Peer Annotation Examples

The SCU label is shown, followed by an extract from a peer where the boldface words are those that match the SCU.

Example 1: An easy case.

SCU LABEL: Plaid Cymru is the Welsh nationalist party
Peer text: The minority parties yesterday joined in the pre-hung parliament manoeuvring, with both the Ulster Unionists and Plaid Cymru, the Welsh nationalist party, outlining . . .


Example 2: Text in the peer that partly matches two different SCUS; note that the discriminator between the two SCUs is mainly that the second pertains to the European Community. Note that there is some overlap in the text that matches each SCU. However, it is not the case that the exact same text matches more than one SCU.

SCU LABEL: Plaid Cymru wants full independence
Peer text: Plaid calls for Welsh self-rule within EC.


SCU LABEL: Plaid Cymru wants Wales to have full standing with the European Community
Peer text: Plaid calls for Welsh self-rule within EC.


Example 3: Similar to Example 2, but with conjunction.

SCU Label: When the British assembly establishes a Welsh parliament, it will have law-making powers
Peer text: calls for a Welsh parliament with law-making and financial powers


SCU Label: When the British assembly establishes a Welsh parliament, it will have financial powers
Peer text: calls for a Welsh parliament with law-making and financial powers


Example 4: Illustrating inexact match. Note that the goal of the matching is to indicate whether information expressed in the pyramid is also conveyed in the summary, so it is important to use careful judgment when there is a partial match. In this example, the SCU label and the peer text overlap very well except for a modifier contained in the label that expresses information not expressed in the peer: taxing powers. Since the main clause of the label and the peer text express the same information, and the missing information is a modifier, it is best to match the text to this SCU. Leave the text unmatched to any SCUs would unfairly penalize the summary. This is especially true in the case of SCUs like this one which have a weight of one (one contributor from the model summary), because the pyramid creators attempt to maintain a balance between reducing the amount of information in each SCU, and minimizing the total number of SCUs.

Here, the label reflects the structure of the source text, where "taxing powers" is a modifier; if syntax and word order had given the phrase taxing powers more prominence, then the balance would shift towards rejecting a match in this case; EG: Labor candidates want taxing powers for a Scottish parliament.

SCU Label: Labour MP candidates want a Scottish parliament with taxing powers
Peer text: Labour government had set up a Scottish parliament


Example 5: A negative example. The peer sentence is too non-specific to match any of the SCU's in the pyramid. Six of the twelve SCUs with the word assembly are listed, along with a reason why the label is too specific to be a match.

SCU Label: The Labour Party advocates an elected assembly
(Peer text doesn't mention/imply "Labour Party")

SCU Label: Labour has committed to the creation of a Welsh assembly
(Peer text doesn't mention/imply "Labour" or "commitment")

SCU Label: The British government will establish an elected Welsh assembly
(Peer text doesn't mention/imply "British government" or "establish")

SCU Label: An elected assembly would give Wales more autonomy
(Peer text doesn't mention/imply "autonomy"; SCU label doesn't mention/imply "favor")

SCU Label: Plaid Cymru's policy is to bypass an assembly
(Peer text doesn't mention/imply "Plaid Cymru" or "bypass"; SCU label doesn't mention/imply "favor")

SCU Label: The Labour Party responded to Welsh nationalist protests by favoring an elected assembly
(Peer text doesn't mention/imply "Labour Party")

Peer text: Elected Welsh assembly gains favour


Example 6: A very difficult one. It requires a close reading of the model summaries to know that calls for council reform occurred in 1994, to observe that the SCU Label below matches model text that says "postponed until at least 1995", and then to infer that the reform was to be delayed a year, and also to infer that "the local government reform" refers to the "Welsh council reform". It is usually not necessary to do so much "research" on the pyramid and model summaries. This is a case illustrating the need for judgment: while it is a legitimate match, given the content of the model summaries, it could also be argued the other way given that the label doesn't mention "delay," and the peer doesn't mention a specifc year.

SCU LABEL: Implementation of the local government reform was scheduled for 1995.

Peer text: Welsh council reform to be delayed a year.

Back to Beginning