Calculating world normalised web indicators (for multiple fields and/or years)

To calculate the world normalised indicators for the web data (e.g. Wikipedia citations), select Calculate MNLCS, gMNCS and NPC for a set of web searches (multiple files with structured names) from the Reports menu. From the dialog box that appears next, select the Wikipedia citations structured names folder containing the Bing API search results files and click OK.  When the report is ready, click No to avoid creating a second report.

The results should be very similar to the structured citation analysis report and Mendeley indicator reports but with many more rows per file. The reason why there are more rows per file is that results are reported counted by unique URL, domain, website STLD (Top Level Domain, such as .com, .ac.uk) and TLD (Top Level Domain, such as .com, uk). Normally, only the URL counting (i.e., counting all URLs matching each query) or domain counting (i.e., counting the number of web domains matching each query) should be used. For Wikipedia, it is best to use URL counting but for other searches it is safer to use domain counting because this eliminates the possibility that the results include duplicate pages or duplicate content within the same website. The (huge) top part, for individual files, is repeated below, with the relevant lines highlighted in bold – the remainder can safely be completely ignored.

Looking at the top Biochemistry Molecular Biology Alcohol 2012 set, for example, these articles have an arithmetic mean of only 0.026 Wikipedia citations per article, and the proportion of articles with at least 1 Wikipedia citation is 0.020080, or 2.0%. For the corresponding Spain set, the proportion of articles with at least 1 Wikipedia citation is marginally higher 0.020725, or 2%, but the confidence interval for Spain (0.008089, 0.052069) comfortably contains the world average so there is insufficient evidence to conclude that Spanish articles are more cited in Wikipedia than the world average. This is confirmed by the world normalised proportion non-zero (NPC) for Spain being slightly higher than 1 at 1.032124, but the confidence interval for it comfortably containing 1: (0.346414, 3.075162).  Also note that right at the bottom some confidence intervals are (NaN, NaN). Here, NaN stands for “Not a Number” and means that the confidence interval is effectively infinite.
World  File: Biochemistry Molecular Biology Alcohol 2012
Queries   : 498
Arithmetic  mean (unique URLs)   : 0.026104
Arithmetic mean (unique domains)   : 0.024096
Arithmetic mean (unique sites)   : 0.020080
 Arithmetic mean (unique STLDs)   : 0.020080
Arithmetic mean (unique TLDs)   : 0.020080
Geometric  mean (95%CI) of unique URLs   : 0.016255  (0.005730, 0.026891)
Geometric mean (95%CI) of unique domains   : 0.015668 (0.005704, 0.025732)
Geometric mean (95%CI) of unique sites   : 0.014016 (0.005297, 0.022810)
Geometric mean (95%CI) of unique STLDs   : 0.014016 (0.005297, 0.022810)
Geometric mean (95%CI) of unique TLDs   : 0.014016 (0.005297, 0.022810)
Mean  (95%CI) of log (1+unique URLs)   :  0.016125 (0.005714, 0.026536)
Mean (95%CI) of log (1+unique domains)   : 0.015547 (0.005687, 0.025407)
Mean (95%CI) of log (1+unique sites)   : 0.013919 (0.005283, 0.022554)
Mean (95%CI) of log (1+unique STLDs)   : 0.013919 (0.005283, 0.022554)
Mean (95%CI) of log (1+unique TLDs)   : 0.013919 (0.005283, 0.022554)
Proportion  non-zero (95%CI)           : 0.020080  (0.010943, 0.036565)
MNLCS  - mean (95%CI) of world normalised log (1+unique URLs)   [Population version]: 1.000000 (0.354343,  1.645657)
MNLCS  - mean (95%CI) of world normalised log (1_unique URLs)       [Sample version]: 1.000000 (0.321747,  3.108037)
MNLCS - mean (95%CI) of world normalised log  (1+unique domains)   [Population  version]: 1.000000 (0.365825, 1.634175)
MNLCS - mean (95%CI) of world normalised log  (1_unique domains)       [Sample  version]: 1.000000 (0.331823, 3.013657)
MNLCS - mean (95%CI) of world normalised log  (1+unique sites)   [Population version]:  1.000000 (0.379564, 1.620436)
MNLCS - mean (95%CI) of world normalised log  (1_unique sites)       [Sample version]:  1.000000 (0.343900, 2.907818)
MNLCS - mean (95%CI) of world normalised log  (1+unique STLDs)   [Population version]:  1.000000 (0.379564, 1.620436)
MNLCS - mean (95%CI) of world normalised log  (1_unique STLDs)       [Sample version]:  1.000000 (0.343900, 2.907818)
MNLCS - mean (95%CI) of world normalised log  (1+unique TLDs)   [Population version]:  1.000000 (0.379564, 1.620436)
MNLCS - mean (95%CI) of world normalised log  (1_unique TLDs)       [Sample version]:  1.000000 (0.343900, 2.907818)
NPC  - world normalised proportion (95%CI) non-zero [ie risk ratio]: 1.000000  (0.428985, 2.331082)

Group  file: Spain_wiki. In set: Biochemistry Molecular Biology Alcohol 2012
Queries   : 193
Arithmetic  mean (unique URLs)   : 0.020725
Arithmetic mean (unique domains)   : 0.020725
Arithmetic mean (unique sites)   : 0.020725
Arithmetic mean (unique STLDs)   : 0.020725
Arithmetic mean (unique TLDs)   : 0.020725
Geometric mean (95%CI) of unique URLs   : 0.014469  (0.000255, 0.028886)
Geometric mean (95%CI) of unique domains   : 0.014469 (0.000255, 0.028886)
Geometric mean (95%CI) of unique sites   : 0.014469 (0.000255, 0.028886)
Geometric mean (95%CI) of unique STLDs   : 0.014469 (0.000255, 0.028886)
Geometric mean (95%CI) of unique TLDs   : 0.014469 (0.000255, 0.028886)
Mean  (95%CI) of log (1+unique URLs)   :  0.014366 (0.000255, 0.028476)
Mean (95%CI) of log (1+unique domains)   : 0.014366 (0.000255, 0.028476)
Mean (95%CI) of log (1+unique sites)   : 0.014366 (0.000255, 0.028476)
Mean (95%CI) of log (1+unique STLDs)   : 0.014366 (0.000255, 0.028476)
Mean (95%CI) of log (1+unique TLDs)   : 0.014366 (0.000255, 0.028476)
Proportion  non-zero (95%CI)           : 0.020725  (0.008089, 0.052069)
MNLCS - mean (95%CI) of world normalised log (1+unique URLs)   [Population version]: 0.890917 (0.015827,  1.766008)
MNLCS - mean (95%CI) of world normalised log (1+unique URLs)       [Sample version]: 0.890917 (0.015768,  3.039885)
MNLCS - mean (95%CI) of world normalised log  (1+unique domains)   [Population  version]: 0.924021 (0.016415, 1.831627)
MNLCS - mean (95%CI) of world normalised log  (1+unique domains)       [Sample  version]: 0.924021 (0.016356, 3.074937)
MNLCS - mean (95%CI) of world normalised log  (1+unique sites)   [Population version]:  1.032124 (0.018336, 2.045913)
MNLCS - mean (95%CI) of world normalised log  (1+unique sites)       [Sample version]:  1.032124 (0.018272, 3.337906)
MNLCS - mean (95%CI) of world normalised log  (1+unique STLDs)   [Population version]:  1.032124 (0.018336, 2.045913)
MNLCS - mean (95%CI) of world normalised log  (1+unique STLDs)       [Sample version]:  1.032124 (0.018272, 3.337906)
MNLCS - mean (95%CI) of world normalised log  (1+unique TLDs)   [Population version]:  1.032124 (0.018336, 2.045913)
MNLCS - mean (95%CI) of world normalised log  (1+unique TLDs)       [Sample version]:  1.032124 (0.018272, 3.337906)
NPC  - world normalised proportion (95%CI) non-zero [ie risk ratio]: 1.032124  (0.346414, 3.075162)

World  File: Chemistry Alcohol 2012
Queries   : 498
Arithmetic mean (unique URLs)   : 0.002008
Arithmetic mean (unique domains)   : 0.002008
Arithmetic mean (unique sites)   : 0.002008
Arithmetic mean (unique STLDs)   : 0.002008
Arithmetic mean (unique TLDs)   : 0.002008
Geometric  mean (95%CI) of unique URLs   : 0.001393  (-0.001363, 0.004156)
Geometric mean (95%CI) of unique domains   : 0.001393 (-0.001363, 0.004156)
Geometric mean (95%CI) of unique sites   : 0.001393 (-0.001363, 0.004156)
Geometric mean (95%CI) of unique STLDs   : 0.001393 (-0.001363, 0.004156)
Geometric mean (95%CI) of unique TLDs   : 0.001393 (-0.001363, 0.004156)
Mean (95%CI) of log (1+unique URLs)   : 0.001392 (-0.001364, 0.004148)
Mean (95%CI) of log (1+unique domains)   : 0.001392 (-0.001364, 0.004148)
Mean (95%CI) of log (1+unique sites)   : 0.001392 (-0.001364, 0.004148)
Mean (95%CI) of log (1+unique STLDs)   : 0.001392 (-0.001364, 0.004148)
Mean (95%CI) of log (1+unique TLDs)   : 0.001392 (-0.001364, 0.004148)
Proportion  non-zero (95%CI)           : 0.002008  (0.000355, 0.011285)
MNLCS - mean (95%CI) of world normalised log (1+unique URLs)   [Population version]: 1.000000 (-0.980000,  2.980000)
MNLCS - mean (95%CI) of world normalised log (1_unique URLs)       [Sample version]: 1.000000 (NaN, NaN)
MNLCS - mean (95%CI) of world normalised log  (1+unique domains)   [Population  version]: 1.000000 (-0.980000, 2.980000)
MNLCS - mean (95%CI) of world normalised log  (1_unique domains)       [Sample  version]: 1.000000 (NaN, NaN)
MNLCS - mean (95%CI) of world normalised log  (1+unique sites)   [Population version]:  1.000000 (-0.980000, 2.980000)
MNLCS - mean (95%CI) of world normalised log  (1_unique sites)       [Sample version]:  1.000000 (NaN, NaN)
MNLCS - mean (95%CI) of world normalised log  (1+unique STLDs)   [Population version]:  1.000000 (-0.980000, 2.980000)
MNLCS - mean (95%CI) of world normalised log  (1_unique STLDs)       [Sample version]:  1.000000 (NaN, NaN)
MNLCS - mean (95%CI) of world normalised log  (1+unique TLDs)   [Population version]:  1.000000 (-0.980000, 2.980000)
MNLCS - mean (95%CI) of world normalised log  (1_unique TLDs)       [Sample version]:  1.000000 (NaN, NaN)
NPC  - world normalised proportion (95%CI) non-zero [ie risk ratio]: 1.000000  (0.104375, 9.580795)

Group  file: Spain_wiki. In set: Chemistry Alcohol 2012
Queries   : 282
Arithmetic  mean (unique URLs)   : 0.003546
Arithmetic mean (unique domains)   : 0.003546
Arithmetic mean (unique sites)   : 0.003546
Arithmetic mean (unique STLDs)   : 0.003546
Arithmetic mean (unique TLDs)   : 0.003546
Geometric mean (95%CI) of unique URLs   : 0.002461  (-0.002406, 0.007352)
Geometric mean (95%CI) of unique domains   : 0.002461 (-0.002406, 0.007352)
Geometric mean (95%CI) of unique sites   : 0.002461 (-0.002406, 0.007352)
Geometric mean (95%CI) of unique STLDs   : 0.002461 (-0.002406, 0.007352)
Geometric mean (95%CI) of unique TLDs   : 0.002461 (-0.002406, 0.007352)
Mean  (95%CI) of log (1+unique URLs)   :  0.002458 (-0.002409, 0.007325)
Mean (95%CI) of log (1+unique domains)   : 0.002458 (-0.002409, 0.007325)
Mean (95%CI) of log (1+unique sites)   : 0.002458 (-0.002409, 0.007325)
Mean (95%CI) of log (1+unique STLDs)   : 0.002458 (-0.002409, 0.007325)
Mean (95%CI) of log (1+unique TLDs)   : 0.002458 (-0.002409, 0.007325)
Proportion  non-zero (95%CI)           : 0.003546  (0.000626, 0.019810)
MNLCS - mean (95%CI) of world normalised log (1+unique URLs)   [Population version]: 1.765957 (-1.730638,  5.262553)
MNLCS - mean (95%CI) of world normalised log (1+unique URLs)       [Sample version]: 1.765957 (NaN, NaN)
MNLCS - mean (95%CI) of world normalised log  (1+unique domains)   [Population  version]: 1.765957 (-1.730638, 5.262553)
MNLCS - mean (95%CI) of world normalised log  (1+unique domains)       [Sample  version]: 1.765957 (NaN, NaN)
MNLCS - mean (95%CI) of world normalised log  (1+unique sites)   [Population version]:  1.765957 (-1.730638, 5.262553)
MNLCS - mean (95%CI) of world normalised log  (1+unique sites)       [Sample version]:  1.765957 (NaN, NaN)
MNLCS - mean (95%CI) of world normalised log  (1+unique STLDs)   [Population version]:  1.765957 (-1.730638, 5.262553)
MNLCS - mean (95%CI) of world normalised log  (1+unique STLDs)       [Sample version]:  1.765957 (NaN, NaN)
MNLCS - mean (95%CI) of world normalised log  (1+unique TLDs)   [Population version]:  1.765957 (-1.730638, 5.262553)
MNLCS - mean (95%CI) of world normalised log  (1+unique TLDs)       [Sample version]:  1.765957 (NaN, NaN)
NPC - world normalised proportion (95%CI) non-zero [ie risk ratio]: 1.765957  (0.184564, 16.897165)

The bottom part of the Wikipedia citations report is again very similar to the bottom of the citation analysis report. Spain is slightly above the world average for Wikipedia citations but the difference is not statistically significant and statistical significance can only be calculated for the NPC statistic. The first table in the report below has been truncated to get rid of all the results except those for URLs.

The tables below aggregate the results from both world files and both Spanish files in an appropriate way. From the first table, although Spain has an MNLCS average citation count score above the world value of 1, confidence intervals cannot be calculated for it (NaN in the Sample confidence limits). From the third table, Spain has a higher (field equalised) proportion of articles with Wikipedia citations than the world average and a NPC above the world average but in both cases the Spain confidence interval includes the world average value and so the differences are not statistically significant.

The mean Normalised Log-transformed Citation Scores (MNLCS) in the table below are the best to use to compare the group overall with the world average if there are multiple different world averages (e.g., different fields and/or years).
For each group they are the average of ln(1+c) values, divided by the world average ln(1+c) for the file (e.g., field and year).
The world average MNLCS should always be 1.
MNLCS values above 1 indicate that the group average is higher than the world average; MNLCS values below 1 indicate that the group average is lower than the world average

WARNING! MNLCS POPULATION confidence limits below are optimistic because they do not take into account the variability in the world average value.
  - Please use only the MNLCS SAMPLE confidence limits. These are adjusted from the population limits using the weighted average Feiller Expansion calculation.
  - NaN in the Sample confidence limits mean that these are impossible to calculate and are effectively infinite.

========================================================================..
Group N     URLMNLCS    L95Sample   U95Samp     L95Pop      U95Pop      ..
World 996   1           NaN         NaN         -0.0407826  2.0407826   ..
Spain 475   1.410414    NaN         NaN         -0.6944824  3.5153120   ..
========================================================================..
         

Overall proportion non-zero calculations - not recommended because biased against groups with more articles in categories with a high world proportion of non-zero values.

============================================
Group N     Positive    Lower95     Upper95
world 996   0.011044    0.006178    0.019668
Spain 475   0.010526    0.004504    0.024402
============================================
 

Field equalised proportion non-zero and NPC calculations - all group sample sizes are set to the arithmetic mean sample size for sets with at least one publication.

==========================================================================
Group N     PropNonzero Lower95     Upper95     NPC   Lower95     Upper95
world 996   0.011044    0.006178    0.019668    1.000 0.443690    2.253826
Spain 475   0.012136    0.005490    0.026609    1.098 0.417754    2.890315
==========================================================================

Back to the overview page.