JP5234543B2 - Answer classification processing system and answer classification processing method - Google Patents

Description

  The present invention relates to an answer classification processing system and an answer classification processing method for classifying answers of a plurality of examinees, and in particular, a check box sheet listing a plurality of alternative question sentences regarding measurement items reflected in a problem The present invention relates to an answer classification processing system and an answer classification processing method for classifying answers of a plurality of examinees.

  In entrance examinations at various schools, a check sheet that lists multiple questions related to specific contents that should be included in the answer to a given question is used to determine whether there is a match or difference with the model answer. Therefore, a scoring method for efficiently scoring examinees' answers is adopted. For example, for each specific content that forms part of the model answer, list multiple question sentences asking whether this is substantially described in the examinee's answer, and the grader will determine this for each question sentence. Give a score. The total score is used to score the answer and determine the superiority or inferiority of the examinee.

  By the way, when scoring such an answer by a plurality of graders, since there are not a few variations in judgment among the graders, the grade becomes inaccurate as the number of graders increases. It was.

  In relation to the above, Patent Document 1 discloses a scoring system that unifies scorers 'judgment criteria and attempts to score examinees' answers more accurately. The scoring results of one grader can be corrected on the system, and if the scoring results do not match among multiple scorers and there is a question about the objectivity of the scoring results, the judgment grounds of other scorers You can reevaluate the scoring results while referring to, and correct them. By sharing the scoring results and judgment criteria among multiple scorers, it is possible to re-evaluate the answers from various viewpoints without being constrained by the subjective scoring results performed by themselves.

On the other hand, in contrast to the above-mentioned exams that determine the superiority or inferiority of examinees, academic achievement exams that establish national and local government educational guidelines, in many ways, demonstrate the tendency of examinees to understand specific learning items. To measure educational issues and extract educational issues. Here, rather than “scoring” that determines the superiority or inferiority of the examinee, it is required to “classify” the examinee's answers into groups related to understanding tendencies. Furthermore, in such a classification test, the number of examinees on the scale of tens of thousands, hundreds of thousands, and sometimes even millions of people, and a “descriptive” test is performed to classify the examinees' answers. It may be necessary to identify educational issues. It is impossible to process such a large number of “descriptive formulas” with only a few measurers due to time constraints, and many measurers with various variations in judgments are adopted and shared and measured. Need to do. Then, since these measurement results must be integrated to classify the examinee's answers, the classification tends to be inaccurate.
JP 2006-277086 A

  In the above-described classification test, as in the system disclosed in Patent Document 1, it is difficult to unify judgment criteria when the number of measurers increases even if an attempt is made to unify judgment criteria by referring to the judgment basis of other measurers. It is. In the classification of answers by many measurers who are difficult to unify the judgment criteria, the classification of the answers becomes inaccurate as the number of measurers increases. Moreover, the work of referring to the judgment basis of other measurers may significantly reduce the efficiency of the answer measurement process.

  Further, even in the above-described classification test, a classification processing method for efficiently classifying the examinee's answers using a check sheet listing a plurality of question sentences is considered. However, the check sheet is a tool for measuring the relative similarity of the answer to one answer example such as “exemplary answer”, and the processing efficiency is not necessarily improved in the classification test classified into a plurality of answer examples. I can not say. In other words, in a classification test, a check sheet is created for each of a plurality of answer examples to be classified, and each answer is checked with these, depending on which answer example is relatively similar to each answer. The entire answer is classified. Therefore, the check sheet must be checked as many times as the number of answer examples to be classified into the number of answers, and as a result, the processing efficiency is never high.

  The present invention has been made in view of the situation as described above. The purpose of the present invention is to provide a method for classifying examinees' answers by a check sheet listing a plurality of question sentences. To provide an answer classification processing system and method capable of efficiently and accurately classifying an answer.

  The inventor of the present application creates a plurality of measurement items in advance by predicting a candidate's understanding tendency with respect to a specific learning item in the classification test as described above, so that these can be described in the examinee's answer, Focusing on the fact that a problem sentence has been created, the present invention has been achieved.

  That is, the system according to the present invention is an answer classification processing system for classifying a plurality of examinees' answers according to a check box sheet that lists a plurality of alternative questions regarding measurement items reflected in a problem, An answer storage area for storing the answer as image data, a check box sheet storage area for storing question sentence data of the check box sheet, and a processing result for storing a two-choice result for the question sentence of the answer A server computer having a storage device including a storage area, and a plurality of client computers connected to the server computer, wherein the server computer allows the creation and correction input of the question text in the check box sheet The question sentence data created or modified by the storage device Over the check box sheet storage area of the check box sheet, and further to the client computer in response to a processing request signal from the client computer, the question sentence data of the check box sheet from the check box sheet storage area of the storage device And one of the image data of the answer is returned from the answer storage area, and the client computer transmits the processing request signal to the server computer and responds accordingly to the question sentence of the answer that is returned The result is allowed to be input, and the two-choice result is transmitted to the server computer and stored in the processing result storage area of the storage device.

  According to such a system, the answers of a plurality of examinees are classified by a single check box sheet that lists a plurality of alternative questions regarding the measurement items reflected in the question, and the question sentences on the check box sheet However, it is possible to proceed with the classification process while the measurer corrects the two-choice results for the specific answer so as not to vary. Therefore, even if measurement is performed by many measurers, the classification process can be performed accurately. In other words, many examinees' answers can be classified more accurately and efficiently.

  In the system described above, the server computer returns one of the image data of the answer stored in the answer storage area of the storage device to a plurality of the client computers, and the processing result storage area of the storage device When a plurality of two-choice results are compared and these are different, the correction of the question text in the check box sheet is allowed and the corrected question text data is overwritten in the check box sheet storage area of the storage device It is characterized by that.

  According to such a system, it is possible to proceed with the classification process while correcting the question text on the check box sheet so that a specific answer among a plurality of measurers, that is, a two-choice result for one image data is always the same. Therefore, even if measurement is performed by many measurers, the classification process can be performed accurately. In other words, many examinees' answers can be classified more accurately and efficiently.

  Further, in the other system described above, the server computer returns one of the unprocessed image data of the answer stored in the answer storage area of the storage device to one of the client computers, and the server computer. Allows the input of the two-choice result for the question sentence of the answer, and compares the two-choice result of the processing result storage area of the storage device and when these are different, the question sentence of the check box sheet The question sentence data that allows correction input and is overwritten in the check box sheet storage area of the storage device may be overwritten.

  According to such a system, the classification process is performed while correcting the question sentence data in the check box sheet so that a specific answer, that is, the two-choice result for one image data is always the same between the problem creator and the measurer. Can proceed. Therefore, even if measurement is performed by many measurers, the classification process can be performed accurately. In other words, many examinees' answers can be classified more accurately and efficiently.

  Furthermore, the method according to the present invention is an answer classification processing method for classifying answers of a plurality of examinees by a check box sheet that lists a plurality of alternative questions regarding measurement items reflected in a problem, An answer storage area for storing the answer as image data, a check box sheet storage area for storing question sentence data of the check box sheet, and a processing result for storing a two-choice result for the question sentence of the answer Preparing a server computer having a storage device including a storage area, and a plurality of client computers connected to the server computer, and allowing the creation of the question text in the check box sheet in the server computer. The question sentence data created thereby is stored in the storage device. Storing in the check box sheet storage area, and the question sentence data and the answer of the check box sheet from the check box sheet storage area of the storage device toward the client computer in response to a processing request signal from the client computer Returning one of the image data of the answer from the storage area; and allowing the correction input of the question text in the check box sheet to allow the correction of the question text to be stored in the check box sheet of the storage device Overwriting the area.

  According to such a method, the answers of a plurality of examinees are classified by a single check box sheet that lists a plurality of alternative questions regarding the measurement items reflected in the question, and the question sentences on the check box sheet. However, it is possible to proceed with the classification process while being corrected so as not to cause variation in the two-choice results for the specific answer by the measurer. Therefore, even if measurement is performed by many measurers, the classification process can be performed accurately. In other words, many examinees' answers can be classified more accurately and efficiently.

  In the method described above, the client computer allows the step of transmitting the processing request signal to the server computer, and the input of the two-choice result for the question sentence of the answer returned in response to the step. Transmitting the result to the server and storing the result in the processing result storage area of the storage device.

  According to this method, since the input process of the two-choice result can be performed only by the client computer that has transmitted the processing request signal to the server computer, the input process of the two-choice result is surely performed according to the processing speed of the measurer. I can do it. In other words, as a result, the answers of many examinees can be classified more accurately and efficiently.

  In the above method, in the server computer, a step of returning one of the image data of the answer stored in the answer storage area of the storage device to a plurality of the client computers; and the processing result storage area of the storage device A step of comparing a plurality of two-choice results and allowing the correction input of the question sentence in the check box sheet to be overwritten in the check box sheet storage area of the storage device when they are different from each other. It may be characterized by including.

  According to such a method, it is possible to proceed with the classification process while correcting the question text on the check box sheet so that a specific answer among a plurality of measurers, that is, a two-choice result for one image data is always the same. Therefore, even if measurement is performed by many measurers, the classification process can be performed accurately. In other words, many examinees' answers can be classified more accurately and efficiently.

  In the above method, in the server computer, one of the image data of the unprocessed answer stored in the answer storage area of the storage device is returned to one of the client computers, and at the same time, the server computer also A step of allowing the input of the two-choice result for the question sentence of the answer, and comparing the inputted two-choice result with the two-choice result of the processing result storage area of the storage device, and when these are different, the check box sheet And allowing the correction input of the question sentence to be overwritten and overwriting the corrected question sentence data in the check box sheet storage area of the storage device.

  According to such a method, the classification process is performed while correcting the question on the check box sheet so that the specific answer between the question creator and the measurer, that is, the two-choice result for one image data is always the same. Can proceed. Therefore, even if measurement is performed by many measurers, the classification process can be performed accurately. In other words, many examinees' answers can be classified more accurately and efficiently.

  As described above, in the classification of answers by a large number of measurers who are difficult to unify judgment criteria, the classification of the answers becomes inaccurate as the number of measurers increases, but the check box sheet according to the present invention was used. The answer classification always has a certain accuracy regardless of the number of measurers.

  First, the principle of the answer classification processing method for classifying the answers of a plurality of examinees by a check box sheet listing a plurality of alternative question texts regarding the measurement items reflected in the problem will be described with reference to the drawings. .

As shown in FIG. 1, a learning item α _m is selected from learning items α _{1 to} α _n that are determined to be acquired in a predetermined grade, and a problem Q that is a basis of measurement is created. Candidates' understanding tendency toward this specific learning item α _m , that is, subdivided learning elements (for example, multiplication, addition, factorization, equation formulation) necessary to acquire the specific learning item α _m , Chinese character, phrase, and various grammatical) to predict the good or weak such understanding of the entire candidates per a ₁ ~a _n, determining a plurality of measurement items a ₁ ~a _m from the learning elements . Problem Q is generated while sequentially modifying the setting conditions to include measurement items a ₁ ~a _m. Along with this, a problem that may be included description of measurement items a ₁ ~a _m in the candidate's answers solving Q, details such as the representation of the problem statement q about the problem Q is fixed and determined.

By the way, in the classification test, it is necessary to predict in advance the examiner's understanding tendency with respect to the specific learning item α _m described above. For example, if the candidate's understanding tendency cannot be predicted, there will be a large difference between the candidate's understanding tendency and the question Q to be asked, and the target answer cannot be classified. For example, in an extreme sense, if the so-called “difficulty” of the problem is extremely high compared to the examiner's understanding tendency, many examinees' answers will be blank.

The measurement items higher similarity a ₁ ~a _m content between, although "texture" Understanding tends to be measured of candidates for a particular learning item alpha _m finer, candidates to be described later too high The answer classification is ambiguous and the measurement accuracy is not necessarily high. Conversely, if the similarity measurement items a ₁ ~a _m content of each other too low, although the classification of answers will become clear, "texture" is rough understanding tends to be measured.

As shown in FIG. 2, wherein for each measurement item _a 1 ~a _m in the candidate's answer, i.e. the candidate's measurement items _a 1 ~a each solution by assuming each handle at _m Example _MA 1 to MA _n create. As noted above, a question sentence q is because assumed to have been created a description of measurement items a ₁ ~a _m in the candidate's answer, conversely, always processes the candidates in the measurement item a ₁ ~a _m Answer examples MA _{1 to} MA _n can be created. At this time, for example, if the description related to the measurement item a ₁ affects the description related to the measurement item a ₂ described later, an answer example reflecting this is created. This means that each example solution _MA 1 to MA _n, the combination of the process in a plurality of measurement items _a 1 ~a _m is reflected.

Referring now also to FIG. 3, measuring whether answers _TA 1 to Ta _n of each examinee are most relatively similar to any of the plurality of sample answers _MA 1 to MA _n by a check box seat 50 And classify.

As shown in FIG. 4, the check box sheet 50 is a question sheet in which a plurality enumerate alternative question sentence _CQ 1 _{~CQ n} respect to the measurement items _a 1 ~a _n reflected in question. Here, the question sentences CQ _{1 to} CQ _n of the check box sheet 50 can be checked by at least two alternatives, and the measurement discretion of the content by the measurer is reduced, and any measurer at a certain level is selected. It is preferable that the measurement is made so that the same judgment result can be obtained for the same answer. Although the details will be described later, the question sentences CQ _{1 to} CQ _n do not ask whether or not specific contents are substantially described in the examinee's answers TA _{1 to} TA _n , but whether or not the contents are described directly. What is asked is preferred.

For example, in terms of questions from a formal aspect, in the case of mathematics, regarding the choice of solutions and formulas, in the case of language, regarding the choice of state and syntax such as passive and active, the description of sentences as not depending on a specific subject field There are things that ask about the presence and quantity of the. In addition, as the contents of questions from a linguistic side, there are things that ask for specific numerical values and expressions in the case of mathematics, and specific words and collocations in the case of language. Two-option results for each question _CQ 1 _{~CQ n} checkbox sheet 50, example solution _MA 1 to to MA _n different for each, question contents as the original question _CQ 1 _{~CQ n} is repeatedly changed . In addition, in the question content of the linguistic aspect described above, the answer examples can be compared with each other and the question content can be created or changed from the difference. After confirming the question contents, question sentences CQ _{1 to} CQ _n are created correspondingly.

Thus, the check box seat 50 in which a plurality of lists are created to choose between question _CQ 1 _{~CQ n} terms reflect the measurement item _a 1 ~a _m problem. Further, the question sentences CQ _{1 to} CQ _n are appropriately modified so that any measurer can obtain the same determination result as long as the measurer has a certain level of the two-choice results for a specific answer.

Two-option results for question _CQ 1 _{~CQ n} in answer _TA 1 to Ta _n of each examinee is collected and compared with the two-option results for question _CQ 1 _{~CQ n} in example solution _MA 1 to MA _n. In other words, answer TA ₃ having the same two-option results as example solution MA ₁ is than it is classified into groups in which the same determination as example solution MA ₁ in the measurement item _a 1 ~a _m. Based on the number of answers classified into each of the answer examples MA _{1 to} MA _n , etc., the examiner's understanding tendency for a specific learning item α _m is determined, and an educational problem is extracted.

  Next, the details of the answer classification processing system 1 for performing the process using the above-described check box sheet 50 and the method of classifying the answers for each examinee by this will be described.

[system]
As shown in FIG. 5, the answer classification processing system 1 is used by a plurality of client computers 21-1 to 21-n and a problem creator connected by a communication line 20 such as an Internet line or a LAN (Local Area Network). A server computer 22 is included. The server computer 22 includes a storage device 11 having a question sentence storage area 12, a processing result storage area 13, an answer storage area 14, a check box sheet storage area 15, and the like. The server computer 22 includes a monitor 22a and a keyboard 22b, and an input device such as a scanner (not shown).

[Create question]
As shown in FIG. 6, for example, a problem Q serving as a basis for measurement is created for a learning item “three square theorem” that is determined to be acquired in a predetermined grade. Here, the learning elements are (1) deriving the formula of the three-square theorem, (2) solving the one-dimensional quadratic equation, (3) accurately performing addition including exponents, (4) There are many factors such as performing the shift accurately. From these, two points were determined as measurement items: (1) deriving the formula of the three-square theorem, and (2) solving the one-dimensional quadratic equation.

  Therefore, regarding the problem Q, regarding the measurement items (1) and (2), the length of two sides sandwiching the right angle of the right triangle is given concretely, and the formula of the three square theorem is described, and then this one-way two-way Create a problem that solves the following equation to determine the length of the hypotenuse. At the same time, a specific numerical value and a specific figure shape are given so that the description of the measurement items (1) and (2) can be included in the answer of the examinee who solves the question Q. "If the sides are 1 cm and 2 cm, respectively, find the length of the hypotenuse." Furthermore, in order for the examinee to calculate the three-square theorem and to give an answer, "If the length of the two sides sandwiching the right angle of the right triangle is 1 cm and 2 cm, respectively, the length of the hypotenuse is xcm. Find the x of the time by deriving the equation. ”The question sentence was changed, and the problem sentence q was obtained by encouraging the description of the process of deriving the answer using a one-dimensional quadratic equation. The question sentence q is input from the keyboard 22b of the server computer 22 or a scanner (not shown) and stored in the problem sentence storage area 12 of the storage device 11.

Next, as shown in FIG. 7, answer examples MA _{1 to} MA ₅ were prepared assuming the processing of the examinee regarding the measurement items (1) and (2). The answer examples MA _{1 to} MA ₅ are inputted from a scanner or the like (not shown) of the server computer 22, and image data identification codes A 1 to A 5 are attached respectively and stored in the answer storage area 14 as image data.

Furthermore, as shown in FIG. 8, alternative question sentences CQ _{1 to} CQ ₅ relating to the measurement items (1) and (2) are created, and a check box sheet 50 listing them is obtained. Question sentence CQ ₁ is a measurement item (2), and question sentence CQ ₅ is a question content from a formal aspect regarding measurement items (1) and (2). The question sentences CQ _{2 to} CQ ₄ are the question contents from the linguistic side regarding the measurement item (1). In checkbox sheet 50, the check column ₅₂ 1-52 ₅ are provided corresponding to the question _CQ 1 _{~CQ 5.} The check box sheet 50 is stored in the check box sheet storage area 15 as data.

As described above, the two-choice results for the question sentences CQ _{1 to} CQ ₅ in the check box sheet 50 are different for each of the answer examples MA _{1 to} MA ₅ . The check pattern of the two-choice result (combination of presence / absence of check) is as follows in order when “Yes” is displayed with check and “×” is not checked.
MA ₁ : (○, ○, ○, ×, ○)
MA ₂ : (○, ○, ×, ×, ○)
MA ₃ : (×, ○, ×, ×, ○)
MA ₄ : (○, ×, ×, ○, ○)
MA ₅ : (x, x, x, x, x)
The two-choice results are stored in the processing result storage area 13 of the storage device 11 with image data identification codes A1 to A5.

After finishing the above, a question sentence q was given to _j examinees, and examinees' answers TA _{1 to} TA _j were obtained. These are input from a scanner (not shown) of the server computer 22 and stored in the answer storage area 14 as image data with image data identification codes A6 to A5 + j.

[Initial setting]
First, as illustrated in FIG. 9, a transmission setting file 53 for setting the number of transmissions to the client computer 21 for each image data is created in a predetermined area of the storage device 11.

  Specifically, the number of transmissions N (x) to the client computer 21 for each image data corresponding to the image data identification code Ax (1 ≦ x ≦ 5 + j) (however, this value is less than or equal to the total number of client computers 21). ) Is input from the keyboard 22b of the server computer 22. That is, the image data input with “0” is not transmitted to the client computer 21.

Here, in order to correct the question _CQ 1 _{~CQ 5,} for example solution _MA 1 to MA ₅ 2 single sample answers arbitrarily selected from MA ₁ and MA ₃ (corresponding image data identification code A1 and A3), A case will be described in which the two-choice results for the question sentences CQ _{1 to} CQ ₅ of the check box sheet 50 are collected by 20 of the n (≧ 20) client computers 21. The administrator operates the keyboard 22b of the server computer 22 to input “20” in each of N (1) and N (3), and input “0” in the other transmission count column N (x). .

[ _First Modification of Question Texts CQ _{1 to} CQ _n ]
As described above, the question sentences CQ _{1 to} CQ ₅ of the check box sheet 50 can be checked by at least one alternative, and any measurer can answer the same answer as long as the measurer has a certain level. It is created so that the same judgment result can be obtained. That is, the question sentences CQ _{1 to} CQ ₅ are corrected by the question creator so that the two-choice result for one answer example MA _m is the same as the two-choice result obtained at the time of question creation, regardless of the measurer. Is done.

[1] Transmission / Reception of Image Data in Server Computer As shown in FIG. 10, transmission and recording of image data corresponding to the image data identification code Ax (1 ≦ x ≦ 5 + j) to each of the client computers 21-1 to 21-n is performed in a matrix M. (1, 1) to M (n, 5 + j). That is, the presence / absence of transmission of image data corresponding to the image data identification code Aq to the client computer 21-x is recorded in the matrix M (x, q). In the matrix to which the corresponding image data has not been transmitted yet, 0 in the initial state is recorded, and when the image data is transmitted, 1 is overwritten. Further, in accordance with the transmission setting file 53 described above, a “*” mark is recorded as inactive (non-rewritable) for a corresponding matrix to which image data is not transmitted. In the right column of the matrix, the number of transmissions NN (x) (1 ≦ x ≦ 5 + j) that has already been transmitted for each image data is automatically tabulated (see also FIG. 11).

  Referring back to FIG. 9, in the transmission setting file 53, the image data whose transmission count N (x) is 1 or more is transmitted to N (x) client computers. That is, as shown in FIG. 11, matrices M (1,1) to M (n, 1) and M (1,3) to M (n, 3) related to the image data identification codes A1 and A3 in the transmission management file 60. The matrix can be transmitted actively (rewritable), and the initial state of no transmission = 0 is recorded. For the other matrices, “*” mark is recorded as inactive. At this time, the transmission counts NN (1) and NN (3) that have already been transmitted are 0, and the others are *.

  Using the transmission management file 60 shown in FIG. 11, the data processing routine S1, the processing data return routine S12, and the processing data storage routine S14 are performed as shown in FIGS. Are sent and received.

  As shown in FIG. 12, in the data processing routine S1, when the server computer receives the processing request signal D1 from the specific client computer 21-U (S11, Y), the process proceeds to the processing data return routine S12. On the other hand, when the server computer receives the processing result transmission signal D2 from the specific client computer 21-V (S13, Y), the processing data storage routine S14 is performed. The processing request signal D1 and the processing result transmission signal D2 respectively include client identification codes U and V for identifying the client computer that transmitted the signal.

  As shown in FIG. 13, in the process data return routine S12, image data to be transmitted to the client computer 21-U is searched using the transmission management file 60 (see FIG. 11). Referring also to FIG. 11, for M (U, I) (I is a parameter and an integer of 1 to 5 + j) corresponding to the client identification code U included in the processing request signal D1, I = 1. In order, that is, the U column of the transmission management file 60 is searched from top to bottom, M (U, I) is 0 (S12-4, Y), and the number of transmissions NN (I) that has been transmitted is set to transmit. When the number of transmissions N (I) set in the file 53 has not been reached (S12-5, Y), the image data corresponding to M (U, I) is retrieved from the answer storage area 14 and the check box sheet 50 is stored. Based on the client identification code U, the data is transmitted from the check box sheet storage area 15 to the client computer 21-U (S12-6). At the same time, “1” for recording that transmission has been completed is recorded in the matrix M (U, I) (S12-7). This process is repeated, and when the parameter I becomes 5 + j, that is, when the image file to be transmitted cannot be searched even when reaching the bottom of the U column of the transmission management file 60 (S12-3, N), Based on the client identification code U, a processing unnecessary signal D3 is transmitted to inform the client computer 21-U that there is no image data that can be returned (S12-8). Thereafter, the process returns to the data processing routine S1, and the repeated processing request signal D1 is received. On the other hand, in the data processing routine S1, the number of image data to be transmitted set in the transmission setting file 53, that is, the total number of N (I) reaches the total number of transmissions that have been transmitted, that is, the total number of NN (I). Is no longer received (S15, Y).

On the other hand, as shown in FIG. 14, in the processing data storage routine S14, the two-choice result for the check box sheet 50 (question sentences CQ _{1 to} CQ ₅ ) received from the client computer 21-V is used as the client identification code V and the image data identification. The processing result storage area 13 of the storage device 11 is attached with a reference numeral (S14-1), returns to the data processing routine S1, and receives the repeated processing result transmission signal D2. In the data processing routine S1, when the number of receptions of D2 has reached the number of image data to be transmitted set in the transmission setting file 53, that is, the total number of N (I) and reception of the processing result is completed, The processing result transmission signal D2 is no longer received (S16, Y).

[2] Image Data Transmission Request and Reception by Client Computer The client computer 21 requests the server computer 22 to transmit image data to be processed and the check box sheet 50, and check the image data returned in response thereto. The sheet 50 is processed and the processing result is transmitted to the server computer 22.

  Specifically, as shown in FIG. 15, in the client processing routine S2, the client computer 21-U transmits a processing request signal D1 including a client identification code U for identifying it to the server computer 22 ( S20) Waiting for the return of data from the server computer 22 (S21).

  If the returned data is attached with the processing unnecessary signal D3 (S22, Y), a message indicating that there is no image data to be processed, for example, “No data to be processed” is displayed (S23). The process is terminated.

On the other hand, when the image data and the check box sheets 50 (question sentences CQ _{1 to} CQ ₅ ) are returned from the server computer 22 (S22, N), the processing data is displayed on the screen as shown in FIG. 16 (S24). ). Then, a check input by the measurer to the check column 52 of the check box sheet 50 is accepted (S25). Input is performed by moving the pointer 30 on the screen by operating an input device such as a mouse.

  When the input completion button 31 and the data transmission button on the screen are operated by the pointer 30 and the fact that the check is completed is input (S26, Y → S27, Y), the processing result transmission signal D2 including the client identification code is processed. The two-choice result attached with the image data identification code of the image data is transmitted to the server computer 22 (S28).

  The client computer 21 repeats this series of operations.

[3] Correction of Question Text in Server Computer As shown in FIG. 17, a verification file 54 for displaying the corrected part of the question text is created in a predetermined area of the storage device 11 and displayed on the monitor 22 a of the server computer 22. To do. Here, the double-option results of sample answers MA ₁ to MA ₅ to the question sentence CQ ₁ ~CQ ₅ created at the same time creating problems are displayed side by side in two-option result display column 35. Below each of the two-choice results, a totaling column 36 is provided, and the results compared with the two-choice results obtained by the client computer 21 stored in the processing result storage area 13 of the storage device 11 are tabulated and displayed. Note that “*” is entered in all the columns corresponding to the image data (image data identification codes A2, A4, and A5) not transmitted to the client computer 21.

Specifically, the two-choice results obtained from the client computer 21 stored in the processing result storage area 13 are sequentially entered into the two-choice results in the corresponding two-choice result display field 35, that is, the two-choice results created at the time of problem creation. And for each question sentence. The number of different two-choice results for the same image data and for each question sentence is displayed in the totaling column 36 of the verification file 54. Here, the two-choice result of the question sentence CQ ₅ for the answer example MA ₃ of the collation file 54 indicates that the one from the two client computers 21 and the one inputted at the time of creating the question are different. Thus, the column in which the number of the total column 36 is 1 or more and the corresponding question sentence column are highlighted (see FIG. 18).

As shown in FIG. 18, “2” in the corresponding totaling column 37 is highlighted, and the answer example MA ₃ and the question sentence CQ ₅ are displayed on the monitor 22 a of the server computer 22. In response to the question sentence “There is one or more formulas written” in CQ ₅ , the calculation example is written in the answer example MA ₃ , so the two-choice result is checked (the two-choice result “◯”). ) And decided at the time of question creation. On the other hand, it is presumed that the two client computers gave different interpretations of the meaning of “expression” to the expression “one or more expressions” in the question sentence CQ ₅ . Therefore, this was corrected to a simpler question sentence CQ ₅ ′ with “expression described” to obtain a corrected check box sheet 50 ′ as shown in FIG. 19. The corrected question sentence CQ ₅ ′ is overwritten in the check box sheet storage area 15.

As mentioned above, although the question sentence CQ ₅ with respect to the answer example MA ₃ is corrected here, the question sentences CQ _{1 to} CQ ₅ may be corrected repeatedly over one or more of the answer examples MA _{1 to} MA _5. .

[Second Modification of Question Texts CQ _{1 to} CQ _n ]
In the “second modification”, two-choice results for the question sentences CQ _{1 to} CQ _n of the check box sheet 50 obtained by the plurality of client computers 21 are collected for _{one of} the examinees' answers TA _{1 to} TA _j. . The question creator or the like again corrects and confirms the question sentences CQ _{1 to} CQ _n having different two-choice results in the plurality of client computers 21.

Although details are omitted, for example, in the above-described initial setting, the transmission setting file 53 (see FIG. 9) so that one image data of the examinees' answers TA _{1 to} TA _j is transmitted to the plurality of client computers 21. The number of times of transmission is input to the number of times of transmission N (x).

  Similarly to the “first modification”, the server computer 22 executes the data processing routine S1 according to the transmission management file 60. That is, the image data of one examinee's answer is transmitted from the answer storage area 14 and the check box sheet is transmitted from the check box sheet storage area 15 to the plurality of client computers 21. On the other hand, the client computer 21 executes a client processing routine S2. As a result, the server computer 22 can obtain a two-choice result on the answers of the same examinee by a plurality of client computers 21. The collation file 54 displays the number of client computers that have obtained different two-choice results with respect to the two-choice results of one examinee's answer. Then, the question sentence CQ obtained with a different two-choice result is corrected, and the corrected question sentence CQ ′ is overwritten in the check box sheet storage area 15.

As described above, the question sentences CQ _{1 to} CQ _n can be corrected repeatedly over one or more of the examinees' answers TA _{1 to} TA _j . By correcting and confirming the question sentences other than the answer examples, it is possible to classify the answers more accurately.

  In the present embodiment, the “second modification” is not performed.

[Answer classification process]
Next, the examinee's answers are classified by the corrected check box sheet 50 ′ (see FIG. 19). The check pattern of each candidate's answer obtained by the check box sheet 50 '(combination of whether or not there is a check) is compared with the check patterns of the answer choices MA _{1 to} MA ₅ and these are compared with each other. The candidate's answer that is the same is classified as an understanding tendency reflected in the answer example.

Specifically, with reference to FIG. 5, the administrator resets the processing result storage area 13 of the storage device 11. After that, the two-choice results for the answer examples MA _{1 to} MA ₅ created at the time of creating the question are stored in the processing result storage area 13. Next, the server computer 22 enters _{1 in} the transmission number column of the image data identification codes A6 to A5 + j corresponding to each of the examinees' answers TA _{1 to} TA _{j in} the transmission setting file 53 (see FIG. 9). Thus, the examinees' answers TA _{1 to} TA _j are transmitted to the client computer 21 once, respectively. Next, the data processing routine S1 (see FIG. 12) is executed to obtain two-choice results for the examinees' answers TA _{1 to} TA _j based on the check box sheet 50 ′. Details are omitted.

Next, a classification routine is executed. In the classification routine, the check patterns of the two-choice results of the examinees' answers TA _{1 to} TA _j are checked in order, and if there is the same check pattern in the answer examples MA _{1 to} MA ₅ , the check patterns are classified. Here, for example, when the same check pattern does not exist in the answer examples MA _{1 to} MA ₅ as in the classification routine S6 shown in FIG. 20, the answers of the same check patterns may be classified together. .

In the classification routine S6, first, a classification file 56 (see FIG. 21), which will be described later, is reset and the “classification destination image data identification code” field is blanked (S60). Next, each of the check patterns (CP) for the examinee's answer of the image data identification codes A6 to A5 + j (AI) is (S61, S69), and the image data identification codes A1 to A5 + j are in this order (S62, S63). The check patterns are compared (S65) until the identification code AI-1 is obtained (S64). When the check patterns match, the latter image data identification code is recorded in the “classification destination image data identification code” column of the former image data identification code of the classification file 56 shown in FIG. Incidentally, when the image data identification code is recorded is A1~A5 indicates that it is identical to the check pattern of the two-option results of sample answers MA ₁ to MA _5. If the check patterns do not match until the check pattern for the examinee's answer of the image data identification code AI is compared with the check pattern for the image data identification code AI-1 (S64, Y), it is shown in FIG. “AI”, which is the image data identification code of itself, is recorded in the “classification destination image data identification code” field of the image data identification code AI of the classification file 56 (S68). Thereafter, if there is the same check pattern, it is classified as “AI”.

  This completes the classification. In addition, as shown in FIG. 22, you may create the total file which filled in the example of the answer used as the classification | category destination, and the number of test takers' answers.

As described above, according to the present embodiment, the question sentences CQ _{1 to} CQ ₅ of the check box sheet 50 can be corrected so as not to cause variation by the measurer in the two-choice results, and the corrected check box sheet 50 is corrected. Candidates 'answers TA _{1 to} TA _j can be classified using'. In addition, one of the answer examples MA _{1 to} MA ₅ is returned to the plurality of client computers 22, and the question sentence CQ of the check box sheet 50 is corrected so that the two-choice results by the plurality of measurers are always the same. Can do. Further, the check box sheet can be modified so that the processing result of the same test taker's answer is the same between the question creator and the measurer. Therefore, the classification process can be accurately performed even if measurement is performed by many measurers. In other words, many examinees' answers can be classified more accurately and efficiently.

  In general, in the classification of examinees 'answers by a large number of measurers whose judgment criteria are difficult to unify, the classification of examinees' answers becomes inaccurate as the number of measurers increases. However, in the present embodiment, the question sentence CQ of the check box sheet 50 can be corrected so as to reduce the determination of the content by the measurer so as not to cause a variation in the determination. Therefore, the accuracy is always constant regardless of the number of measurers. It has.

  In addition, since the next two-choice result input processing is performed sequentially from the client computer that has finished the two-choice result input processing, it is ensured according to the processing speed of each measurer regardless of the processing speed of other measurers. It is possible to perform input processing of two-choice results. In other words, many examinees' answers can be classified more accurately and efficiently.

  The client identification code may be an identification code for identifying the measurer himself. As a result, for example, the measurer may be authenticated by the client computer 21, and any one of the plurality of client computers 21 may be used to measure the answer example or the examinee's answer without duplication of the same image data. Can be implemented.

  As mentioned above, although the typical Example by this invention was described, this invention is not necessarily limited to these, A person skilled in the art will be able to perform various without departing from the gist of the present invention or the scope of the appended claims. Alternative embodiments and modifications can be found.

It is a figure of the principle about the answer classification | category processing method by this invention. It is a figure of the principle about the answer classification | category processing method by this invention. It is a figure about the answer classification processing method by this invention. It is a figure about the answer classification processing method by this invention. It is a figure of the answer classification processing system by this invention. It is a figure about the answer classification processing method by this invention. It is a figure about the answer classification processing method by this invention. It is a figure about the answer classification processing method by this invention. It is a figure about the answer classification processing method by this invention. It is a figure about the answer classification processing method by this invention. It is a figure about the answer classification processing method by this invention. It is a flowchart which shows the answer classification | category processing method by this invention. It is a flowchart which shows the answer classification | category processing method by this invention. It is a flowchart which shows the answer classification | category processing method by this invention. It is a flowchart which shows the answer classification | category processing method by this invention. It is a figure about the answer classification processing method by this invention. It is a figure about the answer classification processing method by this invention. It is a figure about the answer classification processing method by this invention. It is a figure about the answer classification processing method by this invention. It is a flowchart which shows the answer classification | category processing method by this invention. It is a figure about the answer classification processing method by this invention. It is a figure about the answer classification processing method by this invention.

Explanation of symbols

11 storage device 13 processing result storage area 14 answer storage area 15 check box sheet storage area 21 client computer 22 server computer 50 check box sheet 53 transmission setting file 54 collation file 55 summary file 56 classification file 60 transmission management file CQ question sentence MA ₁ -MA _n answer examples TA ₁ -TA _n examinees' answers A1-A5 + j Image data identification code S1 Data processing routine S2 Client processing routine S6 Classification routine S12 Processing data return routine S14 Processing data storage routine

Claims (7)

Multiple test takers for a question created by reflecting the measurement items in a check box sheet that lists multiple alternative questions regarding the measurement items for a specific learning item that is predicted in advance. An answer classification processing system for classifying the answers of
An answer storage area for storing the answer as image data, a check box sheet storage area for storing question sentence data of the check box sheet, and a processing result for storing a two-choice result for the question sentence of the answer A server computer having a storage device including a storage area;
A plurality of client computers connected to the server computer,
The server computer allows creation and correction input of the question text in the check box sheet, and overwrites the question text data created or corrected in the check box sheet storage area of the storage device, and One of the question sentence data of the check box sheet from the check box sheet storage area of the storage device and the image data of the answer from the answer storage area toward the client computer in response to a processing request signal from the client computer Reply
The client computer transmits the processing request signal to the server computer, allows input of a two-choice result for the question sentence of the answer returned in response thereto, and sends the two-choice result to the server computer. Storing in the processing result storage area of the storage device ;
An answer classification processing system , wherein an answer is classified by the check pattern of the two-choice results in the server computer .   The server computer returns one of the image data of the answer stored in the answer storage area of the storage device to a plurality of the client computers, and a plurality of two options of the processing result storage area of the storage device Comparing the results, when these are different, the correction input of the question sentence of the check box sheet is allowed and the corrected question sentence data is overwritten in the check box sheet storage area of the storage device. The answer classification processing system according to claim 1.   The server computer returns one of the unprocessed image data of the answer stored in the answer storage area of the storage device to one of the client computers, and the server computer responds to the question sentence of the answer. The selection result is allowed, and the two choice results in the processing result storage area of the storage device are compared, and when they are different, the correction input of the question sentence in the check box sheet is allowed and corrected. 2. The answer classification processing system according to claim 1, wherein the question sentence data is overwritten in the check box sheet storage area of the storage device. An answer storage area that stores answers as image data; stores question sentence data in a check box sheet that lists multiple alternative question sentences for measurement items for specific learning items predicted in advance of the examinee's understanding tendency A server computer having a storage device including a check box sheet storage area for processing and a processing result storage area for storing a two-choice result of the answer to the question sentence, and a plurality of clients connected to the server computer An answer classification processing method for classifying a plurality of examinees' answers to questions created by reflecting the measurement items, which is executed by a computer ,
In the server computer,
Allowing the creation of the question text in the check box sheet and storing the question text data created thereby in the check box sheet storage area of the storage device;
One of the question sentence data of the check box sheet from the check box sheet storage area of the storage device and the image data of the answer from the answer storage area toward the client computer in response to a processing request signal from the client computer One reply step,
Allowing correction input of the question text in the check box sheet and overwriting the check text data corrected by the check text in the storage area of the check box sheet;
And classifying the answer according to the check pattern of the two-choice results . In the client computer,
Transmitting the processing request signal to the server computer;
Allowing the input of the two-choice result for the question sentence of the answer returned in response thereto, transmitting the two-choice result to the server, and storing the result in the processing result storage area of the storage device. The answer classification processing method according to claim 4, wherein: In the server computer,
Returning one of the image data of the answer stored in the answer storage area of the storage device to a plurality of the client computers;
When a plurality of two-choice results in the processing result storage area of the storage device are compared and these are different, the correction of the question sentence in the check box sheet is allowed and the corrected question sentence data is stored in the storage device. The answer classification processing method according to claim 4, further comprising: overwriting the check box sheet storage area. In the server computer,
One of the image data of the unprocessed answer stored in the answer storage area of the storage device is returned to one of the client computers, and at the same time, the server computer also has a two-choice result for the question sentence of the answer The step of allowing the input of
The question sentence data corrected by allowing correction input of the question sentence of the check box sheet when the two alternative results inputted and the two choice results in the processing result storage area of the storage device are different from each other. 6. The answer classification processing method according to claim 4, further comprising: overwriting the check box sheet storage area of the storage device.