Patent citation - TLS228 and TLS 212

[PEVJ]

I am stuck analyzing patent citation data and I need some help.
I understand that we can use family-family citation (DOCDB family would be my research choice) or patent-patent citation. Family-family citation would be simpler, as the data is in TLS_228_DOCDB_FAMILY_ID. However, I am trying to harmonize these information with TLS212_CITATION, TLS211_PAT_PUBLN and TLS215_CITN_CATEG, because I would like to select some CITN_ORIGIN (TLS212) and PUBLN_KIND (TLS211). [EPO Worldwide Patent Statistical Database CD-room, and I do not have access to the online extension].
To do this harmonization I selected a random sample of 12 patents just to understand the logical model. I do have some questions (I set an example, just to make clearer my doubts):
1) For example, I selected a patent appln_id = 1524 and there are two pat_publn_id (283189889 | 283189896) associated with the appln_id (TLS211). When I search for these two pat_publn_id in the TLS212, it just retrieves one pat_publn_id (283189889). Is there any rule of business to not include pat_publn_id in TLS212 (for example, some publication kind are not included)? Or is just that the pat_publn_id which does not have any citation is excluded from the TLS212?
2)[/b] Using TLS218_DOCDB_FAM, I found the appln_id = 1524´s DOCDB_FAMILY_ID (34957520). From TLS228_DOCDB_FAM_CITN, this family (34957520) has 30 CITED_DOCDB_FAMILY_ID (which I have interpreted that it has 30 family-family citation, which means that it cites 30 different DOCDB families).
I want to find these 30 citations on TLS212, so I first found all patents that is in the 34957520´s DOCDB family. I found 8 simple family members (8 appln_id), being one of those the priority patent (TLS204). The priority patent has 7 citations and the others simple family members together have 51 citations (11 NPL and 40 Patent citation). How can I harmonize these 51 citations with that 30 citations? What is being double count? I am sure I am missing something here....
Thank you very much for your help.
Obs. I use stata software.

[Geert Boedt]

Hello Pevj,
when I read your post, I assume you are using a data base extraction created from PATSTAT Online.
If this is the case, then your observations might be due to way the sample of the extracted data is created. The extraction tool allows you to include cited, citing and also different family members.
But this does not mean that all data from all those applications will be in the extraction. So if you look at a citing document (included through the extraction procedure), the cited documents by the first citing document might not be included anymore in the extraction. If we would not handle some kind of "cut-of", then the data base extraction would simply keep on growing until it ends up being the full PATSTAT database.
So with other words the "integrity" of a data base extraction created with PATSTAT Online will not be as good as working with the full PATSTAT database.
It is therefore important to be sure that all key applications which you need for your research are included in the main query, and are not added through the download extensions.

[PEVJ]

Dear Boedt,
Thank you very much for your answer. In fact, I am using the EPO Patstat (raw data) 2015 spring edition (CD rom), so I am not sure that the differences that I have found are because of the data extraction.
Can I confirm your explanation considering the database that I am using?
Thank you! I am looking forward to hearing from you.
Paula

[Geert Boedt]

Hello Paula,pp
the possible clarification is only valid if you would be using an extraction made with PATSTAT Online.
Kindly upload your list of applications (or family-ID's) for which you do not find citations.
As mentioned in the private note, I would do a short analysis of the distribution over application authority and filing year to have a more general picture of your observation.