated by ×, for example answer α (○, ○, ○, ○, ○, ○), and example answer β (○, ○, ○, × , X, x) and answer example γ are (◯, x, x, x, x, x). When the same check pattern is obtained for the answer examples α, β, γ..., The check box element is changed.

  Now, the n examinees who make up the population are given a question sentence including a plurality of information elements as described above, and are requested to submit a description (answer) of the solution procedure. Check the existence of each of the “terms” (p1, p2, p5, p7, p9, p10) selected as the check box elements for the description of the answering procedure of all n or a part of them. Ask for. Since the check for each element of the check box only checks whether or not the “term” is present in the description of the examination procedure of the examinee, automatic counting by a machine is also possible. Even if it is manual, there is almost no difference in the check pattern results depending on the person in charge of the check. Therefore, a large number of check persons may be employed, and rapid processing can be performed.

  Furthermore, the number of persons for each check box combination pattern is totaled. From this aggregation, it is possible to know the tendency of the population to understand the predetermined learning items. That is, it is predicted that the examinee's answer having the same check box combination pattern as the above answer examples α, β, γ... Has the same specific level of understanding as each answer example. An understanding tendency of a population with respect to a predetermined learning item can be obtained from a ratio of examinees having the same specific degree of understanding as the answer examples α, β, γ. As the number of persons other than the combination of check boxes that are the same as the answer examples α, β, γ,... Therefore, since it is easy to aggregate check patterns even if the number of people making up the population is large, a check box with a changed term is created to increase the “probability” and the points included in a specific learning item. A series of comprehension levels can be measured in many ways.

  From the total number of people for each check pattern, for example, the check pattern with the largest number of people, or what kind of predetermined learning item the specific check pattern has for understanding, is an example of an answer as described above. This can be achieved by creating various check patterns and correcting the example answers until the patterns are the same.

  In the conventional descriptive test, a person who processes the description of the examinee's solution procedure, that is, the judgment and discretion of the grader, intervenes in the description of the examinee's solution procedure, that is, the processing of the answer. Therefore, there has been a problem that the processing result varies depending on the scoring or each grader. According to the above aspect, the grader only checks whether or not there is a “term” in the description of the examinee's solution procedure, so the judgment / discretion of the grader is involved in the processing of the examinee's answer. Without processing, the processing results are hardly different for each grader or grader. In addition, since these “terms” correspond to multiple sub-items that subdivide the learning items, the combination of the presence / absence of each check element reflects a series of comprehension of the points included in the specific learning item. It is a thing. Therefore, if the processing data is aggregated for the combination of presence / absence of each check element, an understanding tendency about a specific learning item of the population can be obtained, so that an educational problem can be extracted.

  Further, one embodiment of the present invention will be described in detail with reference to FIGS.

  In this embodiment, L1: “form a one-dimensional quadratic equation based on the condition of the question sentence, obtain a plurality of solutions of this equation, and then answer according to the preconditions indicated in the question sentence”. One embodiment of a method and a system for analyzing and analyzing the understanding tendency of a plurality of examinees for one mathematics learning item will be described.

  First, as shown in FIG. 3, the content of the learning item L1 is subdivided, and a small item L2 to be questioned in this problem is extracted. That is, <1> One-dimensional quadratic equation can be formulated from the meaning, <2> One-dimensional quadratic equation can be accurately transformed, <3> One-dimensional quadratic equation can be obtained, <4> 2 obtained Four sub-items <1> to <4> are extracted from answering according to the preconditions indicated in the question sentence.

Next, as shown in FIG. 4, a question sentence Q1 corresponding to the small items <1> to <4> described above and an answer example A1 are created as typical answer examples. In addition, it is preferable that the typical answer example is an answer example generally called “exemplary answer”, but the answer is not limited thereto. Here, a calendar, a familiar subject, was selected as a problem for formulating a one-dimensional quadratic equation from an array of numbers. The question sentence Q1 must be created so that the degree of understanding of the examinee can be confirmed in the answer for each of the extracted small items L2. That is, it must be reflected so that at least each of the small items L2 can be confirmed in the answer example A1, and the question sentence is corrected as necessary. Each sub-item <1> to <4> is reflected in the answer example A1 as follows.
<1>: (First line)
<2>: (second line) to (fourth line)
<3>: (5th line)
<4>: Not reflected
That is, in the answer example A1, the two solutions obtained from the one-dimensional quadratic equation in the fifth line are directly transferred to the “answer” column in the sixth line, and the subitem <4> “obtained” is obtained. It does not reflect the content of “answering from the two solutions according to the preconditions indicated in the question sentence”. Therefore, although the same one-dimensional quadratic equation is derived, the correction of the question sentence Q1 is devised so that one of the two solutions is inappropriate from the preconditions indicated in the question sentence.

  As shown in the question sentence Q2 of FIG. 5, for example, a condition for changing one of the two solutions of the one-dimensional quadratic equation by adding a restriction to the value of x by changing the arrangement of numbers in the calendar is added. Then, as shown in the answer example A2 which is a typical answer example for the question sentence Q2, one of the two solutions “3” and “7” “7” is “left” when referring to the calendar of the question sentence Q2. It is excluded from the answer because there is no “neighboring number”. In addition, in the question sentence Q2, the instruction “Please write the required process” is added to urge the students to describe the minimum required procedure for the analysis in this embodiment.

As shown in FIG. 6, once again confirming whether each sub-item <1> to <4> is reflected in the answer example A2,
<1>: (First line)
<2>: (second line) to (fourth line)
<3>: (5th line)
<4>: (6th line) to (7th line)
It becomes.

  Next, an analysis item for analyzing the understanding tendency of the examinee is further selected from the small item L2 extracted from the learning item L1 in the analysis method. That is, when the number of small items L2 is large, it is possible to classify the examinee's answers finely, but on the other hand, the difference for each type becomes small, making it difficult to grasp the characteristics between the types. Therefore, for example, after selecting three sub-items <1> to <3> from four sub-items and obtaining a check pattern of the examinee's answer by an operation described later, further sub-items <2> to <4> You may select and perform the same operation to clarify the difference for each type. In this example, all of the small items <1> to <4> were selected as the analysis items <1> to <4>.

  Next, as shown in FIG. 6, at least one “number of numerical values”, “formula”, “word” from the description part corresponding to the analysis items <1> to <4> of the answer example A2 "Term" such as "sentence" or "sentence" is selected, and CB elements that ask whether or not "term" exists are listed. Here, the CB in which the CB elements are listed is a template for determining whether or not “term” is expressed in the description of the procedure for solving each examinee's answer. Therefore, the CB element is configured in an expression form that can determine whether or not “term” is described in two choices. On the other hand, whether or not “term” is expressed in the description of the solution procedure does not directly ask whether or not a specific item is understood.

  As shown in FIG. 7, here, the presence / absence of “term” is asked by text. For example, CB1 “There is a description of (x + 7)”, a short formula part (x + 7) is selected from the description in the first line of the answer example A2 corresponding to the analysis item <1> (see FIG. 6). Asks in writing whether there is a description of "term". In addition, as for CB0 “No description of numbers / letters”, the presence / absence of description common to all the analysis items <1> to <4> is inquired. In addition, CB11 states that only “3” is written in the answer column, and that only “3” is written instead of “3 and 7” in the “answer” column, not in the answering procedure. I am asking you. Such a judgment of CB element cannot be judged almost different if it is a normal grader, and it is assumed that a plurality of graders have done this for obtaining a check pattern for each answer of each examinee described later. Can also ensure objectivity.

  Next, as illustrated in FIGS. 8 and 9, answer examples A3 to A6 reflecting a specific understanding tendency are created. For example, regarding the sub-item “<4> Answer according to the preconditions from the two solutions obtained”, the answer is that the premise in the question sentence “There must be a number to the left of x” is not noticed. A3 is an example of an answer that does not exclude “7”, which has no number next to it. In addition, regarding the small item “being able to formulate a one-dimensional quadratic equation from the meaning of <1>”, an example of an answer in which the formula for the product is erroneous even though the number directly under x can be understood is A5. Furthermore, regarding the small item “<2> Being able to accurately change the equation of the one-dimensional quadratic equation”, an example of an answer having an error in the sign in the positive and negative products of the one-dimensional quadratic equation is A4. A blank answer example is A6.

Next, as shown in FIG. 10, the CB elements (CB0 to CB11) of the CB are checked with respect to the answer examples A2 to A6, and a combination of correctness of the check (hereinafter referred to as a check pattern) is obtained. Here, “◯” in the figure is a mark (affirmed) that satisfies the condition of each CB element, and “x” is a mark that does not satisfy the condition (denied). Since the check pattern gives two choices of “×” or “◯” to each CB element, if the number of CB elements is n, there are 2 ⁿ check patterns. Since CB element in this embodiment is 11, the number of check pattern is 2 ¹¹ in.

  Here, as will be described later, the CB elements are changed as necessary so that at least the check patterns of the answer examples A2 to A6 to be classified are not the same. In this embodiment, since there is no check pattern that is the same as each other, the CB element is not changed. Let the check patterns of the answer examples A2 to A6 be the types C2 to C6.

Next, for each answer of a plurality of examinees, each CB element is checked to obtain a check pattern and create a database. Check pattern obtained from answers as described above can be classified into up to 2 ¹¹ in. However, since the CB elements created in a single problem have not a few relations with each other, the check results between specific CB elements are related. Therefore, the number of check patterns is much smaller than the maximum number described above. In this embodiment, there are five check patterns to be classified into answer examples A2 to A6. In order to secure five or more check patterns, the number n of CB elements needs to be 5 ≦ 2 ⁿ , that is, n ≧ 3.

  Next, the number of answers that have the same check pattern as the check patterns of types C2 to C6 is obtained. Candidates' understanding tendency can be analyzed from the ratio between the number of these answers and the ratio of each answer to the total number of answers. For example, an answer showing the same check pattern as the check pattern of type C5 is directly below x with respect to the same tendency as in answer example A5, that is, with respect to the sub-item “It is possible to formulate a one-dimensional quadratic equation from <1> subject matter”. Even if you can understand the number, it is estimated that the degree of comprehension of the product is wrong. Similarly, the same answer as the check pattern of type C4 has the same tendency as in answer example A4, that is, the correctness of the one-dimensional quadratic equation with respect to the sub-item “<2> the ability to accurately transform the one-dimensional quadratic equation”. It can be estimated that the degree of comprehension is such that it leads to an expression transformation with an error in the sign in the negative product. If the ratio of the former to the latter in the total number of answers is obtained, the understanding tendency of all examinees can be grasped.

  As described above, in this embodiment, the CB element asks whether or not the short term is described, and the checker can objectively and quickly determine the presence or absence of the term in the test taker's answer. . On the other hand, whether or not there is a short “term” answer does not directly ask whether or not a specific item is understood. In other words, by checking the examinee's answer with a plurality of CB elements related to the learning item, obtaining a check pattern, and comparing this with a check pattern with a specific answer example having a different understanding level, a predetermined probability is obtained. Mutual correlation can be obtained. That is, it is possible to classify the degree of understanding of the learning items of each answer. Therefore, it is possible to obtain an understanding tendency about a specific learning item of the population, and thus it is possible to extract an educational problem.

  Although the analysis method is as described above, once the analysis result is obtained, it is possible to continue the analysis by adding an example of the answer to the answer that has not been classified into the types C2 to C6.

  As shown in FIG. 11, the answer example A7 is an answer example having an error in the modification of the intermediate equation of the one-dimensional quadratic equation with respect to the small item <2> as in the answer example A4. Specifically, the answer example A4 has an incorrect factorization, and the answer example A7 has an incorrect expression expansion. At this time, as shown in FIG. 12, the check pattern for the answer example A7 is different from any of the check patterns of types C2 to C6. Therefore, the check pattern for the answer example A7 can be classified into a type C7, and a part of the answers that are not classified into the types C2 to C6 can be classified into a type C7. Note that the answer classified into the type C7 can be estimated as an answer that tends to cause an error in the expansion of the expression for the small item <2> as described above.

  Furthermore, a plurality of types that are classified into types of answers that are estimated to have similar understanding levels for learning items may be combined into one type.

  As shown in FIG. 13, CB4 to CB8 are CB elements created by terms selected from the description regarding the second item to the fourth row of the answer example A2, which is a typical answer example, that is, the small item <2>. is there. It is estimated that an example of an answer including at least one CB element that is “x” in CB4 to CB8 lacks at least a part of understanding regarding the subitem <2>. In order to categorize the answers that receive such estimation, as categorization CC4, CB4 to CB8 may be created with “at least one place as x” as in Note 1.

  Furthermore, in consideration of the association for each CB element, a plurality of types classified into answer types may be combined into one type.

  As shown in FIG. 13, if the expression is mistakenly modified, the subsequent description should normally be different from the typical answer example A2 or not described. On the other hand, the CB element does not ask for the presence or absence of the description of the entire formula, but asks for the presence or absence of a short “term” that is a part of it, and by chance includes the same “term” as the typical answer example A2. There may be cases. In the answer examples including “x” in CB4 to CB8, CB9 to CB11 are originally different from the typical answer example A2 or should not be described. Therefore, the check results of CB9 to CB11 may be any, and a type represented as “−” may be created as the type CC4. The same applies to the type CC5.

  Next, one embodiment of an answer analysis processing system for executing the analysis processing method will be described with reference to FIG.

  The answer analysis processing system 1 includes a main input device 3, an alarm means 4 and a data storage device 5 connected to the central control unit 2, and a plurality of terminal devices 6 used by the CB scorer. The data storage device 5 includes a CB data storage unit 8 that stores CB data, an answer data storage unit 9 that stores image information or read information of test taker's answer data, and a personal data storage unit that stores personal data such as a CB pattern. 10 The main input device 3 and the terminal device 6 include an image display device 14 and input means 15 such as a keyboard, respectively. These are connected to the central control unit 2 through an Internet line, a LAN line or the like so as to be capable of bidirectional communication.

  Next, a method of using the answer analysis processing system 1 will be described with reference to FIG.

  The created CB data is input from the input means 15 of the main input device 3 by a person who supervises the answer analysis process (hereinafter referred to as a supervisor), and is transferred and stored in the CB data storage unit 8. The answer examples A2 to A6 are also input from the input means 15 of the main input device 3, and are transferred and stored as answer example data in the answer data storage unit 9. Note that CB data and answer example data can be stored as image data by using the input means 15 by a scanner or the like.

  Next, when the supervisor operates the main input device 3, the CB data from the CB data storage unit 8 and the answer example data from the answer data storage unit 9 are transferred to the main input device 3, and are transferred to the image display device 14. Is displayed. The supervisor operates the input means 15 to input “Yes” or “No” of “Term” in two choices of “O” and “X” for each CB element. When input indicating completion of input for all CB elements is performed, a check pattern which is a series of data “O” and “X” is transferred and stored in the personal data storage unit 10 together with a code for identifying the answer examples A2 to A6. Is done.

  When the input of the check pattern is completed for each of the answer examples A2 to A6, the central control unit 2 causes the image display device 14 of the main input device 3 to display the check patterns of the answer examples A2 to A6. When the same check pattern exists in the answer examples A2 to A6, the central control unit 2 displays a warning on the image display device 14 of the main input device 3 and sends a warning from the warning means 4. In response to this, the supervisor operates the main input device 3 to erase the CB data once and re-input the CB element. On the other hand, if the same check pattern does not exist, the central control unit 2 causes the CB data storage unit 8 to rewrite the check patterns corresponding to the answer examples A2 to A6 as types C2 to C6, respectively.

  On the other hand, the question sentence Q2 is given to the examinee, and the answer in which the examination procedure of each examinee is described is collected. By the supervisor, the answer is converted into image data by an electronic information processing means (not shown) and saved as answer data together with a name and an examination number for identifying the examinee in the answer data storage unit 9.

  Next, when a scoring start signal is sent from each terminal device 6 to the central control unit 2, the central control unit 2 sends CB data from the CB data storage unit 8 to each terminal device 6 and from the answer data storage unit 9. 1 answer data is sent out. The CB data and the answer data are displayed on the image display device 14 of the terminal device 6, and input for each CB element of the CB becomes possible. The grader operates the input means 15 of the terminal device 6 to input the presence / absence of “term” for each CB element. When the input indicating that the input to all the CB elements is completed, the check pattern is transferred and stored in the personal data storage unit 10 through the central control unit 2 together with a code for identifying the examinee.

  A check pattern is obtained for the answers of all examinees, and a database is created in the personal data storage unit 10. The database analysis has been described above and will not be described in detail.

  Note that the central control unit 2 may have a function of issuing a warning to the terminal device 6 so as to urge correction of an input of a clearly contradictory check pattern. For example, when CB0 “No description of numbers and characters” is checked, the CB element that asks for other description contents cannot be checked. A warning can be issued to re-input to the device 6, and input errors can be reduced.

It is a flowchart which shows the answer analysis processing method by this invention. It is a flowchart which shows the answer analysis processing method by this invention. It is a figure about the answer analysis processing method by this invention. It is a figure about the answer analysis processing method by this invention. It is a figure about the answer analysis processing method by this invention. It is a figure about the answer analysis processing method by this invention. It is a figure about the answer analysis processing method by this invention. It is a figure about the answer analysis processing method by this invention. It is a figure about the answer analysis processing method by this invention. It is a figure about the answer analysis processing method by this invention. It is a figure about the answer analysis processing method by this invention. It is a figure about the answer analysis processing method by this invention. It is a figure about the answer analysis processing method by this invention. It is a figure of the answer analysis processing system by this invention.

Explanation of symbols

DESCRIPTION OF SYMBOLS 1 Answer analysis processing system 2 Central control part 3 Main input device 4 Alarm means 5 Data storage device 6 Terminal device 8 CB data storage part 9 Answer data storage part 10 Personal data storage part 14 Image display apparatus 15 Input means L1 Learning item L2 Small Item Q2 Question sentence CB Check boxes A1 to A7 Sample answers

Claims (5)

An answer analysis processing method for providing a question sentence including a plurality of information elements to a plurality of examinees and requesting a description of a solution procedure for the information to analyze an understanding tendency of a predetermined learning item of the whole examinee,
A part corresponding to a plurality of small items obtained by subdividing the learning item from a typical answer example which is one answer example of the question sentence, and a set of numerical values, formulas, words, or A check box creation step that creates check boxes by listing check elements consisting of sentences,
Create a plurality of other answer examples consisting of description of the solution procedure giving a difference in the degree of understanding about the predetermined learning item, and for each of the check element of the check box for each of the typical answer example and the answer example A correction step for obtaining an answer example pattern that is a combination of presence and absence and correcting the check elements so that they are not identical to each other;
A personal recording step for obtaining a personal pattern that is a combination of the presence or absence of each check element of the check box and recording the personal pattern together with an identification code for identifying the test taker for the description of the solution procedure for each examinee ,
An answer analysis processing method comprising: a counting step of counting the number of examinees having the same personal pattern.   2. The answer analysis processing method according to claim 1, wherein the counting step includes a step of counting the number of examinees who make the personal pattern the same as the answer example pattern. An answer analysis processing system for giving a question sentence including a plurality of information elements to a plurality of examinees and analyzing the understanding tendency of a predetermined learning item of the whole examinee by requesting description of a solution procedure for the information,
A part corresponding to a plurality of small items obtained by subdividing the learning item from a typical answer example which is one answer example of the question sentence, and a set of numerical values, formulas, words, or A check box display means for accepting and displaying a check element consisting of a sentence;
The check element of the check box displayed by the check box display means that allows an input for specifying the test taker and generates an identification code in accordance with the description of the solution procedure for each test taker A check unit that allows input of presence / absence every time and generates a personal pattern that is a combination of the presence / absence, and records the personal pattern as a personal record together with the identification code;
An answer analysis processing system comprising: a totaling unit that counts the number of examinees having the same personal pattern.   Create a plurality of other answer examples consisting of description of the solution procedure giving a difference in the degree of understanding of the predetermined learning item, and allow input for specifying the typical answer example and the answer example, and accordingly An identification code is generated, an input of the presence / absence of each check element of the check box displayed by the check box display means is allowed, and an answer example pattern that is a combination of the presence / absence is generated, together with the identification code, the answer 4. An answer example check means for recording an example pattern, and an alarm means for urging correction of the check element when the answer example patterns are the same between the answer examples. Answer analysis processing system.   5. The answer analysis processing system according to claim 4, wherein the counting means totals the number of examinees who make the personal pattern the same as the answer example pattern.