A Question on AntConc Collocate Function: Result Count Increases with Higher Min. Freq

[Mandy Wang]

Hi Anthony! I am using the Collocates tool in AntConc and have encountered a phenomenon that I would like to understand better.

When I searched for collocates of the node word "windows" with a window span of 5L to 5R, set the Min. Freq to 1, and sorted by Likelihood, the result returned 11​ collocates.

However, when I changed only the Min. Freq setting to 2​ (keeping all other settings identical) and reran the search, the result surprisingly returned 16​ collocates.

I have checked the KWIC concordance for some of the "new" collocates that appeared in the second search but were absent from the first. They indeed co-occurred with "windows" twice within the defined span, meeting the Min. Freq=2 criterion.

I am curious about why increasing the frequency threshold led to more results being displayed. Thank you very much for your time in helping with this issue.

Best, Mandy

[Laurence Anthony]

Hi Mandy,

Can you send me your corpus (at my email address below), so I can replicate your results?

Laurence.

###############################################################
Laurence ANTHONY, Ph.D.
Professor of Applied Linguistics
Faculty of Science and Engineering
Waseda University
3-4-1 Okubo, Shinjuku-ku, Tokyo 169-8555, Japan
E-mail: antho...@gmail.com
WWW: http://www.laurenceanthony.net/
###############################################################

--
You received this message because you are subscribed to the Google Groups "AntConc-Discussion" group.
To unsubscribe from this group and stop receiving emails from it, send an email to antconc+u...@googlegroups.com.
To view this discussion visit https://groups.google.com/d/msgid/antconc/3a15e14a-09fa-452b-b938-983706d0002dn%40googlegroups.com.

[Laurence Anthony]

Hi Mandy,

Thanks for sending the files. The effect you're seeing is because of the Bonferroni correction to the p-value. More hits will cause a greater impact of the correction. If you use a non-variable p-value threshold (e.g. p<0.05) the results should align with your expectations.

I hope that helps!

Laurence.

###############################################################
Laurence ANTHONY, Ph.D.
Professor of Applied Linguistics
Faculty of Science and Engineering
Waseda University
3-4-1 Okubo, Shinjuku-ku, Tokyo 169-8555, Japan
E-mail: antho...@gmail.com
WWW: http://www.laurenceanthony.net/
###############################################################

[Laurence Anthony]

Just following on from my reply above, it perhaps makes more sense to apply the min-freq threshold to the results *after* the p-value cut-off is applied. Let me think more about this. 

Laurence.