Number of job adverts listing ‘good communication skills’ as essential
The main question I wanted to answer was: How many job adverts demand things like ‘good communication skills’? More specifically, how many job adverts with a person specification list criteria similar to ‘good communication skills’ as essential criteria in the person specification? After exploring the data, I developed 4 search terms that would find criteria containing things like ‘good communication skills’:
-
_____good_____communication_____skills_____
-
_____excellent_____communication_____skills_____
-
_____good_____interpersonal_____skills_____
-
_____excellent_____interpersonal_____skills_____
Each search looked for that sequence of words. The gaps (_____) could contain spaces, words, and/or punctuation. The gaps could also be completely empty (no spaces, no words, and no punctuation).
The search was applied to each essential criterion. For criteria containing a nested list, the search was applied separately to the introductory text of the list, and to each nested list item. For criteria containing conjunctions that were highlighted/set apart, the search was applied separately to each string of text that was separated by one or two of those conjunctions. I applied the search separately to each of these sub-parts of the criteria to reduce false positives (e.g. in a nested list the word “good” could appear in one nested list item and the words “communication” and “skills” could appear inF other nested list items. These words are all part of one criterion, but they refer to separate skills/qualities).
Using the above searches, I found that 927 of the 1,662 jobs listed one or more of the following criteria as essential:
- excellent communication skills;
- good communication skills;
- excellent interpersonal skills;
- good interpersonal skills.
The number of results for each search term was:
- 553 for excellent communication skills;
- 279 for good communication skills;
- 273 for excellent interpersonal skills;
- 166 for good interpersonal skills.
819 jobs returned results for searches for excellent or good communication skills. 434 jobs returned results for searches for excellent or good interpersonal skills.
Collocations
I wanted to know if the words in my search terms collocated (i.e. co-occur more often than would be expected by chance). So, I analysed the Mutual Information (MI) and t-score for each of the following words:
- Excellent;
- Good;
- Communication;
- Interpersonal;
- Skills.
I set a strict minimum frequency of 10 (ƒ ≥ 10). I treated all data as lowercase. The search spanned from 4L to 4R (to allow for other words between collocates). All of the words were associated strongly and significantly.
Table 1. Collocates of ‘excellent’
| Collocate |
Frequency |
MI |
Rank by MI |
T-score |
Rank by t-score |
| interpersonal |
330 |
6.42 |
2 |
17.95 |
4 |
| communication |
650 |
6.34 |
4 |
25.18 |
3 |
| skills |
1247 |
5.66 |
10 |
34.61 |
1 |
| good |
68 |
2.77 |
98 |
7.04 |
25 |
Search term = excellent. MI and t-score correct to 2 d.p.
Table 2. Collocates of ‘good’
| Collocate |
Frequency |
MI |
Rank by MI |
T-score |
Rank by t-score |
| interpersonal |
199 |
5.66 |
6 |
13.83 |
6 |
| communication |
354 |
5.44 |
11 |
18.38 |
4 |
| skills |
899 |
5.16 |
19 |
29.14 |
1 |
| excellent |
68 |
2.77 |
96 |
7.04 |
32 |
Search term = good. MI and t-score correct to 2 d.p.
Table 3. Collocates of ‘communication’
| Collocate |
Frequency |
MI |
Rank by MI |
T-score |
Rank by t-score |
| interpersonal |
282 |
6.48 |
6 |
16.61 |
6 |
| excellent |
650 |
6.34 |
8 |
25.18 |
3 |
| skills |
1351 |
6.06 |
9 |
36.21 |
1 |
| good |
354 |
5.44 |
16 |
18.38 |
5 |
Search term = communication. MI and t-score correct to 2 d.p.
Table 4. Collocates of ‘interpersonal’
| Collocate |
Frequency |
MI |
Rank by MI |
T-score |
Rank by t-score |
| communication |
282 |
6.48 |
3 |
16.61 |
4 |
| excellent |
330 |
6.42 |
4 |
17.95 |
3 |
| skills |
695 |
6.16 |
5 |
25.99 |
1 |
| good |
199 |
5.66 |
6 |
13.83 |
5 |
Search term = interpersonal. MI and t-score correct to 2 d.p.
Table 5. Collocates of ‘skills’
| Collocate |
Frequency |
MI |
Rank by MI |
T-score |
Rank by t-score |
| interpersonal |
695 |
6.16 |
3 |
25.99 |
8 |
| communication |
1351 |
6.06 |
6 |
36.21 |
2 |
| excellent |
1247 |
5.66 |
14 |
34.61 |
4 |
| good |
899 |
5.16 |
25 |
29.14 |
6 |
Search term = skills. MI and t-score correct to 2 d.p.
We can see from these results that there is a strong and significant association between the words ‘excellent’, ‘good’, ‘communication’, ‘interpersonal’, and ‘skills’. This suggests that strings of text like ‘excellent communication skills’ are stock phrases in job adverts.
These results also show that ‘good’ collocates less strongly than the other words. This is particularly the case with ‘excellent’ (see table 1 where rank by MI is 98). Nevertheless, ‘good’ does collocate well with the words ‘communication’, ‘interpersonal’, and ‘skills’, and we would not expect it to be strongly associated with the word ‘excellent’ since they perform the same function in person specifications.
Keywords
I wanted to know if the words in my search terms were keywords (i.e. they occur unusually frequently in the essential criteria in comparison to other texts). I decided to use 2 reference texts. Firstly, I decided to compare the essential criteria with the desirable criteria. This would give an idea of how recruiters use these words within the person specification. Secondly, I decided to compare the essential criteria with the Leeds Internet Corpus frequency list of word forms. This would answer the question of whether these words are key to the person specification. The words I was interested in were:
- Excellent;
- Good;
- Communication;
- Interpersonal;
- Skills.
I analysed the log-likelihood for each of the words above. I treated all data as lowercase. All of the words were keywords when compared to the desirable criteria and the Leeds Internet Corpus frequency list of word forms.
Table 6. Keywords, using ‘desirable criteria’ as a reference corpus
| Keyword |
Frequency |
Keyness (log-likelihood) |
Rank by keyness |
| excellent |
1485 |
516.09 † |
2 |
| skills |
3753 |
482.27 † |
3 |
| communication |
1214 |
337.56 † |
6 |
| interpersonal |
585 |
191.42 † |
10 |
| good |
1513 |
185.74 † |
11 |
Reference corpus = desirable criteria. Keyness correct to 2 d.p. † = significant (p < .0001)
Table 7. Keywords, using the ‘Leeds Internet Corpus frequency list of word forms’ as a reference corpus
| Keyword |
Frequency |
Keyness (log-likelihood) |
Rank by keyness |
| skills |
3753 |
37553.80 † |
3 |
| good |
1513 |
15139.6 † |
6 |
| excellent |
1485 |
14859.42 † |
7 |
| communication |
1214 |
12147.7 † |
11 |
| interpersonal |
585 |
5853.71 † |
31 |
Reference corpus = Leeds Internet Corpus frequency list of word forms. Keyness correct to 2 d.p. † = significant (p < .0001)
These results show that the search words (‘excellent’, ‘good’, ‘communication’, ‘interpersonal’, and ‘skills’) are all keywords in the essential criteria. There is a significant difference between how frequently these words occur in the essential criteria and how frequently they occur in the reference texts. All of these words were top ranking keywords in the essential criteria. That is, these words occur far more frequently in the essential criteria than they do in the desirable criteria and far more frequently than they do outside of a person specification.
We can conclude that recruiters used the words ‘excellent’, ‘good’, ‘communication’, ‘interpersonal’, and ‘skills’ significantly more often in the essential criteria than in the desirable criteria. This suggests that recruiters have a strong preference for stating that things like ‘excellent communication skills’ are essential skills for applicants rather than desirable skills.
We can also conclude that recruiters used the words ‘excellent’, ‘good’, ‘communication’, ‘interpersonal’, and ‘skills’ significantly more often than they are used in a non-recruitment context. This suggests that these words are key to the act of writing a person specification. It suggests that recruiters have a strong preference for using terms like ‘excellent communication skills’ in their job adverts.
Conclusion
The results combined show that the words ‘excellent’, ‘good’, ‘communication’, ‘interpersonal’, and ‘skills’ are individually important to the essential criteria of the person specifications; they have a strong relationship with one another; and they appear in combination in a single criteria in the person specifications of most job adverts. This means that we can expect to find them in a person specification in a phrase similar to ‘excellent communication skills’. These phrases are used, in effect, as stock phrases.
