Further to Quest for the breeding K Parrot

To: <>
Subject: Further to Quest for the breeding K Parrot
From: "Philip Veerman" <>
Date: Sun, 6 Feb 2011 16:20:11 +1100
My apologies for another message about this but it is important (in case anyone has a serious interest in understanding these data) to explain another important reason why Martin's graph did not match mine.
My comments were based on breeding results as tallied by breeding records in the GBS. Martin's graph and comments were based on breeding results as tallied by breeding observations in the GBS. This was clear from Martin's comments and the scale on the graph leaves me with no doubt that is what it shows but even I missed the point and no one has commented on this.
These are potentially very different things, as one breeding record can have from 1 to 52 breeding observations. Although for this species it is typically 1 or 5 to 10. This is all explained in The GBS Report. Basically a record is at least one observation of an event on one chart. An observation is one entry on one week. For example the Ravens breeding at my house this year have 19 breeding observations this year. But from watching them I know there is only one breeding event (one pair raising 3 chicks) and thus one record. In terms of the incidence of local breeding, which surely is the relevant question about the King Parrots, I believe that number of records is far more accurate than the number of observations. It is almost certain that for this species many observations are still just one breeding event, as in repeated noting of the same family over several weeks. Not only that but it is important to note that the number of observations introduces a whole extra huge aspect of survey randomness because it depends on observer activity and how many weeks the observer is present (in my raven case the 2 weeks missing of the year are when I was away).
MS Access is great for counting numbers of records fitting various criteria in the database. So the MS Access report I mentioned: "Breeding records and ratios by species and year" is really easy to output (once the query structure had been built). Martin's output, that counted observations, as in fields (weeks) containing an entry in each record is considerably more complex to extract, although I don't know his method (probably from Countif function after exporting to Excel). Counting observations is very useful for some purposes, such as obviously to produce the graphs showing the timing of the various stages of breeding for the 18 species with the most breeding records shown in the 21 Year editions of The GBS Report (referred to as the Breeding Graphs). However in addition to my earlier note, for these reasons, I recommend that it is flawed in terms of answering the question that Martin was addressing, which was about incidence of local breeding.
My apologies also for earlier writing in terms of "the data", rather than "these data". I often forget that the word "data" is plural.
<Prev in Thread] Current Thread [Next in Thread>

The University of NSW School of Computer and Engineering takes no responsibility for the contents of this archive. It is purely a compilation of material sent by many people to the Canberra Ornithologists Group mailing list. It has not been checked for accuracy nor its content verified in any way. If you wish to get material removed from the archive or have other queries about the list contact David McDonald, list manager, phone (02) 6231 8904 or email . If you can not contact David McDonald e-mail Andrew Taylor at this address: andrewt@cse.unsw.EDU.AU