a quick how-to
|
WebClass
a quick how-to |
|
|
|
|
The construction of test itemsThe following tutorial explains the construction of test items and shows how to create them using the "text-to-items" converter. ► Please note that the tutorial was developed for an earlier version of the testing system on WebClass, and many of its features (as well as some of the CSS styling) have now been updated. However, it should give a general idea of how to construct a variety of different test items.
The converter can be run by clicking either the "Text Converter" link at the top of each set or "Import items" at the bottom (and then selecting "Text"). Before these links can appear, we must (at least) create an empty set of items:
Here are the basic steps to follow when converting text into test items:
More details:
Selected response itemsThe first five types (MC, TF, MC-CL, RW, MR) are created by clicking "Add MC":
Matching (MT) items are created in a slightly different way (see below). Also note that if we create classic fill-gaps items and provide all of the possible responses (e.g. in the rubric), we get a format that should be classified as selected rather than constructed response. As the name suggests, none of these item formats requires test-takers to produce their own answers. Instead, answers are selected from the options given (by clicking the best answer, rearranging the options, or by means of drag-and-drop). Back to the item-type menuMultiple choiceEach MC item consists of:
The options include the key (selected) and distractors (incorrect choices).
Using the text converter to create MC items The options should be enclosed in square brackets, separated by a forward slash. The first one is the key. The order of the key and distractors will be randomized when the items are created. If necessary, the target number of options for all items can be preset using the following select list: Number of MC options: If there are fewer options between any pair of brackets than the target number, the "missing" ones will be replaced with "(...)". If there are more options, the extra ones will be ignored. The following text would be converted into 6 three-option MC items (view the output): 
If no gap is needed within the stem, there should be no text following the options (on the same line), for example:
>> the output
True/falseTF items are created in exactly the same way as MC. The only difference is that they have only two options, and these are the same for each item. For example:
Again, the first option is the key. We can use 't' and 'f' for 'true' and 'false' respectively. The options can be in lower or upper case. Here is the output.
Multiple-choice clozeMethod 1 The cloze passage (with numbered gaps) can be given in the rubric. The items proper follow, but they should have no stems (leave them empty), just the options. If necessary, the text converter can be used to quickly destribute the options. Here is an example:
The cloze as a whole can be seen here. Method 2 The above seems to be OK, but here is a better alternative, especially with longer texts: We can follow the rules of creating standard MC items and then select the following options: Cloze/MR/RW Drop-down MC
Note: The first option alone is used to create RW and MR items. If the second option is unchecked and The second option converts MC choices into select lists and puts them into the corresponding stems. The position of the choices in the stem should be indicated by '__' or '___', etc. (two or more underlines). If we use the text converter, the underlines will be added automatically. The first option makes sure that the items are displayed inline (on the same line) so that they look like a passage rather than a numbered list. If a line break is needed, it can be inserted by pressing ENTER where necessary.
For the converter, we need a passage such as the following: 
A postal carrier is working on a new [beat/blow/hit]. He comes to a garden gate marked BEWARE [OF/FROM/AT] THE PARROT! He looks down the garden and, [sure/certain/obvious] enough, there's a parrot sitting on its perch. He has a little chuckle to himself [at/on/for] the sign and the parrot there on its perch. The mailman opens the gate and walks into the garden. He gets as [far/long/closely] as the parrot's perch, when suddenly, it calls out: "REX, ATTACK!"
This version of the cloze task can be seen here.
Right/wrong (RW)These are single-option items (cf. MR below, which are multi-option items). In general, they are created by selecting Cloze/MR/RW and clicking "Add MC" in the editor. !!! For items of this (and of the following) type, the key is either 'a', meaning that the option is correct, or 'n', meaning that the option is incorrect. For the other MC-like types, the key is a letter corresponding to one of the choices (a-f). Because of that, if we convert regular MC items into RW or MR items, we must bear in mind that the original key will be overwritten in those cases where the correct answer is other than 'a'. The content of the options themselves, however, though invisible, is kept intact, just in case. The construction of RW and MR items is probably somewhat less "intuitive" than that of the other types.
1. Passage If it is a passage of text, the options should be between (rather than within) the stems. This does not apply to the last option, for which there is no stem to follow. Accordingly, the last option must also contain the ending of the passage. If we are using the text converter, all we need to do is select the following option: Right/wrong or Multiple response NOTE: Put an asterisk (*) in front of an incorrect option. Here is an example:
Going to bed the [other] night, I noticed people in my shed [*to steal] things. I phoned the police but was told there was no one in the [area] to help. The policeman said they would send someone [over] as soon as possible. I hung [*down.]
A minute later I rang again. 'Hello', I said, 'I called you a minute ago [*for] there were people in my shed. You don't have to hurry now, because I've shot them.' [Within] five minutes there were half a dozen police cars in the area, plus helicopters and an armed [*answer] unit. They caught the burglars red-handed. One of the officers said: 'I thought you said you [*would] shot them.' To which I replied: 'I thought you said there was no one [available.']
When the items are created, the incorrect options will be unchecked in the editor, like this:
Here is the final version (the colour of the options has been changed in the editor).
2. Statements The stems from the True/false task above can be transformed into options in a Right/wrong task. In the text converter, each option as a whole should be enclosed in brackets:
When the items are created, the stems are empty, and each option contains one of the statements:
When test takers click a statement, the question mark changes into a tick, which in turn changes into a cross.
Multiple response (MR)MR items are exactly like RW (Right/wrong) items, but with multiple options. The difference between MR and MC is that in the case of the latter there is one (and only one) correct answer, whereas here no such restrictions apply. What exactly counts as an item in a muti-response task might be somewhat problematic. Is it the stem with all of the options or should each option be regarded as a single item? Whatever the right answer, there is one thing to remember: on this platform, one point is awarded for each correct decision on every single option. If a test consists of 5 stems, each followed by 4 options, the maximum number of points to score is 20. Here is an example for the text converter (remember to check "Right/wrong or Multiple response"):
With the following settings (in the editor):
we get three subsets, each consisting of a stem followed by four options (click to open). Another example:
Here is the output.
MatchingBefore matching items can be created, the following checkbox must be selected (when adding a new set):
Each item consists of a question (or premise) and a response. In the text converter, the pairs should be formatted in the following way:
In the case of matching items, there is no need to type any text between the brackets. To make the task more difficult for the test takers, one or more extra responses can be added in the editor:
Also, the size of the container should be defined:
Here is the output. Here is a matching task with pictures. Note:
Back to the item-type menu Other selected-response typesWith standard fill-gaps items (see below), we expect test takers to produce their own answers. For example, the following sentences are to be completed with prepositions:
However, the words to be used in the gaps (responses) can be given in the instructions (rubric):
This makes the task look more like matching than gap-filling, and definitely selected rather than constructed response. To make it even more "selected-like", the test takers can be instructed to drag and drop the responses into the gaps, instead of typing them in. For this to be possible, each response should be added as the value of a text input field: 1) Click the button indicated in the figure below.
The test itself: open Back to the item-type menu
Constructed response itemsConstructed-response items of every format are created by clicking "Add FG":
They are all, essentially, modifications of the gap-filling type. They may differ from each other in the length of the input, the length of the expected response, the display style (list vs. passage), and also in whether the gap is empty or not (as for Error correction). An additional setting that is available for constructed-response items is the number of spelling errors permitted for partial score (half point). There is one global setting (at the top of the editor):
and one for each individual item:
Back to the item-type menu Fill gapsThe following text would be converted into a set of 5 gap-filling items. There is no limit on the number of correct responses enclosed in brackets (separated by a forward slash):
>> the output Further options in the editor:
(1) Click to delete one of the keyed responses
ClozeThe term "cloze" is used loosely here to refer to items which are displayed as a passage, not as a numbered list. For example, if the items created in the previous section were displayed inline, we would get the following:
The text-into-items converter can be used to quickly process longer texts. There is a button there which makes it possible to replace every instance of a given word (or words) with a gap. In the text given below, all definite and indefinite articles ("the" & "a") have been enclosed in brackets. In addition, several "empty" gaps have been added (where there should be no article, represented by [x]).
Chimp recognises synthetic speech
By Matt Walker Editor, BBC Nature [A] talented chimpanzee called Panzee can recognise distorted and incomplete words spoken by [a] computer, scientists have discovered. That suggests that apes may be more capable of perceiving spoken sounds than previously thought, and that [the] common ancestor of humans and chimps may also have had this ability. It also refutes [the] idea that humans have brains uniquely adapted to process speech, say [the] scientists who have published their findings in [the] journal Current Biology. Panzee was raised from 8 days old, by humans, and was spoken to and treated as if she were human. At [the] same time, she was taught to use symbols called lexigrams to communicate. "This has resulted in Panzee showing [x] proficiency in understanding approximately 130 English words," researcher Lisa Heimbauer told BBC Nature. That made her an ideal subject to test hypotheses about how well other species, rather than humans, might be able to understand speech. "There is [a] view about [the] human ability to produce and perceive speech that is called 'Speech is Special'," said Ms Heimbauer, who is studying for her PhD. "This argument proposes that, besides humans being [the] only species able to produce speech, due to their anatomy, they also have [a] specialised, cognitive module to process speech." [x] evidence for that comes from studies showing that humans can understand speech even when it is incomplete or highly distorted. "However, an alternative view is that auditory processing is fundamentally similar across most mammals, and that animals therefore have latent abilities for [x] speech perception," said Ms Heimbauer. So she and her colleagues Michael Beran and Michael Owren, all from Georgia State University in Atlanta, US, tested Panzee to find out if she too could recognise incomplete or distorted spoken words. They played Panzee noise-vocoded speech, which alters [the] frequencies of [the] spoken words. This produces [a] sound similar to what people with cochlear implants hear. They also played Panzee so-called sine-wave speech, which is synthesised from just three pure tones. Both types of [x] degraded speech have been shown to be understandable by people. "She is only one of [a] few animals who could be tested in this way, to reveal what [the] speech perception abilities of [a] common chimp/human ancestor may have been," said Ms Heimbauer. [The] researchers discovered that Panzee recognised these degraded spoken words far more often than she should by [x] chance. Her upbringing, say [the] scientists, appears to have given her enough experience of hearing and understanding spoken words to allow her to recognise them when they are distorted. That "highlights [the] importance of [x] early experience in shaping speech perception," said Ms Heimbauer. It also provides evidence that [a] common ancestor of humans and chimpanzees would have had [the] ability to perceive speech, she says. "If humans do possess [a] specialised, cognitive, speech-processing brain module, it would be something that evolved later in [x] humans, making us more efficient at what we do."
This is the result of the conversion (the only thing changed in the editor is the style of the title): open
The "Cloze/MR/RW" option can be used to create items such as: Aspiration can be described as a kind of noise, puff of air, or voiceless ___ consisting of strongly expelled breath between the ___ of the plosive and the onset of the following ___. This sentence will have to be divided into 3 items, one for every gap. Here's how to do it using the text converter: 1) Type in or paste the input text into the editable area, and define the gaps:
2) Click "Create FG/MT". 3) When done, make sure the following options are selected:
>> the result Back to the item-type menu Transformations (word given)For example: NOTE! Each second line is preceded by SHIFT+ENTER, rather than ENTER itself.
A look at the editor: Here, in order to create a new line within a given item, either ENTER or SHIFT+ENTER can be pressed.
>> the final result (with some additional formatting applied to the words given) Back to the item-type menu Transformations (complete)The same rules as above apply here. Example:
Back to the item-type menu Error CorrectionFor the text converter, we need pairs of incorrect-correct sentences. The correct sentences are in brackets because they will be saved as the key. For example:
When the items are created, all we need to do is make sure the following option is selected:
Back to the item-type menu Other constructed-response typesThe item formats described here can be modified in many different ways. Here are some suggestions. Back to the item-type menu Phonetic transcriptionTo make it possible for test takers to type in phonetic symbols, the following option should be selected:
IPA symbols will also appear in the editor:
Here is a sample test: open
Extended answerAn extended-answer task is created by clicking the following option:
Online marking of extended responses is possible. Moreover, all answers can be downloaded to MS Word:
|
