Comparison between Pairs of Key Words

N.B.: The pictures shown in this section have been obtained by using a previous version of T-LAB. These pictures look slightly different in T-LAB 10. Also: a) a new radial diagram is now available which allows to quickly appreciate differences in word associations; b) right clicking on the keyword tables makes additional options available.
Some of these new features are highlighted in the below image; c) a quick access gallery of pictures which works as an additional menu allows one to switch between various outputs with a single click.

This T-LAB tool allows us to compare sets of elementary contexts (i.e. co-occurrence contexts) in which the elements of a pair of key-words are present.

The table on the left shows the list of selected lemmas and their corresponding occurrence values within the whole corpus or a subset of it.
The user is invited to select - one after the other - two of these (a pair) with a single click (see below).

The bar chart (see below) allows us to appreciate the number of elementary contexts in which each lemma co-occurs with the "A" term (red colour), with the "B" term (blue colour) and with both "A" and "B" (green colour).
By double-clicking each label of the chart it is also possible to check its corresponding co-occurrence values and obtain a pie chart with the co-occurrences values concerning each selected word (see below).

The comparisons provided by T-LAB involve the co-occurrences of the elements in the "pair" and each of the words contained in the table (see below).

Let:

A = set of elementary contexts (TOT. E.C. = 86) in which the first word of the pair (e.g. "Terrorist") is present;
B = set of elementary contexts (TOT. E.C. = 16) in which the second word of the pair (e.g. "Muslim") is present.

The first type of comparison concerns the shared associations (see the "intersection" button) and takes into account all words which are present both in "A" and in "B".

In the output table (see above) every row shows the values corresponding to the comparisons of each lemma.

The reading keys are as follows:

• ASS (A) = number of elementary contexts in which each lemma is associated (i.e. co-occurrence) with (A);
• ASS (B) = number of elementary contexts in which each lemma is associated with (B);
• ASS (AB) = number of elementary contexts in which each lemma is associated with (A) and (B);
• CHI2 = Chi-square values;
• (p) = probability associated with the chi square value (def=1).
  

In this case, for each key word (e.g. "country"), T-LAB builds a table like the one below and applies the Chi-square test to it:

In this table:
- the (A) row shows the number of elementary contexts in which "country" is present (10) or absent (76) in the set of contexts (86) containing the first word of the pair ("Terrorist");
- the (B) row shows the number of elementary contexts in which "country" is present (7) or absent (9) in the set of contexts (16) containing the second word of the pair ("Muslim").

N.B.: In this case the Chi-square value is 10.02.

Moreover a double click on each item of the output table allows us to save a HTML file with the number of elementary contexts in the corresponding column.

The second type of comparison involves the differences between A and B (A-B and B-A).

In this case T-LAB offers two tables showing the key words which are exclusively associated with the first (A) "or" second (B) word in the pair.
In both tables the "TOT" column shows the number of elementary contexts in which each lemma is associated with only one of the two terms of the pair.

Eventually, by clicking the appropriate buttons (see the picture below) it is possible to check and export all similarity indexes concerning any word pair.