The generalization of the law factorization of semantic space in relation to all assessment scales
Abstract
The generalization of the law factorization of semantic space in relation to all assessment scales
Incoming article date: 26.08.2016Some experimental regularities obtained on the basis of monitoring and modeling the emotional perception of social processes, which are carried out from 2009 to the present moment by method of semantic differential are given. Total for this period were surveyed several thousand respondents, mainly students of Rostov-on-Don, Grozny, Vladikavkaz and Karachaevsk. And also the politically active part of the population of Rostov-on-Don. The purpose of the survey was to identify the emotional perception of the local and central political orders, own level of relative deprivation, cultural infrastructure and cultural life of the city as a whole, own emotional state. Received partial confirmation law factorization primary space of Ch. Osgood. For all these objects and any groups primary semantic signs reduced to three-four factors. However, we have shown that this regularity is observed only for the classical factorization method, in which the factors are identified based on the minimization of a specific random residue. For other methods of factoring the number of identified indicators of latent increases. Should be noted that this result was not previously observed in the literature. The second part of the Osgood’s law the nature of latent indicators as the factors of assessment, force and activity doesn't possess such commonality. For our subjects splitting primary signs into bunches doesn't occur, all signs are mixed in each factor. Results of the factor analysis, received by a cognitive part of the questionnaire has shown, as for these parameters the same regularity is carried out, as for a semantic part. All set of primary signs reduced to four-five latent factors. Cognitive part of the questionnaire has no relation to a technique of semantic differential. Common for both parts of the questionnaire is rating scales are used. And the result suggests that the observed regularity factorization by Osgood is not connected with the specific semantic space and emotional perception, and is a special case of a general factorization of any rating scale. For verification of this assumption was conducted the research results of poll SFU’s students to identify satisfaction with the educational process carried out in 2015. Questioning had purely cognitive character, 20 primary cognitive signs characterizing educational process from various parties have been used. In total, about 4,000 students were surveyed. Factorization of the classical method showed that the three indicators describe the 95% of the variation of the primary signs, and four - 97%. It convincingly shows the general nature of factorization of both semantic, and cognitive rating scales. We have suggested that the nature of such regularities lies in the psycho-physiological peculiarities of the formation of the respondents estimates. On average the person estimates any object no more than on three – four factors. In attempt to increase the number of estimated parameters is beginning to follow the correlation between them. This result makes for a fresh look at the development of questionnaires and on the adequacy of many of previous polls. It turns out that the use of large scale (the number of parameters to be estimated much more than four) does not make sense, they still rolled up to four factors. In this case the adequacy of the results will be decrease. Identify a limited perception of the estimated parameters is an additional source of errors in the measurement results. Fully it belongs to all earlier received results on a large rating scales.
Keywords: semantic differential, Osgood, semantic space, factor analysis, the classical method, latent factors, rating scales, experimental regularity factorization scales, psycho-physiological nature of the law factorization