﻿ Scores

# Scores

Each of the *col lines in the CCs worksheet is said to define a “subtest”. Subtests may be cognitive or affective. Cognitive tests measure knowledge or achievement, while their affective counterparts attempt to assess such things as attitudes, opinions, and feelings.

There are three *col cards in the Lertap Quiz’s CCs worksheet. The first one points to a total of 25 columns in the Data worksheet, the second points to 10. The third points to the same items as the second; in this case, the third subtest uses the same items as the second subtest, but the items are scored differently.

There’s a *key line shortly after the first *col card, and this tells Lertap that the first subtest, with 25 items, is to be scored as a cognitive test.

The common procedure for scoring responses to cognitive questions is to award one point for each correct answer, and this is what Lertap does. It’s possible to award more points, and it’s possible to have more than one right answer to any cognitive item. These things are accomplished by putting *wts and *mws lines in the CCs worksheet.

(Note: the discussion of the MathsQuiz sample has an example which uses *mws lines. *wts lines, also known as *wgs lines, are mentioned here.)

The first *sub line gives a title of “Knwldge” to the first subtest. What will be the possible range of scores for Knwldge?  From zero to 25. As we’ve just said, this is a “common”, or standard, cognitive run, there are no *wgs or *mws lines in the CCs worksheet.  Consequently respondents get zero points for each incorrect answer, and one point for each right response. There are 25 items, so the maximum possible score for Knwldge is 25.

The second subtest, “Comfort”, is affective in nature, something Lertap detects by the presence of the “Aff” control “word” on the *sub card. There are 10 items in this subtest. Lertap will score each, adding results from each item to derive a total Comfort score.

How? Good question.

Lertap's default assumption is that each affective item will use 5 possible responses. In fact, it assumes that Res=(1,2,3,4,5).  What’s this Res=() thing mean? Two things. The number of characters found within the parentheses tells Lertap how many possible responses there may be to any item, while the characters themselves, five digits in this case, are the responses which will be recognised and scored.

How? If someone has an answer of 1 to the first affective item, how many points are awarded? One. Two points for the second response, which is a 2 in this case. Three for the third, four for the fourth, five for the fifth.

This is a common way of scoring affective items. Lertap allows for many other possibilities, scoring-wise. There can be more than five recognised responses—there may be as many as 26. The responses do not have to be digits. The points given to the responses do not have to correspond to the ordinal position of the response—special lines in the CCs worksheet, such as *mws lines, allow other scoring schemes to be effected.

Okay then, the subtest we’re talking about, Comfort, has 10 items. The minimum possible score on any single item is 1 (one), while the maximum is 5. For the 10 items as a whole, then, the minimum possible score is 10, the maximum 50.

We’re just about ready to look at the scores themselves, but first one final matter. The Comfort subtest has a *pol line associated with it—“pol” means “polarity”. The *pol line has ten plus and minus signs, one for each of the subtest’s items. The first sign in the *pol card is +, which means that a response of 1 (one) on the first item will get one point, while a response of 5 will get five points. This is referred to as “forward” scoring.

Items whose respective entry in the *pol card is minus will be reverse scored. On these items, a response of 1 (one) will get five points, while a response of 5 will get one point.

Why get into this sort of caper?  Because it is not at all unusual for affective tests, or surveys, to consist of a mixture of positive and negative items.  People might be asked to state whether they agree or disagree to a series of statements. The first one might be, for example, “Lertap is great!”, while the second might be “I would not advise anyone to use Lertap.” People who are happy with Lertap, which of course is just about everyone, would be expected to agree with the first statement, and disagree with the second. The use of *pol lines makes it possible to accommodate forward and reverse scoring with some ease.

(To get even more *pol-arised, you could take in another affective example.)

What about missing data, such as the situation which arises when a student does not answer an item? Read all about Lertap's answer to this question by transplanting yourself to here; to quickly whisk away to MDO scoring, as used here for the third subtest, get whisk it.

Okay. Let’s have a squiz at these scores, Knwldge, Comfort, and MDOCmfrt.

Where are they?

They’re in the Scores worksheet. Find its tab at the base of the Excel window. Click on the tab. Look:

What do you make of it, this Scores sheet?  It uses 80 rows, and 4 columns. Lertap presupposes that the first thing you want to rest your peepers on, as far as the Scores sheet goes, is the summary statistics section at the bottom of the sheet. This is why rows 3 through 55 have scrolled off the display.  If you scroll to the top of the sheet, you’ll be able to satisfy yourself that there are 60 sets of scores for each respondent.

Are the summary statistics self-explanatory? Good. You can find out how they’re calculated in Chapter 10 of the manual, "Computational Methods".

A couple of quick comments before moving on. The MinPos and MaxPos scores are the minimum and maximum possible scores on each subtest, while Min and Max are the lowest and highest scores actually earned by the 60 respondents. The correlation coefficients are Pearson product-moment coefficients (the most common kind).

Is it possible to get percentage scores? You bet. Use the “Per” control word on the *sub card. It’s also possible to get a “scale” score if the subtest is affective. Such scores divide the original subtest score by the number of items in the subtest, a procedure which is common to quite a number of internationally-known affective instruments. Scale scores are requested by using the “Scale” control word on a *sub line.

Is it possible to get a total score, one which sums up the scores earned on the individual subtests? Most definitely. In fact, Lertap has to be told not to do this, something which is done by using Wt=0 statements on *sub cards. In the present example, each *sub line has this sort of statement. We didn’t want a total score—we didn’t think it made much sense to add together results from very different subtests, with cognitive items in one, and affective items in the others.

Tidbits:

Don't forget this all-important reference on CCs worksheets and their control lines: the on-line help system.

It is possible to score items which are open-ended, essay, "short answer", "free response", or "constructed response". See this paper, please, but note: the *alt line used  towards the end of the paper is the old style; the *alt line's format was changed after this paper was written (a factor which should not limit your enjoyment of the paper).