Volume 8, Issue 1

(May 2011)

Issue Contents

Past Issues

Audience Research in Extremis: Cold War Broadcasting to the USSR

R. Eugene Parta



How could Western broadcasters during the Cold War learn about their audiences in the USSR when they were denied the possibility of conducting surveys within the country? In response to this quandary, second-best approaches were developed at Radio Liberty employing interviews with travellers outside the country and a sophisticated computer simulation program to draw estimates on audience size, composition and behaviour. In time, it became possible to validate the success of this approach through comparisons with internal survey work, both before and after the breakup of the Soviet Union. This paper overviews the methodology used and a sample of the findings.


Keywords:  Cold War Broadcasting; Audience Research; Radio Liberty; USSR Opinion; Western radio broadcasts; unorthodox survey methodology; Mostellerization.



Cold War broadcasting to the USSR and the countries of the Warsaw Pact has been widely written about and discussed since the fall of the Berlin Wall and the breakup of the Soviet Union. Most of this literature, however, has focused on the broadcasters themselves and not on their audiences [1]. It is the purpose of this paper to shed light on the latter, examining the unique methodology employed to estimate the size and study the behaviour of listeners in the USSR. Arguably, Western broadcasts had an even greater impact on the Warsaw Pact countries of Eastern Europe, but the USSR, the leader of the communist bloc, will be the subject of this article.

Western broadcasters to the Soviet Union faced a unique set of problems in attempting to learn about their audiences, or even confirming that they had significant audiences in the broadcast area.  If Radio Moscow wished to study the audiences to its broadcasts in the West it was free to hire a western survey research firm to conduct a survey.  As this option was closed to Western broadcasters to the Soviet Union, they had to rely on other means and develop new and unorthodox methodologies. At the outset, Radio Liberty, which first broadcast in March 1953, relied on analyzing listener mail.  Mail was actively solicited on-air and this provoked a massive response with thousands of letters arriving despite Soviet censorship and mail interception techniques. (Tests indicated that only about one in ten letters actually succeeded in reaching accommodation addresses in the West).  Although these letters constituted invaluable documentary evidence attesting to the existence of an audience in the USSR, they did not permit projecting estimates of its actual size.

A second method to gauge broadcasting impact was to analyze attacks on Radio Liberty in Soviet media.  These denunciations reached a peak during the years 1966-74 when they appeared in over 1.5 billion copies of Soviet nationally-circulated publications [2]. These media attacks betrayed official concern about Radio Liberty’s impact on the Soviet population (if there were no significant audience it would have made no sense to attack the station and publicize it) but were of little help in learning about the size and composition of the audience.

In order to begin to tackle the problem of gathering data that permitted drawing some general inferences on actual listener patterns, an effort was made to interview Soviet travellers temporarily in the West who intended to return to the Soviet Union. (This distinguished them from defectors or emigrants who would be atypical of the actual listening audience). The first major interviewing effort took place at the Brussels World Fair in 1958 where over 300 Soviet citizens were contacted, of whom 65 turned out to be listeners to Radio Liberty.  Both Voice of America (VOA) and BBC had larger numbers of listeners. This type of ad hoc interviewing of travellers continued through the 1960s but it was not possible to undertake any meaningful statistical analysis and draw inferences about the Soviet population.  This approach did imply, however, that there was considerable listening to Western broadcasts in the USSR. The challenge now was to move the research findings from being ‘interesting’ to being ‘useful and actionable’ in helping to shape effective broadcasting.

As the number of Soviet citizens allowed to travel to the West (almost entirely in supervised groups) steadily increased it became possible to envisage the development of survey procedures and analytical techniques that would permit generalized estimates on the size, composition and listening behaviour of audiences to Western broadcasters to the USSR. From the early 1970s onwards, the audience research department of Radio Liberty began to systematize its data collection methods. Interviews were entrusted to survey research institutes and a standardized questionnaire was developed. By 1990, when it became possible to conduct the first surveys inside the USSR, over 50,000 systematic interviews with travellers had been conducted which provided a unique database for the analysis of the audience to Western broadcasts in the USSR.

This paper will examine: the development of this survey methodology in the field; the computer simulation methodology developed at the Massachusetts Institute of Technology (MIT) for projecting the data back onto the Soviet population to develop broad estimates of audience size; trend listening estimates over time to the major Western broadcasters and generalizations on the composition and behaviour of the audiences; the use of the data by the management and individual language services of Radio Liberty to help shape the broadcasting agenda; the wider sharing of the findings within the larger community of international broadcasters, and finally; validation of the traveller survey project through survey work conducted within the USSR, both secretly during the Cold War period by the USSR Academy of Sciences and openly by local research institutes after the end of the Cold War.

For the sake of simplicity Radio Liberty will be used throughout rather than RFE/RL which became the official name of the organization in 1976 when it was merged with Radio Free Europe. Radio Liberty (RL) became a division of RFE/RL. The station was known by its listeners in the USSR as Radio Liberty throughout. The acronym SAAOR will refer to the RFE/RL Soviet Area Audience and Opinion Research unit based in Paris from 1971, although this name did not become official until 1981.

This study draws heavily on earlier works published by the author, in particular Discovering the Hidden Listener: An Assessment of Radio Liberty and Western Broadcasting During the Cold War published by the Hoover Press at Stanford University in 2007. A more thorough development of the themes in this paper may be found there and in the references listed in the select bibliography.


The Evolution of Systematic Interviewing of Travellers

Soviet travel to the West was highly selective and for the most part strictly supervised. This was especially true in the 1970s. With the onset of the perestroika period in the mid- 1980s it became somewhat more relaxed, although still highly structured. Travel was often organized through the workplace and sometimes given as a reward for good performance.  There is considerable evidence that travellers were screened for loyalty to the regime and Communist party membership in the traveller dataset was considerably higher than in the population as a whole. (approximately 25% of those surveyed were party members compared with about 9% party membership in the adult population).  The traveller population was also biased in favour of males, persons with at least secondary education, those aged between 30 and 50, urban residents and residents of the European Russia, especially Moscow and Leningrad [3]. These structural biases in the traveller dataset were dealt with by means of the MIT computer simulation described below.

Did the media behaviour of Soviet travellers to the West tally with that of their non-travelling compatriots?  Here potential biases might work both ways. On the one hand, travellers might be assumed to come from a more intellectually curious stratum of the population, which takes more interest in foreign events.  Consequently, they might be more active shortwave radio listeners. On the other hand, as noted above, trips might be taken for professional purposes, or granted as a reward for good work, which would broaden the available pool of respondents. Also travellers were screened for political reliability and were disproportionately Communist Party members. This might conceivably lead them to minimize such ideologically dubious activities as foreign radio listening when in conversation with a foreigner.  There is no exact way of knowing to what extent these psychological biases cancelled each other out but we were certainly aware of them. 

The special nature of the project in the tense climate of the Cold War required methods that differ from standard survey research conducted in the West. Soviet travellers were almost always in groups that were supervised by both Soviet and local tour guides. However, travellers were usually given free time for shopping and sightseeing, although even in these instances they tended to travel in groups of threes or more. Consequently, gaining access to individuals to talk freely took some ingenuity and organization. Often this was done with the assistance of local tour guides or in places where travellers were shopping in stores that catered especially to them or in meetings of various types of ‘friendship’ groups.

During the entire 1970s and into the early 1980s Soviet travellers to Western countries were generally intimidated by the sight of a questionnaire in the hands of a Western interviewer and were often unwilling to give candid responses to formally posed questions, although they would often agree to talk to a Westerner in a more relaxed setting. Consequently, no questionnaire was used during the interview in this period but was filled out by the interviewer immediately after the discussion had ended. The interviews took the form of an informal private discussion on the respondents’ media behaviour.  Interviewers were all highly trained and experienced and spoke the language of the respondent. They were instructed to introduce questions on Western radio in the conversation in as structured a way as circumstances permitted and record the responses on the questionnaire immediately after the interview.  It was essential to build rapport with the respondents so that they would be at ease in discussing their media habits, a topic which was generally no longer taboo in the context of détente in the 1970s, although there were still areas of sensitivity. However, interviewers reported that once rapport was established many of the respondents spoke quite candidly and openly, even about political topics.  Some travellers, though, refused to be drawn into discussion. Field reports indicated that the rate of refusals was about 15 per cent although this may well be an under-estimation.

In selecting respondents, interviewers were instructed to interview Soviet citizens from as varied demographic backgrounds as possible, although no formal quota system was applied during this time.  Interviewers were directly employed by independent survey research institutes and the questionnaire was entirely neutral concerning Western broadcasters, i.e. Radio Liberty, BBC and VOA were given identical treatment and the interviewer was unaware of any special interest in the project on the part of any single broadcaster.  The project was designed as a study on media use in the USSR and the role of Western radio in that larger context. During this initial period, most interviews were conducted in various Western European countries.

By the mid-1980s with the advent of perestroika, Soviet travel to the West had increased significantly and the interview situation became more relaxed, permitting an open questionnaire in the interview. The project was accurately explained to respondents as a study of media use in the USSR and interviewers reported that the interview had become significantly easier than hitherto, with respondents often eager to speak at length on their media habits and political developments in the USSR. The questionnaire was later coded for optical character reading and its professional appearance often facilitated the interview process. During this period, approximately 1985-1990, the survey as a result of increased funding had spread beyond the boundaries of Europe into the Asia/Pacific region and south Asia, which improved the geographic distribution of the database. In the late 1980s about. 4,500-5,000 interviews were being conducted a year.

This latter period was characterized by an effort to have the survey work conform as much as possible, given the special circumstances, to standard survey work conducted in Western countries. For example, enhanced quality control measures were introduced with approximately 15-20 per cent of interviews being spot-checked by the employing institute during the interview.  Additionally, a somewhat more formal approach to quota-sampling was undertaken, although the nature of the travelling population still did not permit a rigorous use of this sampling procedure. The MIT computer simulation remained necessary to correct structural biases in the data.

The method of Comparative and Continuing Sampling [4] had originally been developed on the Radio Free Europe (RFE) side of the organization which dealt with interviewing travellers from the Eastern European countries of Poland, Czechoslovakia, Hungary, Bulgaria and Romania. The problems involved in interviewing Eastern European travellers to the West were considerably fewer and more standard interview methods could be applied earlier.

The Comparative and Continuing Sampling methodology was introduced into the Soviet traveller survey in the late 1970s.  This posited that the data from different sampling points was carefully analyzed for internal consistency before being accepted for use in analysis. A chi-square test, to determine statistical consistency of data gathered from different geographic areas, was applied to sub-sets of the data from these areas, and if results were inconsistent the data would be excluded from analysis. The purpose was to ensure that travellers to each sampling point belonged, so to speak, to the same ‘universe,’ and that local differences in interviewing conditions did not impact the findings. Trends over time also had to be consistent over sampling points.


The MIT Computer Simulation

The MIT computer simulation methodology was originally developed under contract from the United States Information Agency to study Soviet audiences for mass communications. The MIT team was headed by Professor Ithiel de Sola Pool and the findings were published in a series of reports in 1975. The input data on western radio listening for this initial study came from SAAOR interviews conducted in 1970-72. Professor Pool gives a detailed account of the methodology used in the first simulation application in the methodological report issued as ‘Simulation Report 4’ in the report series [5].  

The simulation methodology was developed to address the basic issue of how to draw estimates from uneven samples, given that the sample deficiencies could not be corrected in the field. The basic principles were presented in a 1968 article by Professor Frederick Mosteller of Harvard University entitled ‘Association and Estimation in Contingency Tables’ [6]. Professor Pool and his MIT team adapted the procedure to include extensive tests and adjustments for potential internal data inconsistencies and non-convergence to a final statistical solution and dubbed the resulting procedure ‘Mostellerization.’ A more detailed description of the methodology appears in an article that the author of this paper, Prof. Pool and Dr. John Klensin of MIT published in the peer-reviewed international journal Communication Research in 1982 [7].   Space limitations preclude a detailed description here but suffice it to say that the problem can be stated as estimating underlying data from aggregated results, in this case estimating individual cell values in a contingency table (which in some cases might be quite weak or even missing due to sample deficiencies) from the table’s marginals (which would be considerably more robust) employing an iterative procedure until consistent output data are produced.

The first step in the simulation process was to create a structured population model of the USSR based on input from Soviet census data. This population model, derived through the ‘Mostellerization’ process from lower-dimension census tables, was a multi-dimensional table of 240 cells (age-3 categories x gender-2 categories x education-2 categories x rural/urban residence-2 categories x geographic region-10 categories) [8]. Since samples were insufficient to compute listener estimates in all 240 cells of the population model directly, the second stage of the simulation relied on the ‘Mostellerization’ algorithm to derive estimates of underlying cell data from aggregated listenership figures. This procedure first computed listening ratios in a 24-cell table (age x gender x education x rural/urban residence) to derive the basic audience estimate, and then expanded this table through ‘Mostellerization’ to 240 cells to take into account geographic dispersion. The ratings calculated in each cell, multiplied by the population model values for the cell, provided the basic estimate of the audience. In the late 1980s a refined procedure was developed to improve estimates for the geographic regions of the USSR [9].

In the 1970s listening estimates were reported over two year periods with datasets ranging from ca. 2,200 to 2,800 respondents. As annual dataset sizes grew, this reporting period was reduced to one year beginning in 1980. In 1986 the ‘core audience’ concept was introduced. This derived estimates precisely from that part of the sample which was most robust and where most listening to Western radio occurred – the urban, educated, adult population – which made up about a quarter of the overall population. This permitted listening trends to be charted more accurately than using the entire adult population where confidence interval estimates were considerably larger. It also allowed trend estimates to be reported more frequently. In the late 1980s, ‘core audience’ estimates were reported on a quarterly basis, which permitted the radios management to have a more sensitive finger on the pulse of audience shifts in the rapidly evolving media landscape in the perestroika period. 


Generalized findings of the research

Figure 1 below shows the cumulative weekly reach of the major Western broadcasters to the USSR for the period 1980-1990 [10].   Cumulative weekly reach is the percentage of the population reached in the course of an average week. Throughout most of the period Western radio was reaching over 50 million people in the course of an average week.

Figure 1. Weekly Reach of Major Western Broadcasters in the USSR Among the Total Adult Population 16 years and older: 1980-1990


Weekly audience reach figures were more or less steady for the overall Western radio audience, at around 25%, and to the major stations at lesser rates, with VOA having the highest ratings through 1988. Two periods, however, are noteworthy for the fluctuations they show. The first is from 1985 to 1986 following Gorbachev’s election as Chairman of the Communist Party and the beginning of perestroika when Soviet media became livelier and less fettered by official censorship. Listening to Western radio overall dropped during this period as it did for all of the individual broadcasters. It rebounded somewhat in 1987 as Western broadcasters adapted to this new environment and demonstrated their continued relevance and jamming was ended on VOA and BBC. The second major shift took place after jamming ended on Radio Liberty in November 1988. This saw its audience ratings spike upward in 1989 to have the highest ratings among Western broadcasters ahead of VOA, which it had trailed throughout the decade, and BBC, with which it had been more or less equal in terms of weekly reach estimates.

Western radio broadcasts had their greatest appeal to men with a secondary or higher education living in urban areas.  It can be argued that this description also fits those most likely to be interested in politics, as well as those owning radio receivers capable of picking up a signal through jamming.  This profile also corresponds to the type of programs carried by Western radio, which apart from music aimed at youth were designed primarily for the urban intelligentsia. Women were also well represented in the audience, although at lesser rates than men, despite the fact that most of the stations did have special programming for women.Listening rates were highest in the Moscow and Leningrad regions, the Baltic States and the Trans-Caucasus. They were lower in Central Asia, Moldavia and the provincial RSFSR. Communist Party membership, per se, was not a predictor of listening to Western radios. Party members and non-members listened at similar rates [11].

As their primary motivation for tuning in Western radio broadcasts, Soviet respondents cited a desire to hear uncensored news, followed by a need to obtain information unavailable from sources within the USSR (see Figure 2) [12].   A third reason was to learn first-hand about the outside world from non-Soviet sources. Another motive was to verify or disprove information already received from Soviet media. Seeking entertainment was also a motivation for listening but, at 20%, it ranked relatively low in the traveller surveys. This may well underestimate the real interest of Soviet audiences in the music and entertainment programs of Western radios. It is important to note here a difference between this project and a secret internal study conducted by the Institute of Sociology of the USSR Academy of Sciences (ISAN from its Russian acronym) in ca. 1980 which showed very high listening among young people to music broadcasts on Western radio (see below). We hypothesize that this was largely due to the difference in samples. ISAN was able to survey a representative number of young people while SAAOR had to rely on a much smaller group of younger travellers who may have been largely atypical of their age cohort in terms of interest in news and information from the outside world and listening to music on Western stations [13].

Figure 2. Motivations of Soviet Respondents for Listening to Western Radio Broadcasts:1985

Motivations for listening to specific stations generally followed the basic pattern shown here, with some important nuances. Radio Liberty was singled out for its coverage of the USSR and its non-Russian language programming, and was also widely used as a way to verify domestic Soviet information. VOA was often cited for its coverage of the US and its music programs while BBC was widely heralded for its objectivity and professionalism.

Unsurprisingly, personal political orientation was an important predictor of Western radio listening, both in general terms and in choice of individual broadcaster. In 1984, SAAOR published an attitudinal typology of urban Soviet citizens based on over 3,000 interviews with Soviet travellers projected onto the population through the MIT simulation [14].   A factor analysis suggested a scale that broadly segmented the population on a spectrum that I have chosen to label ‘hard-line’ to ‘liberal’ according to their attitudes toward civil liberties in the USSR. ‘Liberals’ supported a more tolerant and open attitude toward civil liberties while the hardliners took an opposite position. This can be broadly understood as indicative of their attitudes toward the Soviet system. Figure 3 gives the breakdown of the urban adult population in terms of these five attitudinal types.

Figure 3. Attitudinal Types in the USSR (Urban Population)


‘Liberals’ and ‘hardliners’ were at roughly equal strength, with one in eight subscribing to one or the other position. ‘Moderates’ and ‘conservatives’ mirrored each other as well, with approximately three in ten in each camp. About one in five could be classified as ‘indifferent’ or ‘neutral’, at the centre of the scale. ‘Liberals’ were significantly more likely to listen to Western radio than any of the other types at 80% listeners. ‘Moderates’ also listened to Western radio at a higher than average rate – 40%.  ‘Hardliners’ tuned in the least, at 10%, and relied on domestic TV as their main source of information [15].

The individual Western stations attracted somewhat different types of audiences in terms of political orientation. Figure 4 shows that half of Radio Liberty’s audience was made up of ‘liberals’ and that another 30% were ‘moderates’ giving it a sharper ideological profile of a ‘proto-democratic’ nature than the other broadcasters. The audiences to VOA, BBC and Deutsche Welle had a plurality of ‘moderates’, who outnumbered the ‘liberals’ two to one in the urban population [16].


Figure 4. Audiences to Western Broadcasters by Attitudinal Type

In terms of information-gathering on current events, Western radio occupied an important place in the Soviet media universe in urban areas (see figure 5).  Although clearly behind the domestic press, TV and radio it was an important source for about 40% of urban residents [17].

Figure 5. Main Sources of Information on Current Events in the USSR: 1978 and 1988 (adult urban population)

The above chart shows the importance of glasnost’ in enhancing the role of Soviet media as an important source of information on current events in 1988 as compared with 1978. While Western radio remains more or less steady at around 40%, the domestic press, which displayed the most freedom of the domestic media sources, jumped considerably. Agitprop meetings, a one-time staple of communist party propaganda, declined by about half as a main source of information in the perestroika environment [18].

In addition to gathering data on general media use and Western radio listening, the interview project was able to place a range of topical opinion questions in the surveys, especially in the 1980s.  This permitted an examination of Western radio listening behaviour in specific circumstances related to concrete issues [19].   Attitudes differ from opinions in that they are more deeply seated and are linked to basic aspects of a country’s political culture while opinions are reactions to short-term events. Underlying attitudes help shape opinions. The typology presented in Figure 3 is an example of an attitudinal study. A dimension underlying both opinions and attitudes is values, an interesting topic which is beyond the scope of this analysis.

Some examples of opinion studies that came from the traveller survey dealt with the fateful Soviet involvement in Afghanistan (1979-1989), the samizdat phenomenon in the USSR, the downing of the Korean airliner off the Siberian coast (1983), the Chernobyl nuclear disaster (1986), glasnost’ and perestroika (1985-1990) and the rise of Solidarity in Poland. Each case examined specific aspects of domestic media and Western radio use.  While it is not possible to examine the details of each study here, suffice it to say that each case demonstrated the measurable role of Western broadcasts in helping to either inform or reinforce positions among listeners that deviated from the official Soviet version of events [20]. Given the total control the Soviet regime exercised over the domestic media, it would be incorrect to say that a true ‘public opinion’ existed in the USSR at the time, but Western broadcasts aided the process of eventually developing one, which began to take shape in the perestroika/glasnost’ period of the late 1980s.


Research and Broadcasting

The ultimate purpose of this research, of course, was to help the broadcasters, their management and their sponsoring organizations to reliably gauge the effectiveness of their broadcasts and to improve them to better serve the audiences.  To this end it was critical to inject the findings of the research into the ‘bloodstream’ of the broadcaster through extensive and timely reporting.  At Radio Liberty this reporting took a variety of forms. Among them were:

  • Trend reports on an annual (and in the late 1980s quarterly) basis provided daily, weekly, monthly and annual listener ratings for all of the major broadcasters. In addition to overall ratings, specific ratings were given for individual demographic categories and geographic regions so that a clear picture of audience trends was available for management decisions. As an unorthodox methodology was being employed, care was taken to caveat all estimates and refer to them as ranges rather than point estimates.
  • Analysis reports examined audience behaviour from various perspectives such as listening times, regularity, span, program preferences overall and among target audience sub-groups, issues of credibility and problem areas, etc. These were important inputs to help the editorial management of individual program services to better understand their listening audiences and improve program output.
  • Broadcast area listener reactions, citing individual quotations from respondents on their reactions to the broadcasts were invaluable in providing direct and uncensored responses from individual listeners.  These quotations were provided in extenso on a monthly basis and were widely read by program staff throughout the broadcast services and helped provide a ‘feel’ for a distant audience which was otherwise lacking.
  • Jamming and Audibility reports.   Monthly reports listing audibility, jamming, and wave-length of listening at specific locations were provided in individual and aggregate formats to the transmission engineers to evaluate the quality of the broadcast signal and enhance counter-jamming strategies..

All of the international broadcasters have an oversight body to which they are responsible. In the case of RFE/RL it was the RFE/RL Board of Directors and the Board for International Broadcasting for the period covered in this study (1970-1990) [21]. The oversight body, in turn, was required by law to report to the US Congress to justify the annual grant for RFE/RL as being in the national interest. The trend listening data, which was provided to the Board on a regular basis was central to this effort and proved on several occasions to be essential in helping to justify the continued existence of RFE/RL. Empirical data, judged to be reliable, showing that the station had a serious audience was crucial to maintaining continued congressional support.

Individual broadcast services and programming executives were by and large most appreciative of the research and analysis and audience feedback, without which they would have been very much operating in the dark; ‘driving at night without headlights’ as it was put by a former president of Radio Liberty, Howland Sargeant [22]. The findings from the data found many practical uses,  such as organizing the program broadcast cycle in terms of audience listening times, determining the proper thematic program mix to meet listeners expectations, choosing the right broadcast frequencies in terms of jamming and audibility conditions, to name just a few. It was also extremely important to broadcaster morale to know that they had a large and appreciative audience and that their programming efforts weren’t in vain.

The above notwithstanding, a few of the broadcast personnel at Radio Liberty who had personal experience of living in the USSR and had suffered under the system were sceptical that it was possible to conduct such detailed research [23]. They doubted that Soviet citizens abroad would freely admit to a Westerner that they listen to Western radio stations, let alone the oft-attacked Radio Liberty. Consequently, they were sometimes reluctant to accept research findings on such issues as ‘broadcast tone’ which ran counter to their personal inclinations. There was a feeling that their personal experience in the broadcast area placed them in a superior position to judge the correctness of their programming decisions than research they felt might be potentially misleading. The issue of broadcast tone is discussed below as one example of how research impacted programming.

With the advent of perestroika in 1985 and with greater flexibility in the Soviet press, Radio Liberty’s broadcast tone was judged to be more critical than friendly by respondents in the survey data (see Figure 6) [24].   The way Radio Liberty’s listeners assessed its tone was probably influenced by the station’s forthright engagement on behalf of human rights and some initial scepticism about perestroika and glasnost.  This was pointed out to the services and broadcast management on the basis of the research and while there were questions on the part of the services they were enjoined by management to be more sensitive to the issue. This appears to have paid off by 1987 when a negative tone score moved into positive territory.  Although Radio Liberty was still considered to be more critical in tone than other Western broadcasters, probably an inevitability given its more sharp-edged coverage of domestic Soviet affairs, it had undergone a significant shift that was directly inspired by the research. 

Figure 6. Trends in Perceived Broadcast ‘Tone’ of International Broadcasters: 1985-1987

Without empirical research the issue of broadcast tone may not have been raised in the first place, and if it had been raised there would have been no objective way to resolve the issue except to choose one set of opinions over another.


Sharing the research with the wider broadcast community

From 1970 onward, RL research was conceived as an objective study of Western broadcasting to the USSR and not as a compilation of interesting anecdotes about Radio Liberty. Consequently, it was only natural to begin to share the findings with other international broadcasters who had no other systematic empirical data sources on their audiences in the USSR. This initial sharing developed into a more active cooperation as the years progressed, especially between Radio Liberty, VOA and BBC.  In addition, to receiving trend listening estimates, both stations subscribed on a monthly basis to the Broadcast Area Listening Evidence report series which provided individual quotes from listeners on the broadcasts. In time, other international broadcasters made use of the data, including Deutsche Welle, Radio Sweden, Radio Canada International, Radio France Internationale, Kol Israel and Vatican Radio. What had started in the 1960s as a project to collect anecdotal evidence on listening to Radio Liberty had developed into a project that serviced all of the major western broadcasters to the USSR during the Cold War.


Validating the Findings

Given the unorthodox methodology, it was of the highest importance to find ways to validate the findings. Although great care was taken to ensure the quality of the survey process and to monitor the internal consistency of the data gathered, this in itself did not guarantee that the results reflected the actual media situation in the USSR. Consequently, a number of methods were used to attempt to gauge the validity of the findings, both during the time of the project during the Cold War and after the break-up of the Soviet Union. Opportunities during the Cold War were limited given that we were unable to conduct surveys within the USSR, but from 1990 on this became possible and the results were largely consistent with earlier SAAOR audience projections [25].

In addition to traveller surveys, SAAOR systematically interviewed Jewish emigrants from the USSR as a separate project (1972-1990). This data was not combined with the traveller data to derive audience estimates, but it provided extensive detailed information on radio listening and programming issues. Although allowed to emigrate because they were Jewish, emigrants were broadly typical of non-emigrants in the same educational, age and geographic categories. Consequently, they were a useful surrogate group to study listening behaviour in greater detail.  Although the emigrants listened to Western radio at twice the rate of travellers, their overall patterns of listening were very similar when the listeners in each dataset were directly compared (see Figure 7) [26].

Figure 7:  Stations Heard Among Western Radio Listeners: A Comparison of USSR Traveller and Emigrant Data: 1990

This similarity in both rank order of stations heard and comparable listening rates was an important validation of the internal consistency of the two datasets.

The first real opportunity to compare SAAOR data on Soviet media behaviour with data from an internal study came in 1975. Data from the traveller survey on domestic TV viewing was compared with that published in a Soviet study conducted in Leningrad and showed similar findings in terms of average hours of TV viewed per week by age and education level and types of programs viewed [27]. This similarity of findings between the two highly different datasets helped increase our confidence in the reliability of SAAOR data in an early period of the project. It was also possible to compare SAAOR findings with several ‘unofficial’ polls conducted within the USSR by Soviet social scientists, presumably of a ‘dissident’ political bent. Unofficial polls conducted in Moscow on attitudes toward Nobel Peace Prize winner Andrei Sakharov and on attitudes toward the Polish Solidarity movement were almost identical with SAAOR Soviet traveller surveys [28]. Of course, unorthodox methodologies cannot scientifically validate each other but the similarity of findings strongly suggested that a reasonably accurate picture of attitudes toward Solidarity and toward Sakharov was being sketched in the SAAOR data.

In the later perestroika period (1988-1989), a few western media organizations managed to include their own questions in internal Soviet opinion polls which were just becoming public. The French news weekly Le Point, Time magazine and CBS-New York Times placed questions on surveys conducted by the Moscow-based Institute of Sociological Research. The Le Point-sponsored studies were consistent with SAAOR findings on perestroika, withdrawal from Afghanistan, emigration from the USSR, and awareness of Andrei Sakharov. Time magazine’s poll with the Institute of Sociology at the USSR Academy of Sciences on attitudes toward glasnost’ and Gorbachev showed almost identical results with SAAOR survey data. The results of the CBS-New York Times poll in 1988 among Muscovites on perestroika and the Soviet-Afghan war were also nearly identical with SAAOR findings in both aggregate and age categories [29].

Meanwhile, in 1990, it was possible for the first time to directly compare data on Western radio listening from the SAAOR traveller survey with surveys conducted inside the USSR [30]. Although the comparison was not exact in terms of question wording and sample design (SAAOR survey data was entire USSR, ISAN sampled three large urban centres, VCIOM sampled the entire USSR, VOX Populi Sampled the RSFSR), the results proved highly instructive and broadly confirmatory (see Figure 8) [31].

Figure 8. Comparative Listening to Western Radio in External and Internal Surveys: SAAOR, Institute of Sociology of the USSR Academy of Sciences and VCIOM/Vox Populi: Annual Reach 1990.

* ISAN survey: Moscow, Leningrad, Kiev, Gus-Khrustalny
**VCIOM Survey: All-Union for Annual Reach: Vox Populi Survey RSFSR only.

Both internal surveys showed slightly higher rates of listening to Western radio than did the SAAOR data, although the results are within the same approximate range. Although annual listening rates for Radio Liberty are somewhat higher in the SAAOR survey, both internal surveys show high rates of listening to Radio Liberty as well. In sum, all three surveys present a relatively consistent picture of listening to Western radio and Radio Liberty in the USSR in 1990.

Surveys in other areas showed broadly similar results between SAAOR data for 1989-early 1990 and internal studies conducted in 1990-1991. It should be noted that listening to Western radio in general declined from the first to the second of these two time periods and that might explain the slight discrepancies in the findings. An internal survey conducted in Leningrad in 1991 showed 20% listening to Radio Liberty, compared to 24% for Radio Liberty in SAAOR 1989 data [32]. In the Baltic States internal surveys conducted in 1990 showed 30% weekly reach for Western radio and 14% for RFE/RL. Comparable SAAOR estimates were 29% for Western radio and 18% for RFE/RL [33].
In all areas where comparisons could be made between internal studies in 1990-91 and SAAOR traveller survey for 1989-early 1991, the estimates were close and the trends identical.

While audiences to Western radio declined in the early years following the breakup of the Soviet Union at the end of 1991, annual surveys on media behaviour measured whether respondents had ever listened to Western radio, even though they might not currently be listeners.  The ‘lapsed’ listener figure was high and together with current listeners always more than encompassed the total listener figure to Western radio derived from SAAOR traveller surveys in the 1980s. 

At the October 2004 conference on Cold War Broadcasting held at the Hoover Institution of Stanford University, Dr. Elena Bashkirova, a former sociologist at the USSR Academy of Sciences presented findings from research carried out by the Institute of Sociology in the late 1970s and early 1980s on listening to Western radio in the USSR [34]. Although the methodology was obviously different from that employed by SAAOR, the Soviet research confirmed SAAOR findings on large audiences to Western radio broadcasts and permitted comparisons in a number of areas. The internal USSR study was based on over 6,000 respondent cases in 6 major cities that were considered by the Soviet researchers to be approximately representative of the urban areas of the Soviet Union. Overall weekly listening rates in the Soviet study and SAAOR data for the period are very close (see Figure 9) with the ISAN rate of 27% only slightly higher than the SAAOR rates [35].

Figure 9. Comparative Weekly Listening Rates to Western Radio: SAAOR and Institute of Sociology, USSR Academy of Sciences 1979-1981

The ISAN study also noted a high rate of occasional listening (less than once a week) to Western broadcasts, i.e. 35% of the urban population. Combined, these two figures showed that ‘by the end of the 1970s more than half of the USSR urban population listened to foreign broadcasting more or less regularly [36].’ This overall reach estimate of 62% is considerably higher than the annual reach of 34% tallied by SAAOR in 1980. One of the reasons that the ISAN rates are higher may be due to the fact that they are based on an urban sample, while the SAAOR rates are based on simulated national samples. Another is likely due to the large number of ‘accidental’ listeners and young people listening for entertainment who were much less likely to be captured in the SAAOR traveller survey.

The ISAN survey was primarily aimed at studying the audience to Western radio in general in terms of its behaviour, motivations and social structure and didn’t compute weekly and overall reach figures for individual stations. Consequently, no comparisons between the two datasets can be made at that level.  ISAN did, however, ask a general question on what stations people heard and here the rank ordering of stations is the same between ISAN and SAAOR with VOA in first place, followed by BBC in second with the heavily jammed Radio Liberty and Deutsche Welle lower. Demographic trends in the two datasets were consistent with the exception of the young group noted above due to sample differences.  Both datasets also showed that large numbers of Soviet citizens sought Western radio broadcasts as alternative sources of information to official Soviet sources. Additionally, trust in the information broadcast on Western radio was high in both the ISAN and SAAOR data. The consistency of these two studies for the same pre-perestroika period is noteworthy.



The unorthodox methodology of interviewing travellers from the USSR to the West and drawing estimates of listening audiences through computer simulation provided Western broadcasters with their only empirical feedback on the size, composition and listening behaviour of their audiences in the Soviet Union during the Cold War.

The audience estimates, based on over 50,000 systematic interviews with travellers (supplemented by over 25,000 interviews with emigrants from the USSR), showed consistency from year to year and fluctuations over time could be correlated to external events, e.g. changes in jamming, political developments, etc.

Validation exercises both before and after the break-up of the USSR showed that comparisons between internal studies and the traveller surveys were close in results and consistent in trends. This congruence of findings, while not scientific proof of the validity of the traveller surveys, is compelling evidence that the SAAOR methodological approach yielded reasonable and useful findings for broadcasters on Western radio audiences, general media use and public attitudes in the USSR.

Despite some scepticism on the part of individual broadcasters, the research was widely used by management and the broadcast services in gauging the effectiveness of the broadcasts, determining the program agenda and topical mix, adjusting programming approaches in terms of content, tone and broadcast style and providing essential reception data to assist in anti-jamming efforts. The research feedback was the main link that programmers had with their distant audiences.  The research was critical for the oversight agencies in their responsibility of determining the effectiveness of the broadcasting effort and eventual allocation of resources.


Select References

The following materials are available in the Audience Research Section of the RFE/RL archive at the Hoover Institution/Stanford University. 

Dawson, Ree, Massachusetts Institute of Technology. ‘Developing a Methodology for Projecting the Audience to Foreign Broadcasts in the Soviet Union.’ Summer 1986.

Dawson, Ree, Massachusetts Institute of Technology. ‘Statistical Treatment of SAAOR Data: A Comparison of Traditional Statistical Methods and Current Audience Simulation Techniques.’ Summer-Fall 1987.

Dawson, Ree, Massachusetts Institute of Technology. ‘Estimates of Geographic Audiences: Imputing Sample Values for SAAOR Interview Data.’ Summer-Fall 1989.

Dawson Ree, Harvard University. ‘Quality-Adjusted Audience Estimates: Russian RSFSR Listening to Foreign Broadcasts 1991-1992. August 1993.

de Sola Pool, Ithiel. ‘Soviet Audiences for Foreign Radio.’ USIA R-17-76. With Summary prepared by the Office of Research, United States Information Agency, Massachusetts Institute of Technology,. September 1976.

de Sola Pool, Ithiel. Director. MIT Communications Research Program. September 1975

  • Preface to Simulation Report Series
  • Simulation Report #1. The Soviet Audience for Foreign Broadcasts
  • Simulation Report #2. The Soviet Audience for Foreign Broadcasts in Minority Regions and Languages
  • Simulation Report #3. The Soviet Audience for Domestic Media
  • Simulation Report #4. Methodology
  • Simulation Report #5. Trends and Variations in Soviet Audiences.

Parta, R. Eugene. .’Civil Liberties and the Soviet Citizen: Attitudinal Types and Western Radio Listening,’ AR 6-84,  Soviet Area Audience and Opinion Research, RFE/RL, Inc.



Abshire, David. International Broadcasting: A New Dimension of Western Diplomacy. The Washington Papers,  Center for Strategic and International Studies. SAGE,  1976.

Alexeyeva, Ludmilla. U.S. Broadcasting to the Soviet Union. Human Rights Watch, September 1986.

Critchlow, James. Radio Liberty: Hole-in-the-Head. American University Press, September 1995.

Cummings, Richard H. Cold War Radio: The Dangerous History Of American Broadcasting In Europe 1950-1989. McFarland & Co., 2009

Hale, Julian. Radio Power: Propaganda and International Broadcasting. Paul Elek, London, 1975.

Holt, Robert T.  Radio Free Europe. University of Minnesota Press,  1958.

Johnson, A. Ross. Radio Free Europe and Radio Liberty: the CIA Years and Beyond, Woodrow Wilson Center Press and Stanford University Press, 2010.

Johnson, A. Ross and Parta, R. Eugene, eds. Cold War Broadcasting: Impact on the Soviet Union and Eastern Europe, Central European University Press, Budapest, 2010.

Kesselman, Leonid. ‘St. Petersburg Audience of Radio Liberty: Through Communication Towards Self-Responsibility,’ in Oleg Manaev & Yuri Prylink, Media in Transition: From Totalitarianism to Democracy.  Minsk, 1993.

Lisann, Maury. Broadcasting to the Soviet Union: International Politics and Radio. Praeger, New York, 1975.

Mickelson, Sig. America’s Other Voice: The Story of Radio Free Europe & Radio Liberty. Praeger,  October 1983.

Mosteller, Frederick. Harvard University. ‘Association and Estimation in Contingency Tables.’  Journal of the American Statistical Association,  March 1968.

Nelson, Michael. War of the Black Heavens: The Battles of Western Broadcasting in the Cold War. Syracuse University Press,  October 1997.

Parta, R. Eugene,  John C. Klensin and Ithiel de Sola Pool: ‘The Shortwave Audience in the USSR: Methods for Improving the Estimates.’ Communications Research: An International Quarterly, Vol. 9, No. 4, October 1982.

Parta, R. Eugene. Discovering the Hidden Listener: An Assessment of Radio Liberty and Western Broadcasting to the USSR During the Cold War. Hoover Press/Stanford, 2007.

Parta, R. Eugene. ‘Soviet Area Audience and Opinion Research at Radio Free Europe/Radio Liberty’ in Kenneth R.M. Short, Western Broadcasting Over the Iron Curtain. Croom Helm, London,  1986.

Picaper, Jean-Paul. Le Pont Invisible: Ces Radios que l’Est Veut Reduire au Silence. Librairie Plon, Paris, 1986.

Presidential Study Commission on International Radio Broadcasting. The Right to Know. U.S. Government Printing Office, Washington, DC, 1973.

Puddington, Arch. Broadcasting Freedom: The Cold War Triumph of Radio Free Europe and Radio Liberty. University of Kentucky,  May 2000

Sosin, Gene. Sparks of Liberty: An Insider’s Memoir of Radio Liberty. Pennsylvania Universty Press, April 1999

Urban, George R.  Radio Free Europe & the Pursuit of Democracy: My War within the Cold War. Yale University Press, January 1998.

 Tyson, James L.
U.S. International Broadcasting and National Security.  National Strategy Information Center, Ramapo Press, 1983.


[1] See A. Ross Johnson,  Radio Free Europe and Radio Liberty: The CIA Years and Beyond for background on the formation of Radio Liberty, Arch Puddington: Broadcasting Freedom: The Cold War Triumph of Radio Free Europe and Radio Liberty for a broad historical development and Gene Sosin, Sparks of Liberty: An Insider’s Memoir of Radio Liberty for an overview of programming.

[2] See R. Eugene Parta. ‘Soviet Area Audience and Opinion Research at Radio Free Europe/Radio Liberty’ in Kenneth R.M. Short, ed. Western Broadcasting Over the Iron Curtain. Croom Helm, London,  1986

[3] Statistical data on the demographic makeup of the surveys are included in the annual Listening Trend Reports, which are on file in the audience research section of the RFE/RL archive at the Hoover Institution, Stanford University. Figures given here are generalizations.

[4] See ‘The Method of Comparative and Continuing Sampling,’ Audience and Public Opinion Research Department, Radio Free Europe, Munich, January 1976. RFE/RL Archives at Hoover Institution/Stanford University.

[5] See Ithiel de Sola Pool, Communist Communications. MIT Communications Research Program, Report #4, Methodology, Cambridge, 1975. RFE/RL Archives at Hoover Institution/Stanford University.

[6] See Frederick Mosteller. ‘Association and Estimation in Contingency Tables.’ Journal of the American Statistical Association. 63:1977.

[7] See Parta, Klensin, Pool: ‘The Shortwave Audience in the USSR: Methods for improving the Estimates.’ Communications Research: An International Quarterly, Vol. 9, No. 4, October 1982.

[8] Parta, Klensin, Pool. pp  592-594.  .

[9] See Parta and Dawson, ‘Revised Geographic Listening Estimates to Foreign Radio in the USSR: Introduction of Log-liner Imputation Techniques for Geographic Estimates,’ AR 2-90, June 1990. SAAOR, RFE/RL, Inc. RFE/RL Archives at Hoover Institution/Stanford University.

[10] See  R. Eugene Parta, Discovering the Hidden Listener: An Assessment of Radio Liberty and Western Broadcasting to the USSR During the Cold War. Hoover Press, Stanford University, 2007, p. 7. The annual estimates are found in the Listener Trend Report series of SAAOR publications which are on file in the RFE/RL archive at the Hoover Institution/Stanford University.

[11] Parta, Discovering the Hidden Listener, pp.  27-30

[12] See Mark Rhodes, ‘Perceptions of Western Radio: How Soviet Citizens View Radio Liberty, VOA, BBC and Deutsche Welle’ AR 3-85, SAAOR, RFE/RL, Inc. RFE/RL Archives at Hoover Institution/Stanford University.

[13] See Rhodes, ‘Perceptions…’., pp. 32-34

[14] R. Eugene Parta, ‘Civil Liberties and the Soviet Citizen: Attitudinal Types and Western Radio Listening,’ AR 6-84,  SAAOR, RFE/RL, Inc. RFE/RL Archives at Hoover Institution/Stanford University.

[15] Parta, Discovering the Hidden Listener, p. 31.

[16] Parta, Discovering the Hidden Listener, p. 32.

[17] See R. Eugene Parta and Mark Rhodes, ‘Information Sources and the Soviet Citizen: Domestic Media and Western Radio.’ AR 5-81. Soviet Area Audience and Opinion Research, RFE/RL, Inc. RFE/RL Archives at Hoover Institution/Stanford University.

[18] Parta, Discovering the Hidden Listener, p. 46.

[19] Opinion research also had a direct impact on broadcasting. In the early years of the Soviet-Afghan war, RL was criticized by many of its listeners, including those opposed to the Soviet invasion, for a perceived tone of ‘schadenfreude’ in some of its broadcasts on the war and for what was seen by some listeners as insufficient empathy for the loss of young Soviet soldiers. This was called to the attention of the broadcast service through top management, programming adjustments were made and subsequent research showed the broadcasts became considerably more effective as a result.

[20] For further details on these case studies see Parta, Discovering the Hidden Listener, pp.  47-59

[21] Since 1994 the oversight body is the Broadcasting Board of Governors, a US federal agency with the eight Governors appointed by the President of the United States on a bi-partisan basis.

[22] Heard by the author in 1965.

[23]   Dr. Vladimir Shlapentokh was a leading Soviet sociologist at the USSR Academy of Sciences who emigrated to the United States in 1979 and became a professor of sociology at Michigan State University. During a BBC visit in 1987 he learned that BBC’s audience research on the USSR came from RFE/RL’s SAAOR unit in Paris and was based on systematic interviewing of travellers from the USSR. Curious, and somewhat sceptical, he was encouraged by the BBC to visit Paris to see for himself. He spent an entire day with SAAOR, getting to know the operation first-hand.  At the end of the day he remarked: ‘This has been a most interesting day for me!’   I queried:  ‘Because our data shows that there is a large audience for Western broadcasts in the USSR?’ Shlapentokh:  ‘No, we’ve known that for a long time. What strikes me is that these travellers are willing to talk to you about it. That means they’re no longer afraid. And that will be very important for future developments in the Soviet Union.’ This was a most prescient observation and an example of how unintended findings are at times as significant as the original research aims.

[24] See report AR 4-87, ‘Comparative Audience Perceptions of Major Western Broadcasters to the USSR: January 1985-June 1987’. Soviet Area Audience and Opinion Research, RFE/RL, Inc. RFE/RL Archives at Hoover Institution/Stanford University.

[25] For a more extensive treatment of these methods see Parta, Discovering the Hidden Listener, pp. 83-93

[26] Parta, Discovering the Hidden Listener, pp. 83-84

[27] See Parta, ‘Listening to Western Radio and Viewing Television in the USSR.’ AR 2-75. SAAOR, RFE/RL, Inc. 1975.  RFE/RL Archives at Hoover Institution/Stanford University.

[28] See Rhodes and Parta, ‘Attitudes of Some Soviet Citizens to Andrei Sakharov: Comparison of SAAOR Data with Unofficial Soviet Poll.’  AR 11-81, SAAOR, RFE/RL, 1981. RFE/RL Archives at Hoover Institution/Stanford University.

[29] For more information on these polls and comparisons with SAAOR data see:  Rhodes, ‘The Recent Joint Soviet-French Opinion Polls and SAAOR Data,’ AR 9-87; Rhodes, ‘Comparison of SAAOR Data with CBS/New York Times Poll,’ Internal SAAOR memorandum, November 15, 1988; Sallie Wise, ‘Comparisons of SAAOR Data and Time Poll,’ SAAOR memorandum, April 14, 1989. RFE/RL Archives at Hoover Institution/Stanford University.

[30] The surveys were acquired by SAAOR in 1990. RFE/RL Archives at Hoover Institution/Stanford University.

[31] See Parta, Discovering the Hidden Listener, p. 91.

[32] see Leonid Kesselman. ‘St. Petersburg Audience of Radio Liberty: Through Communication Towards Self-Responsibility,’ in Media in Transition: From Totalitarianism to Democracy.  Minsk, 1993, p. 264.

[33] Parta, ‘Comparative Baltic Survey Figures.,’  Internal RFE/RL memo from Parta to William W. Marsh, June 25, 1990.  Archived at Hoover Institution/Stanford University.

[34] See Elena Bashkirova, ‘The Foreign Radio Audience in the USSR During the Cold War: An Internal Perspective,’  chapter 6 in Johnson and Parta, eds. Cold War Broadcasting: Impact on the Soviet Union and Eastern Europe. Central European University Press, Budapest, 2010.

[35] 32 Bashkirova., p. 111 for the ISAN survey. SAAOR results are taken from annual Trend Listening Reports for 1979, 1980 and 1981 archived at the Hoover Institution/Stanford University.

[36] Bashkirova, p. 109.


Biographical Note

Russell Eugene (Gene) Parta is the former director of Soviet Area Audience and Opinion Research at RFE/RL and of Media and Opinion Research at the RFE/RL Research Institute.  He is currently a consultant on research issues in international broadcasting.


Last updated 13 July 2011