NATO’s Propaganda War. A case study on Norwegian propaganda in Libya

The Indicter, Vol 2, Nr 28, 10 Feb 2016 By Jón Karl Stefánsson M.A. and Professor Floyd Webster Rudmin Ph.D. Professor Rudmin is member of The Indicter’s editorial board. Norway is probably not the first country that comes to mind when thinking of biased media and war propaganda. Norway is famous for hosting the Nobel […]

The Indicter, Vol 2, Nr 28, 10 Feb 2016

By Jón Karl Stefánsson M.A. and Professor Floyd Webster Rudmin Ph.D.

Professor Rudmin is member of The Indicter’s editorial board.

Norway is probably not the first country that comes to mind when thinking of biased media and war propaganda. Norway is famous for hosting the Nobel Peace Prize each year and boasts of having the world’s highest standards of living. Norway is a shining example of the Nordic welfare society. Norway is, however, a NATO member state and has participated in numerous military endeavours as directed by NATO’s command. In the Libyan war of 2011, Norway was a very active participant. In fact, the Norwegian Foreign Minister, Jonas Gahr Støre, proudly acclaimed to the Norwegian newspaper, Aftenposten, that the Norwegian Air Force was “best in class” in the bombing campaign in Libya. He boasted that his small country contributed 1/8 of the total number of war planes that dropped bombs on the country (Fordal, 2011).

Whenever a country participates in a war of aggression against another country, we must expect some form of propaganda to find its way into the media. But would the press in the “peace-loving” country of Norway succumb to making misleading news reports? If so, how would we measure it?


norway bombs libyaNorway’s national state media channel, NRK, showing the bombing of Libya targets by Norway’s air force

Recently, the entire news coverage of the Libyan war reported by Norway’s national state media channel, NRK, was analysed for one specific type of biased communication. The Norwegian state television network, NRK, was chosen as the sole source for news items for the analysis because of its availability and its large dissemination in Norway. Each article produced by NRK is archived and freely and readily available. NRK is Norway’s largest media outlet (Fordal, 2009). All households in Norway that have a television set are obliged to pay licence fees to NRK. NRK’s evening news show is viewed by more than 700 thousand people each day (NRK, 2011), giving the outlet a special status as a nationally important news medium.

The study specifically focused on the labels used to describe the main agents of action described in the news reports. It must be noted that this is only one of several areas in which media bias can occur. Thus, the study reported here did not examine omissions, framing effects, or the usage of images to create specific impressions. These effects should also be studied in order to get a full picture of the communication that led to the overwhelmingly public and political support in Norway and in other NATO countries for war against the Libyan government.

The analysis of NRK’s use of evaluative labelling was examined in three consecutive studies. In the first study, every article on the Libyan war, produced by NRK, and available in NRK’s official webpage were subjected to content analysis in order to establish which referent terms were used for the agents described in its coverage of the Libyan war. The examined corpus consisted of 689 news articles. The analysis utilized computer programs in order to minimize human error in coding the 361.879 words comprising the corpus.

The end result of this analysis was a cumulative list of all labels used to describe each key agent in the coverage. A key agent was defined as an agent who is mentioned at least once in every other article in average in the corpus of 689 articles. The analysis found 84 unique labels for 15 key agents.

In the second study, the words found to be used as labels for the main agents of NRK’s coverage of the Libyan war were subjected to a semantic differential test. Here, 316 Norwegian subjects rated each word according to the positive or negative affect they experienced by exposure to each word. The measure ranged from -3.00 (very negative affect) through 0 (neutral affect) to + 3.00 (very positive affect). With these measures, the average affective meaning, or emotional connotations, of each label could be calculated.

In a third study, the results of Study 1 and 2 could be combined in order to calculate the average valence of the referent terms found in the content analysis. This was done in order to answer the question, how did the news report’s labels and descriptions of the key persons and institutions cause readers to have emotional reactions towards those persons and institutions? Details of how the study was performed and more detailed explanations of the findings are omitted in this article, but are available in the Norwegian Munin database here.


Study 1: Content analysis

The content analysis set out to identify which actors were represented most frequently in the news corpus, and as such, represented the key agents of the news topic. This was achieved with means derived from Krippendorff’s (1980) suggestions for systematic analysis of content. The study successfully identified which labels were used to denote each agent, as well as providing additional information regarding the usage of auxiliary descriptive terms for them. The analysis identified 15 main agents in the corpus. Agents are whatever is capable of actions, including included people, groups, governments, fighting forces, and events. The analysis provided material for further studies to examine patterns in the association of evaluative labelling of key agents.

Even though content analysis is a very time-consuming endeavour, as is described in most textbooks on the subject matter, it can be recommended as means of characterizing the meanings of discourse elements in a systematic and quantitative fashion (Kaplan, 1943), as well as collecting the raw material for further studies on the psychology of propaganda. Content analysis offers important advantages in data collection for communication studies in psychology. The units of measures are derived from material that has actually been produced, instead of being concocted as hypothetical realities for experimental manipulations. It provides a replicable methodology that can be applied to a broad range of phenomena. Additionally, there is no need to probe the thoughts or intents of the communicators.


Results of content analysis.

The 15 key agents of the content analysis, and the labels used to denote them are listed in Table 1. Also included are the additional descriptive terms that accompanied these labels.

Click on the link below to view Table 1

Table 1: Key agents, their labels and descriptive terms from NRK’s coverage of the Libyan war


Study 2

The second study assessed whether words that were used to label identified main actors in the news stories of Study 1 differed in valence. This was achieved by subjecting these words to an on-line semantic differential test. By this method, the words found to be used as referent terms for the main agents were administered to 316 Norwegian subjects. The subjects were asked to rate words according to the affect they associated with them, using only the Evaluation dimension of the semantic differential.

In the study, a word’s evaluative affective meaning was calculated as its mean rating on the 7-point Semantic differential scale, with a maximum negative evaluation of-3.00 and a maximum positive evaluation of + 3.00. Thus, the word that has the most negative evaluative affective meaning used as a label in NRK’s corpus was “krigføring” (warfare), which had an average negative affective evaluation of – 2.61, followed by “tyrann”, which had an average negative affective evaluation of -2.60. The word that was most positively evaluated in terms of affective meaning was “frihet” (freedom), which had a mean positive evaluation of + 2.71.

Results of semantic differential measures:

The mean evaluative affective meaning of each studied word, alongside number of appearances and standard deviation, are presented in Table 2. The words are listed in rank order, on the basis of their mean scores for evaluative affective meaning. The emotionally positive words are listed in the left column, and the emotionally negative words are listed in the right column. Thus “frihet” (freedom) had the most positive emotional connotation, and “krigføring” (warfare) had the most negative. “Opprører” (rebels) had neutral emotional connotation for the Norwegian sample evaluating these words.

A different sample of Norwegians might produce different evaluative norms for these words. Similarly, a sample of people from a different nation evaluating the same words in their own language, would produce different evaluative norms for these words.


Table 2: Words used as labels in NRK’s Libya-corpus, alongside their average affective meaning as measured in Study 2, with standard deviation and standard error of the mean.

Approbative   Pejorative
Word Mean_E    SD     SE   Word Mean_E    SD    SE
Frihet 2,71 ,630 0,038 Krigføring -2,61 ,778 0,183
Feire 2,44 ,794 0,048 Tyrann -2,60 ,828 0,051
Juble 2,43 ,859 0,052 Bombe -2,57 ,844 0,051
Lojal 2,21 ,900 0,054 Kreft -2,40 ,968 0,215
Beskytte 2,19 ,956 0,058 Diktatur -2,40 1,062 0,064
Humanitær 2,04 1,070 0,065 Geriljakrig -2,18 1,118 0,067
Mat 2,04 1,155 0,067 Diktator -2,17 1,425 0,336
Bidrag 1,67 1,029 0,243 Hevn -2,16 1,053 0,064
Demokrati 1,67 1,495 0,352 Luftangrep -2,15 1,021 0,062
Innsats 1,63 ,964 0,058 Krigshandling -2,14 1,028 0,062
Oss 1,58 1,311 0,079 Pest -2,06 1,514 0,357
Folket 1,55 1,124 0,068 Militært_angrep -1,94 1,162 0,07
Sivile 1,50 1,098 0,259 Hersker -1,70 1,057 0,064
Frigjøringskamp 1,36 1,163 0,07 militær_operasjon -1,61 1,037 0,244
Deltakelse 1,35 1,086 0,066 Offer -1,40 1,437 0,087
Råd 1,25 1,049 0,063 Konflikt -1,35 ,927 0,056
Befolkning 1,23 1,114 0,067 Hæren -1,28 1,127 0,266
Internasjonal 1,08 1,132 0,068 Militært_oppdrag -1,22 ,943 0,943
Mål 1,03 1,253 0,076 Statskupp -1,20 1,084 0,065
Tro ,95 1,506 0,091 Regime -1,17 1,043 0,065
Legitim ,91 1,044 0,063 Militæraksjon -1,10 1,164 0,07
Allierte ,84 1,264 0,076 Milits -1,08 1,169 0,071
Mannskap ,81 ,996 0,06 Geriljasoldat -1,06 1,731 0,408
Forsvar ,80 1,198 0,072 Makthaverne -,97 1,014 0,061
Representant ,75 ,890 0,054 Despot -,88 1,204 0,073
Verdenssamfunnet ,75 1,207 0,073 Flyforbudssone -,74 1,080 0,065
Alliance ,72 1,227 0,289 Opprør -,68 1,172 0,071
revolusjonsforkjemper ,68 1,152 0,069 Makten -,60 1,007 0,061
Vestlig ,59 1,079 0,065 Sanksjoner -,44 1,066 0,065
Regjering ,59 1,061 0,064 Motstandere -,43 ,978 0,059
Leder ,58 1,138 0,069 Opprørsbevegelse -,42 1,151 0,069
Stat ,56 1,046 0,063 Soldater -,41 1,318 0,08
Regimekritiker ,45 1,191 0,072 Styresmakter -,32 1,048 0,063
Mandate ,42 ,906 0,055 Militæret -,14 1,342 0,081
Oppdrag ,40 ,963 0,058 Dissident -,12 ,749 0,045
Pakt ,39 1,092 0,257 Demonstranter -,12 1,073 0,065
Revolusjon ,37 1,251 0,076 Styrker -,09 1,286 0,078
Sjef ,29 1,059 0,064 Oberst -,08 1,095 0,066
Aktivist ,28 1,102 0,067 Opposisjon -,02 1,052 0,064
Koalisjon ,25 ,855 0,052 Opprører ,00 1,328 0,313
Styre ,22 ,983 0,059
Myndigheter ,18 ,997 0,06
Intervensjon ,16 ,952 0,057
Overgangsråd ,12 ,576 0,035
Aksjon ,09 1,027 0,062
Demonstrasjon ,07 1,112 0,067
Operasjon ,04 1,051 0,063

Study 3

All appearances of the 84 evaluative labels found in Study 1. Thus, in total, 12758 appearances for 15 key agents were analyzed. The average valence of labels for the agents overall was -.102, SD = .94, SE = .00832. The descriptive statistics associated with evaluative labeling across the fifteen key agents, with means, number of appearances of evaluative labels, standard deviations and confidence intervals are reported in Table 3.

Table 3. A Between-group Comparison of Evaluation Scores for Actors in News Coverage of the Libyan Conflict

          95% Confidence Interval
Agent N  Mean  SD  SE Lower bound Upper bound
Gaddafi 595 -0,59 1,39 0,06 -0,70 -0,47
LibyanGvt 1229 -0,41 0,82 0,02 -0,45 -0,36
LibyanMlt 1511 -0,25 0,38 0,01 -0,26 -0,23
Civilians 1069 1,50 0,06 0,00 1,50 1,51
RebelMlt 3186 -0,10 0,26 0,00 -0,11 -0,09
RebelGvt 1226 0,18 0,26 0,01 0,16 0,19
Coalition     530 0,43 0,34 0,01 0,40 0,46
Nato     120 -0,10 0,04 0,00 -0,10 -0,09
UN     237 0,54 0,63 0,04 0,46 0,62
NorwegianGvt      59 0,42 0,20 0,03 0,37 0,47
NorwegianMlt    480 0,13 0,22 0,01 0,11 0,15
USA      32 0,26 0,24 0,04 0,18 0,35
InternationalAttacks  1111 -1,08 0,98 0,03 -1,14 -1,03
NorwegianAttacks    508 0,77 1,12 0,05 0,68 0,87
War    865 -1,40 0,80 0,03 -1,46 -1,35
Total 12758 -0,10 0,94 0,01 -0,12 -0,09

Own names were excluded in the calculation of evaluative affective meaning, as they were not measured in Study 2. Figure 1 shows the key agents of the analysis according to the attitudinal valence, or average evaluative affective meaning. Figure 1 is a graph of the mean attitudinal valence of labels for governments, as they were found in NRK’s corpus.

Figure 1. The mean evaluative affective meaning of the labels for each key agent of NRKs news corpus for the correspondence of the Libyan war of 2011, ordered from the most negatively evaluated agent (the war) to the most positively evaluated agent (civilians).

Fig 1 Setaffanson & Rudmin article

Evaluative labelling of persons and groups:

The agent which was labeled with the most negatively laden terms, as well as having the largest numbers of labels of descriptive terms in general, was Muammar Gaddafi. The ten different terms which were used in the 5231 occasions in which he was mentioned in the corpus had a mean negative affective evaluation of -.59. The range of labels in terms of affective meaning for Gaddafi was from -2.60 (“tyrann”) to 0.58 (“leder”). The agent which was labeled with the most positively-laden terms was the group “civilians”. The average affective meaning of the terms used to label this agent was + 1.50. The second most positively evaluated agent was “FN” (the UN). The terms used to label this agent had an average affective evaluation of + 0.54.

Evaluative labeling of events:

The study also examined coverage of the main events described in the news corpus. It was found that coverage of these events was also biased in terms of evaluative labeling. Comparable events used labels differently depending on the valence that they conveyed. For example, Norway’s’, NATO’s and the coalition’s participation in the Libyan conflict consisted of the same basic acts. However, Norway’s participation was labeled with terms that have positive evaluative connotations (average affective evaluation, +0.87), in stark contrast to what the words actually referred to. This is obvious considering the most negatively labelled event, “the war” itself (average valence -1.40). It is also in contrast to the words used to label the exact same acts committed by the other nations of the “coalition”. Here, the words used to label the bombing raids had average affective evaluations that amounted to -1.08. This wording suggests to the reader that Norway’s bombing attacks were somehow not the same as other bombing attacks.

Evaluative labeling of armies:

The choice of labels was biased in terms of evaluative affective meaning depending on the specific army which was being referred to. Thus, the words used collectively to label the Libyan military had a mean evaluative affective valence ( -.15). The words used to label the rebel army had a mean evaluative valence ( -.43). The labels for NATO’s army had a neutral valence ( -0,09). The labels used for the Norwegian army had a mean evaluative affective valence ( +.33) The “coalition” forces had a mean evaluative affective valence ( + .47). The same concepts were used for each of these agents, but in different frequency depending on the agent.

Evaluative labeling of governments:

The choice of labels for the governments of different parties is not less interesting. The words used to label the government of Libya had the average affective valence of -0,41, while the labels used to describe the rebel government had the average affective valence of +0,15. The US government was labeled with terms that had a positive evaluation of +0,10 and the Norwegian government with +0,13. That the journalists of NRK chose to use negatively laden terms such as “regime” more often for the Libyan government than positively laden terms such as “government”, but vice versa for the rebel government, the US government, or the Norwegian government is a clear example of bias in the usage of evaluative labeling.


Libya - The_Indicter article by J Stefansson & F Rudmin 1 During the invasion of Libya, over a million people, or the overwhelming majority of the population of Tripoli, gathered in downtown Tripoli to support the Jamahiriya system and government. Pictures of such support demonstrations were not reported by NRK. Instead, the headlines read “jubilee in Libya after the fall of Gaddafi”, followed by pictures such as this one here below:[Foto: ISMAIL ZITOUNI / Reuters]

Libya - The_Indicter article by J Stefansson & F Rudmin 2


Discussion of the Three Studies:


The content analysis conducted in Study 1 and the semantic differential measure of the labels found in it, made possible a quantitative comparison of evaluative labelling in NRK’s coverage of the Libyan war. It was found that NRK was overall heavily biased in its labelling of the key agents. Discounting events, the agent who was labelled with the most negatively laden terms was the same agent who was mentioned most often, “Gaddafi”, followed by “the Libyan government” and the “Libyan military”. The “Norwegian military”, “the government of rebels”, “USA”, “Norwegian government” and “the military coalition” that aided various groups of anti-governmental forces taking power in Libya were all awarded with labels that had positive valence overall.

NRK’s reporters could have chosen to use neutral labels, such as the denotative proper names for the agents, or distributed their positive and negative labels evenly across the agents involved in the events. By these means, the events reported would have spoken for themselves, and the readers would have been let to make their own evaluative judgements on the events and agents in the discourse. Instead, as we have seen, the reporters in general added these evaluative labels that signalled how negatively or positively we, the readers should evaluate the agents.

Consider the following illustrative examples. A factual statement with relatively neutral emotional valence might be:

Yesterday, a spokesperson for the rebels alleged that

Libya’s army had killed 12 rebels and 3 other people.

The same information using biased word choices that arouse readers’ emotions against the Libyan government might be:

Yesterday, a reliable source confirmed that Gaddifi’s thugs had

slaughtered 12 pro-democracy freedom fighters and 3 innocent civilians.

The same information using biased word choices that lead readers favour the Libyan government might be:

Yesterday, came the good news that Libya’s national defense forces

had eliminated 12 terrorists and caused little collateral damage.

Though exaggerated in these examples, such manipulations by word choices are happening routinely in news reports, often unnoticed by the public.


Implications of the findings:

The data accumulated from the content analysis and subsequent studies makes it possible to make certain inferences about the expected reactions from the readers of these texts. As has been demonstrated by several theorists (e.g., Osgood,Suci, & Tennenbaum, 1957; Heise, 1980), the choice of labels for focal objects on the basis of the affects they convey can predictably influence how people understand and evaluate news events. Thus, when a highly negatively-laden word is used to label a referent, the reader can be expected to evaluate that referent more negatively than if a more positively evaluated term is used, even when these terms are factually interchangeable. For example, “regime” and “government” are synonyms, but the former has a strong negative affect. When a highly positively-laden term is used instead, the reader can be expected to evaluate that referent more positively. Thus, NRKs discourse on the Libyan crisis is likely to have contributed in creating negative affective evaluations towards certain agents, and positive towards others.

Among the mechanisms for such effects include evaluative conditioning. Here, repeated pairings of a focal object with an emotionally potent stimulus can result in a learned association that comes about through change in valence of evaluative responses towards that focal object (e.g., Staats & Staats, 1958; Gresham & Shimp, 1985). Thus, “when an initially neutral stimulus immediately precedes another stimulus that already has positive or negative associations, the neutral stimulus can come to be positively or negatively evaluated itself” (Petty, Wheeler & Tormala, 2003, p 362). Furthermore, it has been demonstrated that evaluatively biased labels can serve as signals to the receiver that the focal object has either been condemned or been highly valued. The labels can then serve as queues for what opinions towards the focal object are generally accepted, and thus which opinions are safe to communicate, which are correct, and which are safe to hold (e.g., Cialdini 2007).


Of the range of terms that the journalists who authored the corpus could have used to refer to each key agent, they systematically chose negatively laden terms more often for targeted agents who were to be attacked and eventually destroyed, and more positively laden terms for others doing the attacks and the destruction. Since the affective nuances words carry can affect how people react to, and understand, messages that they are embedded in, such choice of labels can be considered propaganda. It is clear that NRK’s reporting was neither fair nor balanced in terms of evaluative labeling.

Other Findings from the Content Analysis

The three studies described here had the objective to quantitatively measure differences in the affective evaluations of the labels used to describe the main agents in NRKs coverage of the Libyan war. However, several other interesting findings appeared in the course of the analysis. These additional findings might be found to appear in biased communication regarding other similar events. NRKs coverage of the Libyan war may therefore serve as a case study for biased coverage of major political and military events in general.

Note that all of the following cases of biased communication are only based on simple word manipulations. Other types of bias have yet to be examined; if accompanying pictures were selected for propagandistic purposes, if statements from the Libyan government were generally dismissed, etc. However, even these simple word biases in themselves tell a propagandistic tale. It told the simplistic tale of innocent people who stood up against a bad person and of the heroic “us” who came to their rescue.

Regime or government?

Bias in the tone of terminology can occur with synonymous terms. For example, the term “regime” is sometimes used instead of synonymous words such as “government” in news reports. However, the usage obviously depends on which government is being referred to. For example, in NRKs news corpus, the word “regime” was used 772 times in total. Of these, it was used to refer to the Libyan government 674 times, or in 87% of the times it appeared. In the remaining 13% of their appearances, the word was mainly used to denote the governments of countries that were not directly linked to the coverage, such as the “Syrian regime”. Of other main agents in the coverage, the word “regime” was used only once to denote the rebel government, after Gaddafi had been sodomized and murdered.

If one looks for the definitions of these terms, one finds that they are in fact denotatively synonymous[1], minding that these words were used interchangeably to refer to the same agent (the Libyan government was also referred to as “regjering”, “myndigheter” etc). However, the term “regime” appears to be used only for governments that are for some reason disliked. To use the term “the Norwegian regime” or “the Stoltenberg regime” does sound different than “the Norwegian government” or “the Stoltenberg government”. A real difference between these words, “regime” and “government”, does not lie in denotative meaning, but in affective meaning. Thus, the word “regime” had a score of -1.17 in terms of the emotional evaluation as measured in Study 2, while the word “regjering” had a positive score of +.59. Here, NRK systematically chose the negatively laden term to describe the Libyan government, and positive labels for other governments. Perhaps we should always be sceptical when we see the word “regime” being used for a government of a country that has been designated as target to be destroyed.

The bad guy

With the exception of the word “Libya”, “Gaddafi” was the most frequently mentioned content word in the corpus of 689 NRK news reports on Libya that were examined. Gaddafi’s name, or one of the terms used to identify him (including “the dictator” (diktatoren) and “the Libyan leader” (den libyske lederen) appeared 6466 times in the corpus, which amounts to 9,38 times on average in each article. Both the Libyan government and military were directly linked to Gaddafi, through descriptive terms such as “The Gaddafi regime” (Gaddafi regimet) and “Gaddafis troops” (Gaddafi styrkene). In fact, 70% of the terms used to refer to the government and 81% of the terms used to denote the Libyan military were connected to Gaddafi.

Soldiers and civilians alike who supported the government were labeled as “Gaddafi loyalists”, as if there was not even a society or governing system outside the man. This demonstrates that the presentation to Norwegians of the crisis and war in Libya was very focused on the person of the national leader. Personification of a nation seems to be one of the main methods of justifying an attack on nation, first, by placing all blame for all events on the personality of the leader, and at the same time side-stepping all issues of national sovereignty. The destruction of Libya was merely for the purpose of removing a criminal person. Similar justification was used for the 1989-1990 invasion of Panama to remove Manuel Noriega, code-named “Operation Just Cause”, and the 2003 invasion of Iraq to remove Saddam Hussein, code-named “Operation Iraqi Freedom.” These were presented, not as wars or as invasions of foreign nations, but as operations to find and punish a bad person.

In itself, this type of personification is relevant to persuasion. Rojo (1995, p. 49) contended that similar personification observed in the coverage of the Gulf war of 1991 to Saddam Hussein had the effect of “establishing an inclusive us and exclusive them, in this case him.” Furthermore,

This move therefore entails a personification of the conflict, which produces immediate feelings of identification or rejection, and simplifies how the war is understood. Both effects are reinforced by the second move of the exclusion procedure: `rejection’. Once the two camps are established, an imaginary dimension, related to the ideology of consensus and ethnic prejudices, is evoked in order to create an image of Saddam Hussein in which he plays the stranger, the irrational being, the madman, the beast and, ultimately, the personification of evil. These are the villain’s attributes in the fairy tale of the just war … which is the script of the event activated by the newspaper for the conceptualization of the conflict and its protagonists. On the other hand, a positive image is created for a unique and ideological us, in which, as readers, we are included and absorb. (Rojo, 1995, p. 49).

Aggression as defense:

The Libyan military was mentioned 1498 times in the corpus, or 2.17 times per article. The military of the Libyan state were most frequently referred to as “forces” (styrker – 699), “soldiers” (525) and “the military” (hæren – 105). Notably, the word “the defense forces” (forsvaret) which appeared 606 times in the corpus, only applied to the Libyan defense force 19 times in total in the corpus, or in 3% of the word’s appearances. In contrast, the Norwegian forces were labelled as the “defense force” (forsvaret) 287 times of the 480 in which it was mentioned in the corpus. This amounts to just under 60% of appearances. This discrepancy gives the reader the impression that the Libyan army was not defending the state, but the invading forces were. The term “defense” was thus turned on its head in reporting on Norway’s attack on Libya.

[1] Thus, the word “government” is defined as “1 the power to govern… 2 the method or system of governing… 3 … the group of people governing a State” (“Government”, Oxford Student Dictionary of Current English, 1992, p. 278). The word “regime” is defined as “1 a method or system of government or administration … 2 a set of rules for diet, exercise etc for improving one’s health and physical well-being” (“Regime”, Oxford Student Dictionary of Current English, 1992, p. 522). Thus, the terms “government” and “regime” are, except when “regime” refers to exercise, denotatively synonymous.


The Norwegian Air Force was a very active agent in the war. It had extensive participation in “Operation Unified Protector”, which was the official name for the NATO-led bombing campaign against Libya. According to official statements, Norwegian fighter jets had 583 missions and dropped 569 bombs on various targets in Libya, a country that has roughly the same population as Norway. However, in the corpus, these bombing raids are labelled with various euphemisms. They were not “bombing raids” or “killings”, but “contribution” (bidrag – 197 times in total), “participation” (deltakelse – 39 times) and “efforts” (innsats – 38 times).

These “contributions” were accompanied by descriptive terms such as “sharp” (13), “military” (9), “our” (6), “complicated” (5), “demanding” (5), “extensive” (3), “humanitarian” (2) and “precise” (2). The terms most often associated with the Libyan military, in contrast, were “abuse” (overgrep – 42 mentions), “were killed” and “rape”. The Libyan military was early on accused of not only raping civilians, but of being supplied with Viagra pills provided by Gaddafi himself. That story was later exposed as false. The actions of the Libyan military certainly were not described as “humanitarian” or “difficult”.

The military attacks on Libya, coordinated by the coalition as a whole, were treated similarly. The words used to refer to this military operation were ”operation”, “intervention”, “action” “effort” (inssatsen), “air attacks” (luftangrepene), “warfare”, “participation” and “no fly zone”, in nominalized form, as well as several compound terms. Thus, when “we” bomb, we contribute. When “they” bomb, they slaughter.

A new euphemism is born:

The term “no-fly zone” (flyforbudssonen) demands special notice. The term appeared 139 times in total. The term stems from the legal origins of the military attacks on Libya. The United Nations Security Council released a resolution (no. 1973-2011) which denied Libya the right to fly airplanes over the country, and simultaneously gave the UN the rights to use “any necessary means” to uphold this resolution. “Flyforbudssonen” was accompanied by the term “uphold” (håndheve) 32 times, and “introduce” (innføre – 22 times). In these 54 instances, the compound terms referred to the “coalitions” bombing campaign against Libya.

The usage of the term in the corpus was not consistent. In some instances, it was used in a manner consistent with what the word “zone” implies, in its geographical sense, as in the sentence “a no-fly zone over the country” (en flyforbudssone over landet). However in certain instances, the word was used differently. Here, the term “flyforbudssone” was used as a direct reference to the coalition’s bombing campaign against Libya: “The no-fly zone over Libya is underway” (Flyforbudssonen over Libya er på plass). Thus, the term was nominalized; morphed into a noun describing a “humanitarian” bombing campaign.

Blacks = foreigners:

Five agents that appeared in the corpus were closely related to the agent identified as the Libyan military. What these agents shared with the military was that they were described as being opposed to the rebel army and NATO. These were, however, separate agents. The frequency of mention of these did not exceed the minimum 0.5 mentions per article and were therefore not included in further analyses. These include “foreign mercenaries” (leiesoldater – 70 mentions). In early stories, it was alleged that Gaddafi himself hired soldiers from other parts of Africa to fight the rebels. It was later revealed that these men were not hired at all. These were black Libyans who strongly supported the government. Black Libyans were especially targeted by the rebels in the war. Also mentioned were “security forces” (sikkerhetsstyrker – 54 mentions), “sharp shooters” (snikskyttere – 53), the police (52 mentions), and the “elite forces”. Gaddafis’ female bodyguards were mentioned mostly in the latter part of the coverage, as they were killed during the fall of Tripoli.

The “rebels” didn’t do it, the “militias” did!

In the news corpus, the opponents of the Libyan government were identifiable through different terms depending on the time frame in which these appeared. In the earliest articles, opponents of the government were frequently referred to as “demonstrators”, “activists” and “opponents” (“demonstranter”, “aktivister” and “motstandere”). These groups largely disappeared from the corpus in early march 2011, when they were replaced by the term “rebels” (opprørere). After August 20, 2011, the rebels had taken over the governance of Libya and were replaced by terms such as “the government (regjeringen) and the “authorities” (myndigheter), but notably not “regime”. At this time, a new term began to appear frequently. These were “militias” (milits). These groups appear to be composed of the same persons as the “rebels”. However, as “militias”, the groups were now reported to be responsible for horrendous acts, such as torture and mass murder of prisoners and bystanders.

The change of word used to denote the rebels in time can be exemplified through the following concordance plots of the news corpus for the terms “demonstrant*”, “opprører*”, “overgangs*” and “milits*”. In the figures below, the earliest appearances of the target word in articles are shown furthest to the left. Each article that the word appeared in is marked by a thin black line.

Fig 2 Setaffanson & Rudmin article

These trends must be understood in terms of the history of the conflict, according to the news corpus. The conflict had four main time periods in the news corpus. The first period, from February 16 to March 17 spanned the time frame between the supposed start of the unrest in Libya to the start of NATO’s bombing campaign against the government of the country. This period involved 64 articles. The second period spanned from March 18, when NATO forces started to aid in the political takeover of Libya by an armed group of rebels, to August 21, when the rebels conquered the capital of the country. This period constituted of 376 articles. The third period was exemplified by a hunt for the country’s alleged leader, Muammar Gaddafi, which resulted in him being brutally killed in October 20 by rebels after his convoy was hit by NATO airplanes. This period included 167 articles. The fourth, and final, period, can be viewed as a period of a new government in Libya and the situation in the country after the ousting and killing of its former political figures. This period lasted from October to February 17, 2012 and included 82 articles. Notably, the term “rebels” (opprører) was replaced with the term “militia”.

Careful reading did not find a factual distinction between the groups “rebels” and “militia”. In other words, the groups “demonstrators”, “rebels” and “militias” appear to be the same physical people, contextually divided by selective labeling. However, there were now numerous accounts of torture, killings and looting committed by these militias. The new label, however, freed the rebels from being associated with the obvious atrocities committed by them. As to how these acts of violence were explained, the word “revenge” (hevn) was used 58 times to describe actions committed by the militias. The word “revenge” in itself implies reaction to prior wrongdoings, thus it does imply that the militias were somewhat right in their ultra-violence.

Legitimate rule of happy, democratic rebels:

Certain terms that appeared relatively frequently in the corpus were almost exclusively associated with actions and motives for the rebel groups. The most notable of these were “freedom” (fri* – 218 mentions), “celebrate” (feire – 143 mentions), “democratic” (demokrati* – 133), “legitimate” (legitime – 96), “cheers” (Jubel – 79), “happiness” (glede -43 mentions), and “rightful” (rettmessig – 25 association). These word associations are in themselves noteworthy. These are in effect the motives and emotions connoted to the rebels in general. They were fighting for “freedom” and “democracy”; their rule was “legitimate” and “rightful”, their emotions were “cheer” and “happiness”. These are in stark contrast to the alleged motives of the government’s forces, which were never described as “legitimate” or fighting for freedom and democracy. Not only did this use of words imply an obvious dichotomy of “good” vs “bad”, it also made a direct statement on international politics: Legitimacy of governance is reserved for who we see as the “good guys”, not by international laws, at least according to NRK.

Freedom and democracy:

Usage of the terms “free” and “democratic” in NRK’s coverage demand special scrutiny. The word fri* had in all 22 forms (“fri”, “frihet”, “frigjort” etc.). It appeared 218 times in total in the corpus of 689 articles from NRK. In the instances in which it was not used to describe the motives of any agent, the word “fri” roughly denotes the act of freeing something from some sort of custody, such as freeing a boat from a harbour, freeing assets from frozen accounts or releasing prisoners. Therefore, in the instances where the word was used to denote motives of agents, mostly the rebels, it can be interpreted as a metaphor, where the ruling system before the rebel victories is likened to a prison. Thus, the word was used 108 times to describe territorial gains made by the rebels, or used as a vague term to describe their motives. Thus, when members of the government army and police had been killed, or forced to surrender or flee, and the rebels had taken power, the town was described as being “frigjort” (freed). The word was never used when the government forces had won a territory back.

In other instances it was used in a vague manner that implies that the old system was in effect a prison. Thus, there were numerous sentences such as: “the rebels have tasted freedom” ([Opprørerne] har smakt frihet) in the later part of the corpus. The term “fritt” was frequently used in the context “Libya er fritt” (Libya is free) in the context of the victory of the rebels and the coalition, after Gaddafi had been murdered. The Libyan army was never described as “freeing” a territory.


The word “demokrati*” (democracy, in all word forms) appeared 133 times in the corpus. In all but 8 appearances the word was used to describe the motives of the rebels or the NATO-led coalition of armies. In 6 occasions the term was used to describe Gadhafi’s ideas of direct democracy, always followed by the specification that Gaddafi was being quoted, and in all instances the quote was followed by a clause that stated that in reality, Gaddafi had all power.

In 14 instances, the rebels were described as fighting, or wishing, for democracy. In 29 occasions Western powers (USA, NATO, Norway and more) were described as working to “bring in” (innføre / “jobbe for en overgang til”) “democracy” in Libya with their military campaigns. Of these, in 11 occasions, the conflict was described as being about the transfer from “dictatorship” (diktatur) to democracy. In summary, the term “democracy” was mostly used in descriptions of motives of both the NATO alliance and the rebels, in 125 of 133 mentions (94%). Democracy and dictatorship were described as the main dichotomy in the conflict, and democracy was ascribed as motives for the rebels or the coalition. There was no attempt made to define democracy, or how the rebels would make or allow democracy.


The “international community”, c’est nous:

One of the most frequently mentioned agents in the corpus was most commonly referred to as “the coalition” (koalisjon). This agent was composed of smaller sub-groups such as “NATO”, “The Arab League”, and the armies of countries such as the USA, Great Britain, Qatar, Norway and France. This compound agent was responsible for military attacks on the Libyan government to support the rebel forces in Libya.

Some countries which were frequently mentioned as being a part of this coalition were sometimes used individually, without mention of this coalition, some international institutions such as the UN Security Council, the International Criminal Court, and others were sometimes depicted as being a part of this coalition, although other mentions of these agents conflicted with this usage. For example, many countries in the United Nations (“FN”) were described as being opposed to the actions of the coalition. Among these countries were all countries of the African Union (AU), India, Russia, China, Germany and all mentioned South-American states. Obviously, these “exceptions” represent the vast majority of the world’s population. This discrepancy in the usage of the agent was especially apparent in other terms that were sometimes used synonymously with this “coalition”, such as “the international community” (det internasjonale samfunnet) or ”the world community” (verdenssamfunnet).

The term “world community” was used in this noteworthy manor throughout the corpus. In some instances, the word was used to refer to the UN, roughly. However, in several instances (14 of 25 appearances for the term “verdenssamfunnet” and 17 of 43 times for the term “internasjonale samfunnet”) this phenomenon was described as committing actions that were in fact committed by the Coalition, such as supporting the rebels with arms and using “all necessary means” to militarily fight the Libyan government. In these instances, this “world society” did not refer to all nations in the United Nations, as might be implied by the direct term. Instead, this term was used to denote specific nations. These nations were in essence the same nation-states as the “coalition” and “alliance”, and can be treated as a synonym to these terms. This was obvious when considering that this world society was described as taking actions against the Libyan government, while most nations in the UN did not participate in, or in fact opposed, the actions of this “world community”. Thus, the entire African Union, most South American and Asian countries, Russia, Germany etc. apparently did not belong to this group of world society, as they were not part of the coalition.

“Civilians” (sivile):

Civilians or Libyan people that did not belong to any armed forces or identified institutions, were mentioned 1039 times in the corpus, or 1,5 mentions in average in each news story. The verb “killed” (drept) appeared 645 times in the corpus. Thus, the term “killed” referred to civilians 496 times, or in 77% of the occasions the word appeared in the corpus. Similarly, the term “protect” (beskytte) appeared 200 times in the corpus. Of these appearances, the term ”protect civilians” (example ”beskytte sivilbefolkningen”) appeared 164 times. Thus, in 82% of the appearances of the term ”protect”, it related to the protection of this group. The agents responsible for the protection of civilians were the same agents that performed the bombing raids on Libya during the civil war: The coalition of NATO and Arab-league countries under the UNs mandate nr. 1973. The Libyan government was never described as protecting people. “Civilians”, however, was a term that was not used to describe every non-military persons in Libya.

Some people deserved it:

Analysis of specific words revealed that not all were equal in being labelled as civilians. Specifically, the victims of different agents were designated different labels. When the Libyan military was reported as killing people, NRKs coverage was rather specific in describing who was targeted. The victims of other agents, such as “the rebels”, “NATO”, “the Coalition” and the “Norwegian military”, were labelled the vague terms “persons” and “people”. This type of bias, which has been described as subject deletion, is a case in which the reader is left to infer as to who were subjected to violence.

For example, after the overthrow of the Libyan government, it was reported that the rebels had imprisoned 8000 or 9000 people (“personer”) in Tripoli, without mentioning if these people were civilians, soldiers or some other group. The same applies to sentences such as “six persons have been killed in a NATO attack” (“seks personer er drept i et Nato-angrep”).

Agent depletion, in which the reader is left to infer who the perpetrator was, also occurred in NRKs coverage. In sentences such as “41 persons were killed in the riots in Libya (“41 personer er drept i opptøyene i Libya”), it was neither specified who had died, nor who had been the perpetrators, supporters or opponents of the government. In 105 appearances, the word “person” appeared in a context that made it impossible to determine which agent was being referred to.

“People” are reliable sources:

In numerous occasions (79), the word “people” was used to refer to unidentified sources for information, through sentences such as “people say that…” (“folk sier at…”) or “ without specifying who these people are, except when they were specified as belonging to the “rebels”.

Government (Gaddafi) supporters are not people:

In 69 occasions, the word “government” was directly used to denote the rebels, rioters or the NTC. In 60 occasions, the word was used to denote the agent “civilians”. The Libyan armed forces were called “Gaddafis people” 15 times in the corpus, further underlining the extent to which the theme story was personified. Similarly, the government of Libya was called “Gaddafi-folk” 6 times. The government was linked with the word “folk” 23 times in total. Civilian supporters of the government were referred to as “people” 6 times in total in the corpus. That particular group, which repeatedly filled the streets of major cities in Libya in support of the Green flag, was hardly ever mentioned in NRKs coverage.

Final Words


These studies and analyses make it clear and certain that NRK news reports showed bias in the selection of words, especially the labels used to describe the key agents. Even in this specific part of the coverage, NRK was found to show serious breaches of the professional and ethical standards expected for journalism, as put forth by organizations such as the International Federation of Journalists (IFJ), by the American Society of Newspaper Editors (ASNE) and by the Norwegian Press Association (Norsk Presseforbund). Amongst these demands are accuracy, fairness, contextuality and absence of bias. (e.g., ASNE, 1975).

Such requirements are hardly controversial. Without accuracy, fairness, contextuality and absence of bias, news reports cannot be considered to convey the information that is needed in order to make informed decisions regarding important topics. If serious decisions, such as the decision to wage war against another country, are based on false, biased or contextually incomplete information and reports, the result can be absolutely catastrophic, as the NATO attack on Libya shows.

Democracy is not possible in a society that does not have high standards of truth in reporting.  At best, biased news is meaningless; at worst, it is a useful tool for those who can control media to create the illusion that the public gives their endeavours its blessings.  If democracy is so highly valued, as suggested by constant lofty praise of democracy and as suggested by our support of war against other nations in order to help them obtain democracy, then surely we must be willing to safeguard democracy against information manipulation.  Routine and systematic analyses of news reports are necessary, as is the education of the public on the very subtle and subconscious methods by which word choices can predictably shape our thinking and emotions about world events.  This is critically important, especially now, when NATO, including Norway, is preparing the public to accept more war in Syria and in Ukraine.



ASNE. (1975). Statement of principles. Retrieved from

Cialdini, R. B. (2007). Influence: The psychology of persuasion. New York: Harper Collins.

Fordal, J. A. (2009). NRKs brede medietilbud. Oslo: NRK. Retrieved from

Gresham, L. G., & Shimp, T. A. (1985). Attitude toward the advertisement and brand attitude: A classical conditioning perspective. Journal of Advertising, 14, 10 – 17.

Heise, D. R. (1970). The Semantic differential and attitude research. In G. F. Summers (Ed.), Attitude measurement (pp. 235-253). Chicago: Rand McNally Staats, A. W., & Staats, C. K. (1958). Attitudes established by classical conditioning. Journal of Abnormal and Social Psychology, 57, 37 – 40.

Krippendorff, K. (1980). Validity in content analysis. In E. Mochmann (Ed.), Computerstrategien für die Kommunikationsanalyse (pp. 60 – 112). Frankfurt: Campus.

Norsk Presseforbund. (2013, 01.07). Vær Varsom-plakaten: Etiske normer for pressen. Retrieved from

NRK. (2011). Nyhetsformidling in endring. Oslo: NRK. Retrieved from

Osgood, C. E., Suci, G. J., & Tennenbaum, P. H. (1957). The measurement of meaning. Urbana: University of Illinois Press.

Per Anders Johansen. “Norge bomber mest i Libya”. Aftenposten, 12.10.2011.

Petty, R. E., Wheeler, C, & Tormala, Z. L. (2003). Persuasion and attitude change. In T. Millon & M. J. Lerner (Eds.), Comprehensive handbook of psychology (2nd ed). New York: John Wiley & Sons.

Rojo, L. M. (1995). Division and rejection: from the personification of the Gulf Conflict to the demonization of Saddam Hussein. Discourse and Society, 6, 49-80.


The authors

4-Jón picJón Karl Stefánsson [photo at left] is from Iceland, where he studies computer science and psychology in theProf Floyd Webster Rudmin - The_Indicter University of Iceland. He received his MA in psychology in the University of Tromso, Norway. He is co-founder of two independent news outlets in Iceland and has since 2003 written several articles independently in mostly Icelandic newspapers and independent news outlets. Additionally he has been active in the establishment of cooperatively run workplaces. His focus is on biased language and how it affects democratic societies.

Floyd Webster Rudmin PhD, (Canada) is emeritus Professor of Social and Community Psychology at UIT Norway’s Arctic University, located in Tromø, Norway. He publishes in prominent online sites such as Common Dreams, Counter Punch, Global Research, and Counter Currents. His writing has been referred to, among other places, in Huffington Post. He has also contributed to The Professors’ Blog. [Photo at right]