Anagram solution as a function of pronounceability and difficulty

Abstract

The present study investigates the effect of pronounceability (P), as rated by Ss, on anagram solution probability under two levels of difficulty (D) Mayzner & Tresselt (1965) presented evidence that there is no difference in solution times of word and non-word anagrams. This discrepancy of results suggests that perhaps an alternate approach is needed to elucidate the problem of word vs. non-word anagrams.It is reasonable to consider words more pronounceable than non-words. It is equally reasonable to consider some non-words more pronounceable than other nonwords. In light of the Beilin & Horn (1962) and Ekstrand & Dominowski (1965) findings, it is suggested that pronounceability (P) may interfere with word anagram solution. If this is the case, it follows that P should have a similar, though perhaps reduced, effect on the solution of non-word anagrams. For example, if one takes two anagrams of equal letter order difficulty and Thorndike-Lorge frequency, e.g., crove (solutioncover) and mcusi (solution-music), Ssshouldrate crove higher in P than mc u s i. Differences in solution probabilities may then be attributable to P. This study investigates the effect of P, as rated by Ss, upon the probability of solution of non-word anagrams under two levels of difficulty (D), or number of letters moved. Method Materials. Thirty Thorndike-Lorge AA nouns were randomly chosen as potential solution words from a list of five-letter nouns beginning with consonants. Sixty anagrams (30 one-letter move and 30 two-letter move problems) were then constructed by E so as to lend themselves easily to division into high pronounceability (HP) and low pronounceability (LP) categories.Psychon. Sci., 1966. Vol. '" (10) JOHN A. HEBERT AND CECIL A. ROGERS, JR.UNIVERSITY OF ARIZONA Two hundred and ten Ss then rated the 60 anagrams on a 1 to 5 scale for P (low to high). Each S rated a list of 10 anagrams presented in a random sequence or its inverse. All lists contained expected HP and LP problems. Each anagram was rated by 35 Ss. This procedure yielded four types of anagrams: (1) HP oneletter move or "easy" anagrams (HP-E), (2) HP twoletter move or "hard" anagrams (HP-H), (3) LP oneletter move anagrams (LP-E), and (4) LP two-letter move anagrams (LP-H). The 10 items having the smallest variance in each of the four categories of anagrams were then selected for one of four lists, since no overlap in ratings of the expectedHP and LP items occurred.The HP-E list, (e.g., patin, crove, and larbo with solutions-paint, cover, and labor) had a mean Prating of 4.66, and a range of 4.91 to 4.34, while in the HP-H list (e.g., rocut, sumic, anddoblowithsolutions-court, music, and blood) the mean=4.59, and range = 4.94 to 4.29. In the LP-E list (e.g., sgaru, mcusi, and trnai with solutions-sugar, musiC, and train) themean=2.29,and range = 2.63 to 1.86, while iIi the LP-H list (e.g., drkni, ifgth, and mnoye withsolutions-drink,fight, and money) mean = 1.98, range = 2.57 to 1.45. Some of the solutions, e.g., sugar, were used in more than one list. Note

The Effects of Superficial and Structural Information on Online Problem Solving for Good versus Poor Anagram Solvers

“…Replicating earlier research (Dominowski, 1969;Hebert & Rogers, 1966), there was a reliable effect of pronounceability, as participants spent more time on easier to pronounce anagrams (Ms of 17.05 s and 15.40 s, respectively; SDs of 5.32 and 4.76). As predicted, however, this effect was much larger for poor solvers.…”

“…The degree to which an anagram letter string is pronounceable is not predictive of the solution because, for example, (a) many pronounceable letter combinations are not words (e.g., zelba, eckho), and (b) a contiguous subset of letters that is pronounceable in an anagram (e.g., aft in aftes) may not make up part of the solution in combination with the remaining letters (the solution is feast). Previous research indicates that anagrams that are easier to pronounce (e.g., aftin as opposed to infta) are harder to solve (the solution is faint; see Dominowski, 1969, andHebert &Rogers, 1966). These studies used unselected college students, who our earlier research suggests are typically poor to fair at solving anagrams (see the Experiment 1 Participants section).…”

The effect of pronunciation practice on anagram difficulty

“…If it could be shown that mere presentation or spelling practice produces minimal changes in pronunciation speed, a comparison of these types of training with pronunciation practice would be informative. FinaHy, as Hebert & Rogers (1966) suggested,. a consideration of pronounceability.effects provides a·plausible explanation for the finding that word anagrams are more difficult than nonsense anagrams(e.g., Ekstrand & Dorninowski,1%8).…”

“…With stimuli having weaker populars, however, the idiosyncratic responders tended not to repeat their responses, producing a low degree ofTRR. Hebert & Rogers (1966) reported that anagrams which are easy to pronounce (EP) are more difficult to solve than anagrams that are hard to pronounce (HP). However, the design of their study did not permit an unambiguous conclusion concerning the effect of anagram pronounceability since EP and HP anagrams had different solution words.…”

The Syllable Effect in Anagram Solution: Unrecognised Evidence from Past Studies

“…Novick and Sherman (2008) describe it as "superficial" and something which is more likely to affect poor solvers. Yet the detrimental effect of pronounceability on anagram solution has been demonstrated on many occasions (Herbert & Rogers, 1966;Dominowski, 1966;Gilhooly & Johnson, 1978). It clearly suggests that Anagrams and Syllables phonemic encoding of anagrams takes place and is involved in some way in their solution, just as phonemic encoding plays a part in visual search, word recognition and reading processes (Conrad, 1964).…”