Running Head Word Association Spaces 1 Word Association Spaces Semantic Spaces based on Free Association that Predict Memory Performance Send correspondence to
responses. By applying scaling methods on the matrix, words are placed in a high dimensional
space such that words with similar associative patterns are placed in similar regions of the space.
--------------------------------------------------------------------
Insert Figure 1 and Table 1 about here
--------------------------------------------------------------------
In total, more than 6000 people participated in the collection of the free association
database of Nelson et al. (1999). An average of 149 (SD = 15) participants were each presented
with 100-120 English words. These words served as cues (e.g. “cat”) for which participants had
to write down the first word that came to mind (e.g. “dog”). For each cue the proportion of
subjects that elicited the response to the cue was calculated (e.g. 60% responded with “dog”,
15% with “pet”, 10% with “tiger”, etc). We will first explain how SVD is applied to the norms and then continue to the metric
MDS method. The method of SVD can be applied to any matrix containing some measure of
strength or co-occurrence between two words. Although many different ways have been
proposed to calculate an index of associative strength between two words (e.g., Marshall &
Cofer, 1963; Nelson & Zhang, 2000), we will restrict ourselves to two simple measures of
associative strength. Let Aij represent the proportion of subjects that gave the response j to the
cue i. The simplest measure would be to take Aij itself. In the norms, the associative strengths Aij
are often highly asymmetric where the associative strength in one direction is strong while it is
weak or zero in the other direction. Even though SVD can be easily applied to asymmetric
matrices, the results are more interpretable when it is applied to symmetric matrices3. In a later
section in the paper, we will show that symmetrizing the norms does not necessarily mean that
热门文档