Credo Reference offers various search options for both experienced and less experienced searchers. Described below are ways that you can use the full capabilities of Credo's search. It also provides helpful details for metasearch systems or others that are creating search forms for the Credo Reference search engine.
How to enter advanced queries
The query examples described below may be entered directly into the search input box found on the main search screen as well as on the search box that appears below the main menu on other pages. You may also enter this syntax on Concept Map searches.
Credo Reference also offers an assisted search form called Advanced Search. This form is designed to help less experienced users take advantage of some of the advanced search features offered by Credo Reference. To power search Credo Reference, you should enter your terms directly into the regular search box not the Advanced Search form.
Credo Reference also offers targeted search forms through its Gadgets. You can think of these as pre-configured power searches, again mainly for the benefit of users unfamiliar with the advanced syntax. For the most part, the advanced syntaxes described below are not intended for use in the gadget search forms.
Prefix operators are symbols that go on the front of a word or phrase and indicate how the word must occur within the retrieved entries. This syntax can be used by itself or in combination with Boolean syntax.
|+||Must occur||All retrieved entries must contain this word or phrase. If a term or phrase has no prefix operator, + is implied.|
|^||Should occur||Retrieved entries with this word should rank higher than those without this word or phrase.|
|!||Shout not occur||Retrieved entries with this word should rank lower than those without this word or phrase.|
|-||Must not occur||Retrieved entries must not contain this word or phrase.|
|"a b"||Phrase||Surround two or more words with double quotes to search that phrase. Word order in the retrieved entries is the same as word order in the phrase.|
|*||Truncation||Search word variations allowing any number of letters to occur where the * is located. Can be used in front of the word, inside the word or at the end of the word.|
|?||Wildcard||Search word variations allowing any single letter to occur where the ? is located. Can be used in front of the word, inside the word or at the end of the word.|
|~||Sounds like||Retrieve entries with words that sound like this word. Useful when the exact spelling of a term is unknown.|
1 Assuming all other score factors are equal.
2 If more than 200 words match the requested pattern, only the first 200 will be searched.
- +"global warming" ^"climate change"
Credo supports Boolean operations and nested queries which allows you to broaden or narrow your search by combining words or phrases using the operators AND, OR, and NOT.
Operations are evaluated left to right. You can change the order of evaluation by using nested parentheses to specify the evaluation order.
|Operator||Function||Diagram of results (shown in purple)|
|and||Both words or phrases must exist. Example: lowell and industrial revolution|
|or||Either word or phrases must exist. Example: USSR or Soviet Union|
|not||The left word or phrase must exist and the right must not exist. Also recognizes "and not" as this operator. Example: football not soccer|
Note: Words and phrases in a Boolean search can use field qualifiers (explained below) or prefix operators (explained above), except the sounds like operator.
- sun or saturn and moon
- sun or (saturn and moon)
- h:george and h:washington not carver
By default, searches find hits within three fields: the entry heading, entry section headings and the full text of an entry. You may control the fields searched by prefixing the search term with the field letter and a colon. The fields are assigned the following letters in the search syntax:
|s:||Entry Section Heading|
Examples: To find entries with global warming in the entry heading and climate change mentioned in the entry body:
h:"global warming" and b:"climate change"
To find entries with a References section heading that mention Milton in the body:
b:Milton and s:references
About the relevance ranking
Several factors determine the final score that each retrieved entry receives. The result set, by default, will list the retrieved entries in a descending relevance order. These are the factors that determine the relevance score for an entry:
|Word frequency||Query terms that occur more often in the retrieved entry will boost the score|
|Hit density||If two entries have the same number of hits, the one with fewer words will have a higher "hit density", causing a boost in the score|
|Word proximity||If the query has two or more words, those entries that have the terms closer together will score higher then those where the terms are farther apart|
|Inverse document frequency||If the query has two or more words, the more common words will have less of an impact on the score than words that are more unique.|
|Entry length||If all other factors are equal, longer entries are boosted over shorter entries|
|Image presence||If all other factors are equal, entries with images are boosted over entries without images.|
|See references||If the retrieved entry is a see reference, its score is lowered|
|Date updated||If the final score is equal, the entry most recently loaded will show first. This favors more recently published material over older material.|
There are no biases in the ranking. Credo does not favor one set of titles over another other than the factors described above
About proximity searching
The knowledge of the ranking rules can sometimes help the power searcher. For example, the Credo query syntax supports phrase searching but not proximity searching (where the phrase words can be allowed in either order or with a specified distance of separation). But the ranking rules will boost entries with closer proximity anyway. So if your desire is to do an 'either – order' proximity phrase search, you might be inclined to enter "a b" or "b a" as your search (given the lack of a proximity operation). But if you just search for a b (not as a phrase, but as two individual words) then you might find that the ranking rules will produce the same result order that a proximity search would have.
For example, person names are usually surname-first in the entry heading but first name- first in the full text prose. Compare the results of these three searches:
"Nixon, Richard" (a phrase query requires word order)
"Nixon, Richard" or "Richard Nixon" (an attempt to allow either order)
Richard Nixon (eliminate the quotes and let the search engine find either order)
All three searches find a top-ten set of entries all about Richard Nixon. But the third query allows for the intervening M. or Milhous that is sometimes in the name and sometimes not. And even though proximity was not enforced, the scoring is such that the top ten all have immediate or nearly immediate proximity anyway. Some new and relevant hits are now part of the top ten that the first two searches excluded or ranked lower.
Credo Reference does not use automatic stemming in either its indexing or query parsing. If you want to search for stemmed variations of a word, use the truncation or sounds like operators described above.
About stop words
Credo Reference does not use a stop word list in its indexing. Any word, including individual letters, prepositions and other common words are all searchable.
Credo Reference does use a list of approximately 25 commonly occurring words (plus the individual letters) as a 'noise word' list. When presented with a query containing noise words, Credo will change the default operator from 'must occur' to 'should occur' for that word. You can override this behavior by prefixing the noise word with a plus (+) to indicate that it must occur. This noise word strategy improves results when Credo is presented with queries that are expressed as questions, like "why is the sky blue?".
Noise Word List
- a..z (all the individual letters)