Appendix A: What was actually submitted to the Web search services.

Return to the Table of Contents.
Services: The actual form used to submit to the search engine:

----------------------------------------------------------------------

Alta Vista http://www.altavista.digital.com/cgi-bin/query?pg=q

Alta Vista a http://www.altavista.digital.com/cgi-bin/query?pg=aq

Excite http://www.excite.com/?xam

Hotbot http://www.hotbot.com/index.html

Infoseek http://www.infoseek.com/Home?pg=ultra_home.html

Lycos http://www.lycos.com/index.html

Lycos a http://www.lycos.com/customsearch.html

General criteria for assigning grades 0 through 3:
  3     A three is for a site that would likely be of use to anyone
	who entered the search query. Such a site would be a one that
	contains a lengthy and general discussion of the topic, or
	one that contains many links to other sites that deal with the
	topic (a webliography, if you will).

  2     A two is for a site that could be of some conceivable use to 
	someone searching the topic, but not necessarily. Twos are given
	to sites that focus on a narrow aspect of the subject, or that 
	deal with the subject, but not thoroughly. For instance, in 
	"unemployment rate" +Illinois, a site that link to many pages
	containing Illinois unemployment rates might get a three, but a page 
	that reported the Illinois unemployment rate for only one month would
	only get a two. Also, a page with a link to a three, but without 
	other qualifications, will get a two. 

  1     A site that is technically correct, but not useful. The html code for 
	the page did satisfy the required components of the query. However, 
	the components may not have been in the correct context. To continue 
	with #2, the unemployment rate, if a page gave the unemployment rate 
	of some other place, but then happened to mention the word "Illinois," 
	it technically satisfied the query without being useful. In mechanical
	terms, the incorrect subject is more likely to have the components
	of the query farther apart in proximity. The query might be Boolean
	or unstructured, but the useful information had higher proximity. 
	Search engines would occassionally return a dictionary page as the 
	result to a query. Indeed, all of the words in the query were found
	in the dictionary, but it was of no conceivable use to someone who
	had issued the query. Also, pages that mentioned the correct topic,
	but contained only a minimum of actual information that could be 
	useful to someone conducting the search were given ones. 

  0     A site that did not satisfy the query as submitted was given a zero.
	In the unemployment case, if the either the word Illinois or the 
	phrase "unemployment rate" was missing, a zero was given. In #3, a 
	zero was avoided if the word "athletes" was present and either of 
	the phrases "wife abuse" or "domestic violence" was present.


PART I: Queries that could be used because of how they were selected 
and examined.


Query 1: Simple: Information on the Individuals with Disabilities Education 
		  Act
	Source: Obtained locally
	Searched: 1/31/97
----------------------------------------------------------------------
Alta Vista      Individuals with Disabilities Education Act
Excite          Individuals with Disabilities Education Act
Hotbot          All the words: Individuals with Disabilities Education Act
Infoseek        Individuals with Disabilities Education Act
Lycos           Individuals with Disabilities Education Act

Criteria:
  3     A site with a lengthy explanation of the Act or a site
	with links to many other resources.
  2     A site with some discussion of the Act, or focusing on some
	aspect of the act. 
  1     A site that contained all of the words but did not deal with the
	topic, or one that mentioned the topic, but was insubstantial.
  0     HTML Code for the page lacked some of the key words of the 
	search, like "Individuals" or "Disabilities" or "Education".

Query 2: Structured: Find unemployment rate statistics for Illinois
	Source: Obtained locally
	Searched: 2/1/97
----------------------------------------------------------------------                                               
Alta Vista a    +"unemployment rate" AND +Illinois
Excite          +"unemployment rate" AND +Illinois
Hotbot          All of the words: unemployment rate Illinois
		Modify: must contain the exact phrase: "unemployment rate"
Infoseek        +"unemployment rate"  +Illinois
Lycos      a    Match all words, close match: unemployment rate Illinois

Criteria:
  3     A site with a full analysis of the unemployment rate of Illinois, or
	with links to pages that listed the rate for different time periods.
  2     A site that cited the unemployment rate for Illinois, regardless of 
	how recent, or that mentioned the unemployment rate for Illinois. 
  1     A site that contained the phrase "unemployment rate" and the word 
	Illinois, but did not deal with the rate in Illinois.
  0     HTML Code for the page lacked some of the key words of the 
	search, like "unemployment rate" or "Illinois".


Query 3: Structured: Information on the problem of domestic violence among
	athletes
	Source: Obtained locally
	Searched: 2/8/97
----------------------------------------------------------------------
Alta Vista a    +athletes AND ("domestic violence" OR "wife abuse")
Excite          +athletes AND ("domestic violence" OR "wife abuse")
Hotbot          Any of the words: athletes "domestic violence" "wife abuse"
		Modify: must contain the word: athletes
		Modify: should contain the exact phrase: "domestic violence"
		Modify: should contain the exact phrase: "wife abuse"
Infoseek        +athletes "domestic violence"  "wife abuse"
Lycos      a    Match 3 words: athletes sports domestic violence wife abuse


Criteria:
  3     A site with a lengthy treatment of the domestic violence among
	athletes, or with links to many pages that discussed it.
  2     A site that made some mention of the topic, or dealt with some
	aspect of it.
  1     A site that contained the phrase "domestic violence" or "wife abuse" 
	and the word athletes, but did not deal with the two together.
  0     HTML Code for the page lacked some of the key words of the 
	search, like "domestic violence" or "wife abuse" or "athletes".


Query 4: Simple: animal husbandry
	Source: Tomaiuolo and Packer study
	Searched: 2/14/97
----------------------------------------------------------------------
Alta Vista      animal husbandry
Excite          animal husbandry
Hotbot          All of the words: animal husbandry
Infoseek        animal husbandry
Lycos           animal husbandry

Criteria:
  3     A site which had a lengthy treatment of the topic of animal husbandry, 
	or which provided links to many pages that discussed it.
  2     A site that made some mention of the topic, or dealt with some
	aspect of it.
  1     A site that contained the words animal and husbandry, but did not 
	deal with the topic meant by two together in a substantial way.
  0     HTML Code for the page lacked some of the key words of the 
	search, like "animal" or "husbandry".


Query 5: Simple: cable television regulation
	Source: Tomaiuolo and Packer study
	Searched: 2/14/97
----------------------------------------------------------------------
Alta Vista      cable television regulation
Excite          cable television regulation
Hotbot          All of the words: cable television regulation
Infoseek        cable television regulation
Lycos           cable television regulation
   
Criteria:
  3     A site which had a lengthy treatment of the topic of cable television
	regulation, or which provided links to many pages that discussed it.
  2     A site that made some mention of the topic, or dealt with some
	aspect of it.
  1     A site that contained the words cable, television and regulation, 
	but did not deal with the topic in a substantial way.
  0     HTML Code for the page lacked some of the key words of the 
	search, like "cable" or "regulation".


Query 6: Simple: MTX International Corporation
	Source: Obtained locally
	Searched: 2/14/97
----------------------------------------------------------------------
Alta Vista      MTX International 
Excite          MTX International 
Hotbot          All of the Words: MTX International 
Infoseek        MTX International 
Lycos           MTX International 
   
Criteria:
  3     A site which had substantial information on the MTX International
	corporation.
  2     A site that provided some substantial information about the company,
	or one that provided a link to the MTX International homepage.
  1     A site that happened to mention the company, say in a list of 
	products supplied, but no information about the company, or that 
	contained the words MTX and Internation, but did not deal with the 
	company "MTX International".
  0     HTML Code for the page lacked some of the key words of the 
	search, like "MTX" or "International".


Query 7: Simple: cell cryopreservation
	Source: Tomaiuolo and Packer study
	Searched: 2/20/97
----------------------------------------------------------------------
Alta Vista      cell cryopreservation
Excite          cell cryopreservation
Hotbot          All of the words: cell cryopreservation
Infoseek        cell cryopreservation
Lycos           cell cryopreservation
   
Criteria:
  3     A site which had a lengthy treatment of the topic of cell 
	cryopreservation, or which provided links to many pages that 
	discussed it.
  2     A site that made some mention of the topic, or dealt with some
	aspect of it.
  1     A site that contained the words cell and cryopreservation, 
	but did not deal with the topic in a substantial way.
  0     HTML Code for the page lacked some of the key words of the 
	search, like "cell" or "cryopreservation".


Query 8: Simple: classical architecture
	Source: Tomaiuolo and Packer study
	Searched: 2/21/97
----------------------------------------------------------------------
Alta Vista      classical architecture
Excite          classical architecture
Hotbot          All of the words: classical architecture
Infoseek        classical architecture
Lycos           classical architecture
   
Criteria:
  3     A site which had a lengthy treatment of the topic of classical 
	architecture, or which provided links to many pages that dealt with 
	it.
  2     A site that gave some treatment of the topic, or dealt with some
	aspect of it, such as a thorough discussion of an academic program
	that focused on classical architecture. "Classical" was left to 
	the site to define. Roman, Greek and Turkish classical pages were
	discovered. 
  1     A site that contained the words classical and architecture, 
	but did not deal with the topic in a substantial way.
  0     HTML Code for the page lacked some of the key words of the 
	search, like "classical" or "architecture".



Query 9: Simple: ecotourism
	Source: Tomaiuolo and Packer study
	Searched: 2/24/97
----------------------------------------------------------------------
Alta Vista      ecotourism
Excite          ecotourism
Hotbot          ecotourism
Infoseek        ecotourism
Lycos           ecotourism
   
Criteria:
  3     A site which had a lengthy treatment of the topic of ecotourism, 
	or which provided links to pages from more than one tour company
	that organizes ecotours.
  2     A site that discussed the topic of ecotourism, or provided links
	to pages describing several different ecotours offered by one tour
	company. 
  1     A site that contained the word ecotourism, but had little discussion 
	of the topic, or only described one tour offered by a commercial 
	tour provider. 
  0     HTML Code for the page lacked the word "ecotourism".



Query 10: Simple: In Focus Systems  
	Source: Obtained locally
	Searched: 2/28/97
----------------------------------------------------------------------
Alta Vista      In Focus Systems  
Excite          In Focus Systems  
Hotbot          All of the Words: In Focus Systems  
Infoseek        In Focus Systems  
Lycos           In Focus Systems  
   
Criteria:
  3     A site which had substantial information on the In Focus Systems
	corporation, such as the In Focus Systems homepage.
  2     A site that provided some substantial information about the company,
	or one that provided a link to the In Focus Systems homepage.
  1     A site that happened to mention the company, say in a list of 
	products supplied, but no information about the company.
  0     HTML Code for the page lacked some of the key words of the 
	search, like "In" or "Systems".



Query 11: Structured: Find information on the National Gallery of Prague
	Source: Obtained locally
	Searched: 3/1/97
----------------------------------------------------------------------
Alta Vista a    "National Gallery" AND +Prague
Excite          "National Gallery" AND +Prague
Hotbot          Exact phrase: National Gallery
		Modify: must contain the word: Prague
Infoseek        "National Gallery" +Prague
Lycos      a    Match 2 words, Loose Match: National Gallery Prague  
   
Criteria:
  3     The homepage for National Gallery of Prague, or a page that had
	an extensive discussion of it or links to exhibits from it
	(only the National Gallery of Prague homepage qualified).
  2     A site that gave information about the National Gallery of Prague 
	or one that provided a link to the National Gallery of Prague
	homepage.
  1     A site that contained the phrase "National Gallery" and word Prague, 
	but did not deal with the two together.
  0     HTML Code for the page lacked some of the key words or 
	phrases of the search, like "National Gallery" or "Prague".



Query 12: Structured: The effects of caffeine on aerobic exercise
	Source: Obtained locally
	Searched: 2/27/97
----------------------------------------------------------------------
Alta Vista a    +caffeine AND +aerobic AND exercise AND effects
Excite          +caffeine AND +aerobic AND exercise AND effects
Hotbot          All the words: caffeine aerobic exercise effects
		Modify: must contain the word: caffeine
		Modify: must contain the word: aerobic
Infoseek        +caffeine +aerobic exercise effects
Lycos      a    Match three words, loose: caffeine aerobic exercise effects

Criteria:
  3     A site with a lengthy treatment of the effects of caffeine on human or
	animal performance doing aerobic exercise, or with links to many 
	pages that discussed it.
  2     A site that made some mention of the topic, or dealt with some
	aspect of it.
  1     A site that contained the words "caffeine" and "aerobic" 
	but did not deal with the effects of caffeine on aerobic exercise.
  0     HTML Code for the page lacked some of the key words of the 
	search, like "caffeine" or "aerobic."


Query 13: Structured: Find information on the group called "Queer Nation"
	Source: Obtained locally
	Searched: 3/5/97
----------------------------------------------------------------------
Alta Vista a    "Queer Nation"
Excite          "Queer Nation"
Hotbot          Exact Phrase: Queer Nation
Infoseek        "Queer Nation"
Lycos      a    Match All Words, Loose Match: Queer Nation

Criteria:
  3     A site with a lengthy treatment of the group called Queer Nation,
	or provide many links to pages that discussed it.
  2     A site that made some mention of the group, or dealt with some
	aspect of it, or dealt with the Australian club called Queer Nation.
  1     A site that contained the phrase "Queer Nation" but did not discuss
	it, or one that contained the phrase but had no substantial 
	information.
  0     HTML Code for the page lacked the phrase "Queer Nation".



Query 14: Simple: Find information about prozac, but not the rock group
	Source: Obtained locally
	Searched: 3/12/97
----------------------------------------------------------------------
Alta Vista      prozac -rock
Excite          prozac -rock
Hotbot          All of the Words: prozac 
		Modify: must not contain the word: rock
Infoseek        prozac -rock
Lycos           prozac -rock
   
Criteria:
  3     A site which had a lengthy treatment of the topic of prozac, 
	or which provided links to many pages that discussed it, without
	containing the word "rock," as in rock group.
  2     A site that made some mention of the topic, or dealt with some
	aspect of it, without containing the word "rock," as in rock group.
  1     A site that contained the word prozac and not the word "rock,"
	but did not deal with the topic of prozac, the drug, in a substantial 
	way. (This situation was common, as the word prozac has become a
	popular term in popular culture.)
  0     HTML Code for the page lacked the key word "prozac" or contained 
	the word "rock".



Query 15: Find information on Maria Luisa Bombal
	Source: Obtained locally
	Searched: 3/12/97
----------------------------------------------------------------------
Alta Vista      Maria Luisa Bombal
Excite          Maria Luisa Bombal
Hotbot          The Person: Maria Luisa Bombal
Infoseek        Maria Luisa Bombal
Lycos           Maria Luisa Bombal
   
Criteria:
  3     A site which had a lengthy treatment of the writer, or which 
	provided links to many pages that discussed her (there were none).
  2     A site that had any information at all on her, other than citations 
	to books by her or books about her. This category was stretched
	a little further for this query than for others, because of the 
	lack of pages that discussed her. 
  1     A site that contained the name "Maria Luisa Bombal" but had little
	substantial infomation about her. Included here were pages that
	cited one of her books. 
  0     HTML Code for the page lacked the key name Maria Luisa Bombal.
	Lyocs had the most zeros, because it found many pages about other
	women who were named "Maria Luisa."


Appendix B: Raw scores for the Experiments and the Numbers of Inactive, Duplicate and Zero links

Table B1a: Score of 1, 2 or 3 in the first: 
three results/ ten results/ twenty results

Query \ Serv.   Alta V   Excite   HotBot    Infoseek    Lycos
---------------------------------------------------------------------------
1 IDEA          2/9/19   3/10/20  2/5/6     3/8/18      3/8/14
2 u. rate       3/9/16   3/9/16   2/7/12    3/10/18     0/0/0 (0)
3 athletes      3/9/18   3/10/20  3/10/18   1/5/13      1/1/1
4 animal h      3/10/17  3/9/19   2/6/13    3/9/17      2/6/15
5 cable t       3/9/18   3/9/18   2/7/15    3/8/17      2/8/13
6 MTXI          3/10/11  3/10/20  2/9/14    3/9/17      3/3/3
7 cell cry      3/10/20  3/10/20  3/9/19    3/10/20     2/7/14
8 classical     3/10/20  3/10/20  2/5/12    3/8/17      3/10/20
9 ecotour       2/9/19   2/8/18   2/8/17    3/9/15      3/9/17
10 IFS          3/8/16   3/10/20  0/2/3     3/10/20     1/1/2
11 Nat Gall     2/6/15   3/10/19  1/7/13    2/7/14      0/0/0
12 caffeine     3/8/17   3/10/20  3/9/12    3/7/12      0/1/3 (16)
13 QN           2/7/14   3/5/13   2/5/11    3/6/14      1/6/8
14 prozac -r    3/10/18  3/8/17   3/8/12    2/7/17      3/9/14
15 ML Bombal    3/10/19  3/10/18  3/9/14    2/8/11 (14) 2/3/3
                              (18)

Table B1b: The weighted score from test 1 

Query \ Serv.   Alta V  Excit   HotBot  Infosk  Lycos
-------------------------------------------------------
1 IDEA          .93     1.00    .36     .88     .73
2 u. rate       .83      .83    .63     .93      0
3 athletes      .90     1.00    .93     .60     .07
4 animal h      .89      .94    .64     .87     .71
5 cable t       .90      .90    .73     .84     .69
6 MTXI          .68     1.00    .75     .87     .22
7 cell cry     1.00     1.00    .94    1.00     .70
8 classical    1.00     1.00    .58     .84    1.00
9 ecotour       .93      .87    .83     .80     .87
10 IFS          .81     1.00    .16    1.00     .11
11 Nat Gall     .71      .96    .65     .70      0
12 caffeine     .84     1.00    .69     .64     .15
13 QN           .70      .62    .54     .68     .45
14 prozac -r    .93      .84    .66     .81     .76
15 ML Bombal    .96     1.00    .76     .79     .20
       

Table B1c: Mean, Standard Deviation and Standard Error of Mean for Services 
for Test 1

Serv\Statistic  N       Mean            Stan. Dev.      St. Error
-----------------------------------------------------------------
Alta Vista      15      0.8676          0.1048          0.0271
Excite          15      0.9314          0.1057          0.0273
Hotbot          15      0.6566          0.2013          0.0520
Infoseek        15      0.8158          0.1197          0.0309
Lycos           15      0.4440          0.3492          0.0902

All together    75      0.7431          0.2621          0.0303


Table B2a: Score 2 or 3 in the first: three results/ ten results/ 
twenty results

Query \ Serv.   Alta V   Excite    HotBot    Infoseek    Lycos
---------------------------------------------------------------------------
1 IDEA          2/8/15   3/7/11    2/5/5     3/7/15      3/7/11
2 u. rate       1/2/4    3/8/14    1/5/5     3/9/13      0/0/0 (0)
3 athletes      0/0/3    2/7/11    1/1/3     1/5/11      1/1/1
4 animal h      3/8/12   2/5/7     1/3/5     3/8/13      0/4/7
5 cable t       2/6/9    3/7/11    2/3/6     0/2/4       1/5/10
6 MTXI          2/7/7    2/2/3     1/6/8     2/5/11      1/1/1
7 cell cry      1/6/12   1/8/12    0/1/6     1/2/8       0/0/2
8 classical     3/10/18  1/5/12    0/0/4     3/5/9       3/8/17
9 ecotour       2/8/13   2/5/7     0/2/5     1/5/8       2/5/7
10 IFS          3/8/16   3/10/18   0/1/1     1/8/16      1/1/2
11 Nat Gall     0/1/4    3/8/15    1/6/12    2/5/10      0/0/0
12 caffeine     0/1/2    2/2/3     1/4/4     1/3/6       0/0/2 (16)
13 QN           1/2/6    3/5/7     2/4/10    3/5/10      0/2/4
14 prozac -r    2/7/10   0/4/10    0/2/5     2/5/15      1/4/9
15 ML Bombal    2/3/4    0/1/2(18) 0/1/2     1/2/3 (14)  0/0/0


Table B2b: The weighted score from test 2

Query \ Serv.   Alta V  Excit   HotBot  Infosk  Lycos
-------------------------------------------------------
1 IDEA          .76     .60     .33     .75     .60
2 u. rate       .20     .73     .32     .72      0
3 athletes      .11     .59     .14     .53     .07
4 animal h      .66     .40     .27     .70     .35
5 cable t       .49     .60     .31     .19     .49
6 MTXI          .45     .18     .45     .54     .07
7 cell cry      .59     .64     .24     .35     .07
8 classical     .93     .57     .14     .48     .84
9 ecotour       .69     .40     .23     .42     .40
10 IFS          .81     .93     .06     .78     .11
11 Nat Gall     .17     .77     .59     .51      0
12 caffeine     .10     .18     .25     .27     .07     
13 QN           .28     .41     .48     .52     .19
14 prozac -r    .56     .46     .23     .68     .43
15 ML Bombal    .24     .14     .10     .21      0


Table B2c: Mean, Standard Deviation and Standard Error of Mean for Services 
for Test 2

Serv\Statistic  N       Mean            Stan. Dev.      St. Error
-----------------------------------------------------------------
Alta Vista      15      0.4686          0.2723          0.0703
Excite          15      0.5042          0.2328          0.0601
Hotbot          15      0.2757          0.1454          0.0375
Infoseek        15      0.5128          0.1909          0.0493
Lycos           15      0.2481          0.2584          0.0667

All together    75      0.4019          0.2477          0.0286


Table B3a: Score of 3 in the first: 
three results/ ten results/ twenty results

Query \ Serv.   Alta V   Excite   HotBot    Infoseek    Lycos
--------------------------------------------------------------------------
1 IDEA          1/4/4    2/2/2    1/2/2     1/2/3       2/2/2
2 u. rate       1/1/1    1/1/1    0/0/0     3/5/5       0/0/0 (0)
3 athletes      0/0/2    2/2/2    0/0/2     0/1/3       1/1/1  
4 animal h      0/2/4    1/2/2    1/1/3     3/6/9       0/1/1
5 cable t       1/2/3    0/1/2    1/1/1     0/0/1       1/3/3
6 MTXI          0/0/0    1/1/1    0/0/0     1/1/2       0/0/0
7 cell cry      1/2/4    0/0/0    0/0/1     1/1/2       0/0/0
8 classical     1/1/1    0/3/4    0/0/2     0/1/2       0/1/2
9 ecotour       1/2/3    2/3/4    0/0/0     1/2/2       0/0/0
10 IFS          1/1/1    1/1/2    0/0/0     1/1/2       0/0/0
11 Nat Gall     0/0/0    0/0/1    1/1/1     1/1/1       0/0/0
12 caffeine     0/0/0    1/1/1    1/2/2     0/1/1       0/0/0
13 QN           0/0/0    0/0/0    0/0/0     0/0/1       0/0/0
14 prozac -r    0/0/1    0/0/0    0/0/0     1/2/4       0/0/1
15 ML Bombal    0/0/0    0/0/0    0/0/0     0/0/0       0/0/0


Table B3b: The weighted score from test 3

Query \ Serv.   Alta V  Excit   HotBot  Infosk  Lycos
-------------------------------------------------------
1 IDEA          .25     .14     .13     .17     .14
2 u. rate       .07     .07     0       .34     0
3 athletes      .07     .14     .07     .13     .07
4 animal h      .19     .13     .14     .51     .06
5 cable t       .17     .10     .07     .04     .19
6 MTXI          0       .07     0       .11     0
7 cell cry      .20     0       .04     .11     0
8 classical     .07     .22     .07     .10     .10
9 ecotour       .17     .24     0       .13     0
10 IFS          .07     .11     0       .11     0
11 Nat Gall     0       .04     .07     .07     0
12 caffeine     0       .07     .13     .06     0 
13 QN           0       0       0       .04     0
14 prozac -r    .04     0       0       .20     .04
15 ML Bombal    0       0       0       0       0       


Table B3c: Mean, Standard Deviation and Standard Error of Mean for Services
for Test 3

Serv\Statistic  N       Mean            Stan. Dev.      St. Error
-----------------------------------------------------------------
Alta Vista      15      0.0875          0.0878          0.0227
Excite          15      0.0889          0.0771          0.0199
Hotbot          15      0.0487          0.0547          0.0141
Infoseek        15      0.1403          0.1294          0.0334
Lycos           15      0.0401          0.0617          0.0160

All together    75      0.0811          0.0912          0.0105


Table B4a (HotBot special): Score 2 or 3 minus duplicates in the first:
        three results/ ten results/ twenty results

Query \ Serv.   Alta V   Excite      HotBot       Infoseek    Lycos
--------------------------------------------------------------------------
1 IDEA          2/9/19   3/10/20     3/6/6 (9)    3/8/18      3/8/14
2 u. rate       3/9/16   3/9/16      3/7/12 (18)  3/10/18     0/0/0 (0)
3 athletes      3/9/18   3/10/20     3/10/18      1/5/13      1/1/1
4 animal h      3/10/17  3/9/19      2/7/13 (16)  3/9/17      2/6/15
5 cable t       3/9/18   3/9/18      2/7/15 (18)  3/8/17      2/8/13
6 MTXI          3/10/11  3/10/20     2/9/14 (16)  3/9/17      3/3/3
7 cell cry      3/10/20  3/10/20     3/9/19 (19)  3/10/20     2/7/14
8 classical     3/10/20  3/10/20     3/8/12 (16)  3/8/17      3/10/20
9 ecotour       2/9/19   2/8/18      2/8/17       3/9/15      3/9/17
10 IFS          3/8/16   3/10/20     0/2/3 (19)   3/10/20     1/1/2
11 Nat Gall     2/6/15   3/10/19(19) 2/9/13 (15)  2/7/14(17)  0/0/0
12 caffeine     3/8/17   3/10/20     3/10/12 (13) 3/7/12(18)  0/1/3 (16)
13 QN           2/7/14   3/5/13      2/6/11 (16)  3/6/14      1/6/8
14 prozac -r    3/10/18  3/8/17      3/9/12 (13)  2/7/17      3/9/14
15 ML Bombal    3/10/19  3/10/18     3/9/14 (19)  2/8/11 (14) 2/3/3
                              

Table B4b: The weighted score from test 4

Query \ Serv.   Alta V  Excit   HotBot  Infosk  Lycos
-------------------------------------------------------
1 IDEA          .93     1.00    .65     .88     .73
2 u. rate       .83      .83    .69     .93      0
3 athletes      .90     1.00    .93     .60     .07
4 animal h      .89      .94    .77     .87     .71
5 cable t       .90      .90    .79     .84     .69
6 MTXI          .68     1.00    .87     .87     .22
7 cell cry     1.00     1.00   1.00    1.00     .70
8 classical    1.00     1.00    .77     .84    1.00
9 ecotour       .93      .87    .83     .80     .87
10 IFS          .81     1.00    .16    1.00     .11
11 Nat Gall     .71     1.00    .87     .81      0
12 caffeine     .84     1.00    .95     .64     .15
13 QN           .70      .62    .66     .68     .45
14 prozac -r    .93      .84    .92     .81     .76
15 ML Bombal    .96     1.00    .79     .79     .20


Table B4c: Mean, Standard Deviation and Standard Error of Mean for Services 
for Test 4

Serv\Statistic  N       Mean            Stan. Dev.      St. Error
-----------------------------------------------------------------
Alta Vista      15      0.8676          0.1048          0.0271
Excite          15      0.9338          0.1068          0.0275
Hotbot          15      0.7692          0.1947          0.0503
Infoseek        15      0.8309          0.1055          0.0273
Lycos           15      0.4440          0.3492          0.0902

All together    75      0.7691          0.2574          0.0297


Table B5a (HotBot special): Score 2 or 3 minus duplicates in the first:
        three results/ ten results/ twenty results

Query \ Serv.   Alta V   Excite      HotBot       Infoseek    Lycos
--------------------------------------------------------------------------
1 IDEA          2/8/15   3/7/11      2/5/5 (9)    3/7/15      3/7/11
2 u. rate       1/2/4    3/8/14      2/5/5 (18)   3/9/13      0/0/0 (0)
3 athletes      0/0/3    2/7/11      1/1/3        1/5/11      1/1/1
4 animal h      3/8/12   2/5/7       1/3/5 (16)   3/8/13      0/4/7
5 cable t       2/6/9    3/7/11      2/3/6 (18)   0/2/4       1/5/10
6 MTXI          2/7/7    2/2/3       1/6/8 (16)   2/5/11      1/1/1
7 cell cry      1/6/12   1/8/12      0/1/6 (19)   1/2/8       0/0/2
8 classical     3/10/18  1/5/12      0/2/4 (16)   3/5/9       3/8/17
9 ecotour       2/8/13   2/5/7       0/2/5        1/5/8       2/5/7
10 IFS          3/8/16   3/10/18     0/1/1 (19)   1/8/16      1/1/2
11 Nat Gall     0/1/4    3/8/15      2/8/12 (15)  2/5/10      0/0/0
12 caffeine     0/1/2    2/2/3       1/4/4 (17)   1/3/6       0/0/2 (16)
13 QN           1/2/6    3/5/7       2/5/10 (16)  3/5/10      0/2/4
14 prozac -r    2/7/10   0/4/10      0/3/5 (13)   2/5/15      1/4/9
15 ML Bombal    2/3/4    0/1/2 (18)  0/1/2 (19)   1/2/3 (14)  0/0/0


Table B5b: The weighted score from test 5

Query \ Serv.   Alta V  Excit   HotBot  Infosk  Lycos
-------------------------------------------------------
1 IDEA          .76     .60     .53     .75     .60
2 u. rate       .20     .73     .35     .72      0
3 athletes      .11     .59     .14     .53     .07
4 animal h      .66     .40     .31     .70     .35
5 cable t       .49     .60     .33     .19     .49
6 MTXI          .45     .18     .52     .54     .07
7 cell cry      .59     .64     .25     .35     .07
8 classical     .93     .57     .23     .48     .84
9 ecotour       .69     .40     .23     .42     .40
10 IFS          .81     .93     .06     .78     .11
11 Nat Gall     .17     .80     .79     .59      0
12 caffeine     .10     .18     .29     .32     .07     
13 QN           .28     .41     .55     .52     .19
14 prozac -r    .56     .46     .34     .68     .43
15 ML Bombal    .24     .14     .10     .21      0


Table B5c: Mean, Standard Deviation and Standard Error of Mean for Services 
for Test 5

Serv\Statistic  N       Mean            Stan. Dev.      St. Error
-----------------------------------------------------------------
Alta Vista      15      0.4686          0.2723          0.0703
Excite          15      0.5061          0.2352          0.0607
Hotbot          15      0.3401          0.1992          0.0514
Infoseek        15      0.5202          0.1902          0.0491
Lycos           15      0.2481          0.2584          0.0667

All together    75      0.4166          0.2505          0.0289


Table B6a: First Twenty--dead links, duplicates, zeros

Query \ Serv.   Alta V   Excite   HotBot    Infoseek    Lycos
-----------------------------------------------------------------------------------
1 IDEA          1/0/0    0/0/0     1/11/2    2/0/0       2/0/4
2 u. rate       3/0/1    3/0/1     5/2/1     2/0/0       0/0/0 (0)
3 athletes      0/0/2    0/0/0     0/0/2     6/0/1       2/0/17
4 animal h      3/0/0    1/0/0     1/4/2     3/0/0       5/0/0
5 cable t       1/0/1    2/0/0     2/2/1     3/0/0       5/0/2
6 MTXI          0/0/9    0/0/0     1/4/1     2/0/1       6/0/11
7 cell cry      0/0/0    0/0/0     0/1/0     0/0/0       5/0/1
8 classical     0/0/0    0/0/0     2/4/2     3/0/0       0/0/0
9 ecotour       1/0/0    2/0/0     3/0/0     4/0/0       3/0/0
10 IFS          1/0/3    0/0/0     4/1/12    0/0/0       7/0/11
11 Nat Gall     3/0/2    0/1/0     1/5/1     3/3/0       7/0/13
12 caffeine     2/0/1    0/0/0     1/7/0     6/2/0       4/0/9 (16)
13 QN           6/0/0    6/0/1     1/4/4     5/0/1       7/0/5
14 prozac -r    2/0/0    2/0/1     1/7/0     2/0/1       6/0/0
15 ML Bombal    0/0/1    0/0/0(18) 4/1/1     2/0/1 (14)  5/0/12


Table B6b: Median, Mean and Mode for Dead, Duplicate and Zero Links

Query \ Serv.   Alta V   Excite   HotBot    Infoseek    Lycos
-----------------------------------------------------------------------------------
Dead Median     1        0       1           3[2..5]*    4.5[3..6]* 
Dead Mean       1.53     1.07    1.80        2.87        4.27
Dead Mode       0        0       1           2           5


Dup Median      0        0       4           0           0 
Dup Mean        0        0.06    3.53        0.33        0
Dup Mode        0        0       4           0           0


Zero Median     1        0       1           0           5.5 [1..8.5]*
Zero Mean       1.33     0.20    1.93        0.33        5.67
Zero Mode       0        0       1           0           0


*Used an estimated population median and a confidence interval for
the median based on the Wilcoxon test.

Appendix C: The Numerators and Denominators for the Five Experiments

QUERY ENGINE NUM1 NUM2 NUM3 DEN123 NUM4 NUM5 DEN45
1 1 259 212 71 279 259 212 279
2 1 232 57 20 279 232 57 279
3 1 252 30 20 279 252 30 279
4 1 249 185 54 279 249 185 279
5 1 252 138 47 279 252 138 279
6 1 189 125 0 279 189 125 279
7 1 279 165 57 279 279 165 279
8 1 279 259 20 279 279 259 279
9 1 259 192 47 279 259 192 279
10 1 225 225 20 279 225 225 279
11 1 198 47 0 279 198 47 279
12 1 235 27 0 279 235 27 279
13 1 195 77 0 279 195 77 279
14 1 259 155 10 279 259 155 279
15 1 269 67 0 279 269 67 279
1 2 279 168 40 279 279 168 279
2 2 232 205 20 279 232 205 279
3 2 279 165 40 279 279 165 279
4 2 262 111 37 279 262 111 279
5 2 252 168 27 279 252 168 279
6 2 279 50 20 279 279 50 279
7 2 279 179 0 279 279 179 279
8 2 279 158 61 279 279 158 279
9 2 242 111 67 279 242 111 279
10 2 279 259 30 279 279 259 279
11 2 269 215 10 279 269 215 269
12 2 279 50 20 279 279 50 279
13 2 174 114 0 279 174 114 279
14 2 235 128 0 279 235 128 279
15 2 259 27 0 259 259 27 259
1 3 101 91 37 279 111 94 172
2 3 175 88 0 279 178 91 259
3 3 259 40 20 279 259 40 279
4 3 178 74 40 279 185 74 239
5 3 205 87 20 279 205 87 249
6 3 209 125 0 279 209 125 239
7 3 262 67 10 279 269 67 269
8 3 161 40 20 279 185 54 239
9 3 232 64 0 279 232 64 279
10 3 44 17 0 279 44 17 269 
11 3 182 165 20 279 199 182 229
12 3 192 71 37 279 199 71 249
13 3 151 134 0 279 158 141 239
14 3 185 64 0 279 192 71 209
15 3 212 27 0 279 212 27 269
1 4 245 208 47 279 245 208 279
2 4 259 202 94 279 259 202 279
3 4 168 148 37 279 168 148 279
4 4 242 195 141 279 242 195 279
5 4 235 54 10 279 235 54 279
6 4 242 151 30 279 242 151 279
7 4 279 97 30 279 279 97 279
8 4 235 134 27 279 235 134 279
9 4 222 118 37 279 222 118 279
10 4 279 219 30 279 279 219 279
11 4 195 141 20 279 202 148 249
12 4 178 84 17 279 195 84 259
13 4 191 144 10 279 191 144 279
14 4 225 191 57 279 225 191 279
15 4 172 47 0 219 172 47 219
1 5 205 168 40 279 205 168 279
2 5 0 0 0 79 0 0 79
3 5 20 20 20 279 20 20 279
4 5 198 98 17 279 198 98 279
5 5 192 138 54 279 192 138 279
6 5 60 20 0 279 60 20 279
7 5 195 20 0 279 195 20 279
8 5 279 235 27 279 279 235 279
9 5 242 111 0 279 242 111 279
10 5 30 30 0 279 30 30 279
11 5 0 0 0 279 0 0 279
12 5 37 20 0 239 37 20 239
13 5 125 54 0 279 125 54 279
14 5 212 121 10 279 212 121 279
15 5 57 0 0 279 57 0 279

Return to the Table of Contents.
Questions: Please contact Vernon Leighton at vleighton@winona.edu Last updated 6/16/97