Informational models and their uses

Binder, Arnold; Wolin, Burton R.

doi:10.1007/bf02289566

Cited by 7 publications

(3 citation statements)

References 16 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Although something like this analysis is described in Attneave (1959), the representation of the transition matrices and uncertainties used here follows the Markov model as presented by Binder and Wolin (1964), in which H=-:SPi P i jlog2 Pij' where Pi and Pij are the absolute probability of being in state i and the transition probability of going to state j after being in state i, respectively.…”

Section: Modell: H (K-span)mentioning

confidence: 99%

Information, run structure and binary pattern complexity

Vitz¹

1968

Perception & Psychophysics

View full text Add to dashboard Cite

show abstract

Section: Modell: H (K-span)mentioning

confidence: 99%

Information, run structure and binary pattern complexity

Vitz¹

1968

Perception & Psychophysics

View full text Add to dashboard Cite

show abstract

“…The theory of Markov processes applies to a sequence of elements extended without limit. Pattern generation makes use of small segments of a process, and Binder and Wolin (1964) have pointed out that segments may not be very representative of the process as a whole. In addition, if a collection of generated patterns is to be statistically independent, the segments must not be consecutive.…”

Section: Obtaining Special Propertiesmentioning

confidence: 99%

Vargus 7: Computed patterns from markov processes

Evans

1967

Syst. Res.

View full text Add to dashboard Cite

show abstract

“…However, the earlier equations and those of the present paper lead to identical numerical results. For an interesting discussion of the different assumptions underlying the two sets of equations, see Binder and Wolin (1964). Table 1 shows the constraints existing between pairs of letters that are contiguous (C2) and pairs that are separated by nine intervening characters (Cn) for first, third, and fifth readers of the two series.…”

mentioning

confidence: 99%

Statistical Comparison of Two Series of Graded Readers

Carterette

Jones

1965

American Educational Research Journal

View full text Add to dashboard Cite

In the search for a useful method of quantifying the difficulty level of text intended for children, a redundancy measure derived from information theory was considered. Redundancy of letters in written English is a function of the relative frequency of each letter and of the constraints that operate in the language to produce combinations of letters with vastly different frequencies, including, of course, many of zero frequency (Shannon, 1951;Garner, 1962). When this measure of redundancy was applied to measurement of the sequential constraints in adult texts between pairs of letters, a very regular mathematical function describing per cent of redundancy resulted (Newman and Gerstman, 1952). Furthermore, these curves appeared to discriminate among texts of varying difficulty (Newman and Waugh, 1960). Application of these measures to children's graded readers yielded similar smooth curves of growth of sequential constraint, and good discrimination of difficulty among readers at Levels 1, 2, 3, and 5 (Carterette and Jones, 1963). The redundancy found in children's freereading choices was likewise related to difficulty of text but not to mere differences in style (Jones and Carterette, 1963).In view of the stability of these measures with a sample size of six thousand letters (Jones and Carterette, 1963), it is of some interest to apply them to a comparison of two series of graded readers. The important question for the educator is whether it is possible to design readers for a given level of difficulty in such a way that a level in one series is equivalent to the same level in another series and, especially, so that a regular progression of difficulty is achieved. Results are presented here for a new series of readers, and these are compared with data previously published for another series of readers and for free-reading choices. METHODThe texts of three levels (1, 3, and 5) of two widely used series of basic readers, Sheldon Basic Reading Series and Ginn Basic Readers, (Sheldon and others, 1957a, b, c; Russell and others, 1957; Russell and others, 1956) were key-punched on IBM cards, using a 28-character alphabet: the 26 letters of the English alphabet plus word mark and sentence mark. Redundancy statistics were computed by 13 at Yale University Library on July 6, 2015 http://aerj.aera.net Downloaded from

show abstract

Informational models and their uses

Cited by 7 publications

References 16 publications

Information, run structure and binary pattern complexity

Information, run structure and binary pattern complexity

Vargus 7: Computed patterns from markov processes

Statistical Comparison of Two Series of Graded Readers

Contact Info

Product

Resources

About