Systematic errors in the construction of the scale, Context...

Systematic errors in the scale construction

Score points

An elegant mathematical model, proposed by Torgerson, ideally can give an interval scale. The problem, however, is that the scores themselves, reflecting the relations of order, can be largely subject to distortions. These distortions are caused by systematic errors that experts can commit by assigning scores to one or another stimulus. Let's consider the most common mistakes of such a plan.

Error mitigating the judgment. This error can occur when an expert evaluates well-known incentives. In this case, either an understatement or an overestimation can occur. In the first case, they speak about the error of negative mitigation, in the second case, about a positive softening. Often there is a positive softening. For example, a teacher can evaluate the answer of a student well-known to him is less critical.

The result of a mitigation error is a systematic shift of the scale, either to the positive or to the negative side. As a result, the distribution of experts' answers will assume a clearly expressed asymmetric character. In order for this not to happen, you can use more fractional grades of the scale in the middle so that the subject less often chooses the extreme values. In this case, the distribution of answers will become more of a normal distribution. However, in this case, again, it remains unclear whether the test feature actually has such a distribution or whether it is the result of the scaling procedure that was set by the experimenter.

Centering error. This error to a certain extent is the opposite of the one we have just considered to be a mistake in mitigating judgment. If an expert is unaware of the assessed object, he may be inclined to shift his estimates to the middle of the scale. The result of this is also the deviation of the distribution of expert estimates from the normal and the occurrence of a kurtosis distribution. As a guideline for avoiding such errors, it may be possible to ask the subjects to avoid, as far as possible, moderate judgments or completely remove this possibility from the categories assigned to the subjects, as was the case in the example of the weight sense scale that we cited, where the middle point is completely absent. Another recommendation can be the location of scale gradations in the middle with large gaps than at the edges.

Halo effect. There is an error that is determined by the influence of the whole assessed object as a whole or the whole personality of an individual on the evaluation of its individual characteristics or features. This error of the general impression is usually referred to as the halo effect & quot ;, or halo effect & quot ;. So, if the estimated person as a whole gives us positive impressions, then we will tend to evaluate him on a number of positive characteristics higher, and his negative characteristics, on the contrary, will be ignored. The general negative attitude towards the person will force the expert to give him lower scores on these scales, strengthening the ratings on negative characteristics.

Since the halo effect reflects the general positive or negative attitude of the expert in relation to the person being evaluated or to some other object, it is not possible to completely eliminate this error. Nevertheless, there are a number of standard recommendations that allow, if not completely eliminate the probability of such an error, then at least minimize it. It is considered that one should avoid assessing the hard to explain features, exotic, unconventional signs, insufficiently clearly defined characteristics, signs involving connections with other people, and characteristics of high moral value. It is also necessary to ensure that the subject evaluates only one characteristic, not their combination.

A logical error. If any characteristics are presented to the subject logically connected, he will be inclined to give close or even identical estimates to the measured objects by these characteristics. Unlike the halo effect, this error does not depend on the ratio of the expert to the measured object or person, but only reflects his tendency to perceive the evaluated characteristics as logically or semantically coincident. For example, a subject may feel that such characteristics as "beautiful" and attractive & quot ;, or modern and "liberally minded", are synonymous with each other. In this case, his estimates for these characteristics will largely coincide. An exit from this situation can be a recommendation to examine the subject only on the directly observed links, rather than on speculative logical or semantic coincidences.

Contrast error. A number of expert's personal attitudes may make him underestimate or, conversely, overestimate the presence of certain characteristics in the subjects. For example, knowing the content of the taught subject at a very high level, the teacher will underestimate the level of preparation of students on this subject, considering it insignificant.

Since this error is related to the expert's personal settings, it is extremely difficult to control.

Context Impact and Adaptation Level

F. Helson suggested that all subjective assessments within the psychological continuum should be viewed not as absolute, but as relative. The basis for it was the following formulation of the Weber-Fechner law:

Normally, the value of the lower absolute threshold is chosen as the value of S 0 , thus taking this value as the zero value of the scale. Helson suggested using the level of adaptation of an individual ( AL ) as such a value. This level represents the amount of sensation that the subject is inclined to regard as a feeling of average magnitude. The very value that determines the level of adaptation depends on the past experience of the individual, as well as on the incentives that the individual evaluates at the moment and which constitute the context of the experience, influencing the subject. The stimulus corresponding to the level of adaptation, the subject will assign the average scores of the scale. Incentives that exceed the level of adaptation will be evaluated higher, and those incentives that are below the adaptation level will receive lower estimates.

In the simplest case, AL can be estimated as the average logarithmic value of the estimated incentives. For example, when estimating the mass from 200 to 400 g, AL will be about 250 grams. Incentives less than 250 g will be assessed by the subject as light, and stimuli of greater weight - How heavy. If the series includes objects weighing from 400 to 600 g, the adaptation level is already 475 g. If a stimulus of 900 g is added to the first or second range of estimated stimuli, then the adaptation level in the first case will increase to 350 g, and in the second - up to 550 g.

The level of adaptation affects not only psycho-physical assessments. It can also be manifested in situations where incentives are used, which have a pronounced physical nature. Thus, in one study, subjects were asked to evaluate different ages, placing them on a scale from the "very young" to very old & quot ;. It turned out that the 10-year-old subjects choose the age of 36 as the average age, the twenty-year-olds 42 years, the seventy-year-olds 52 years. Interestingly, these results almost completely coincided with theoretical calculations based on the theory of the level of adaptation of Helson.

Thus, Helson argues that scores can be considered as an explicit scale of equal intervals, rather than a hidden interval scale, as implied in the law of categorical judgments.

thematic pictures

Also We Can Offer!

Other services that we offer

If you don’t see the necessary subject, paper type, or topic in our list of available services and examples, don’t worry! We have a number of other academic disciplines to suit the needs of anyone who visits this website looking for help.

How to ...

We made your life easier with putting together a big number of articles and guidelines on how to plan and write different types of assignments (Essay, Research Paper, Dissertation etc)