IMPACT SURVEY OF VOTER
KNOWLEDGE AND AWARENESS
Introduction
One specific aim of this Impact Survey was to test whether any changes in awareness and perceptions had occurred among voters since the National Elections of 1998. This test involved a comparison with the findings of our Baseline Survey conducted before the election[1] and included questions on the quality of participation in the elections and the mechanisms of vote-getting.
Another aim of this Impact Survey was to assess the effectiveness of the voter education campaign in reaching the electorate with apt, effective messages. This assessment included questions referring to voter education broadcasts and publications produced by Cambodian government and non-government organizations, with the help of international donors.
A third aim of the Impact Survey was to anticipate some of the challenges that the voter education community may face during the process of creating institutions of good governance at the local level. This involved questions regarding the constituencies of the new elected bodies that will be formed in the forthcoming commune level elections scheduled for late 2000. There are likely to be fewer donor funds available for commune elections than were available for the 1998 National Elections, so the voter education community will need information of the kind presented in this study to make their efforts most effective.
In this Impact Survey we did not test for changes in attitude toward the participation of women, the poor and the illiterate as voters in elections, since the Baseline Survey showed overwhelmingly positive support for the voting rights of these citizens, which is unlikely to decrease.[2] We did, however, pose questions in the Impact Survey about the ability of these categories of citizens to serve in elected commune office.
We did retest voter attitudes toward the voting rights of ethnic minorities. The Impact Survey aimed to determine if the experience of the voter education campaign had any impact on the very large percentage of “don’t know” responses to questions about the rights of these categories of people in the Baseline Survey.
In the Impact Survey we also compared pre-election expectations and actual experience in regard to difficulties and influence on the electorate associated with registration or voting. We also considered voter knowledge of their elected officials, following up on questions in the Baseline Survey.
Early interim results of this Impact Survey were made available to Cambodian voter education and development NGOs and to election monitoring organizations in late 1999 and early 2000. We were able to bring our Survey findings to the civil society groups that were formed in the Fall of 1999 to discuss the drafts of the Commune Council Administration law that were currently being deliberated in the Council of Ministers. The results of our research were able to inform the efforts of these civil society organizations at a crucial time as they attempted to consult with government officials regarding the process of decentralization reflected in the formation and function of elected commune councils.
In accord with the original design of the project, an important workshop was held in February 2000 at which the results of our Impact Survey were presented and discussed in Khmer (with simultaneous translation in English). The workshop was held jointly with COFFEL, COMFREL and NICFEC who also presented their findings on the specific topic of voter preferences for election mechanisms to form commune councils. The event attracted a number of participants from the National Assembly and Senate, spokespersons from the Ministry of Interior, and a large audience from embassies, from international organizations and from NGOs and the public.
One of the results of the workshop was to bring public attitudes, gathered at the grassroots by means of social research, into a high profile discussion with government officials that received considerable TV attention. Another result of the workshop was to raise the issue of the kind of voter education campaign that would be required, once the Commune Laws were finalized, in order to meet the needs of the electorate to participate in these local elections with understanding.
Methodology
The Impact Survey instrument was developed in English and Khmer, using the Baseline Survey as a guide, especially where follow-up questions were concerned. The written questions were composed in a brief, simple language to avoid problems of formality and status that are encoded in literary Khmer. The CAS team of trained survey researchers was especially sensitive to the need to avoid “translating,” or elaborating on the written questions, when dealing with illiterate respondents.
We utilized “semi-open” questions by which respondents were asked to volunteer examples or instances in the order that came to mind, while the researcher prompted the respondent to give more examples. These replies were pre-coded and entered by the researcher on the questionnaire. The pre-codes were generated by means of an extensive pretest of the instrument. (The Khmer and English versions of the instrument are attached in the Annex.)
Our sample came from seventeen Provinces and Municipalities and included respondents from fifty-six districts/khan. Extensive census had not yet become available when we designed our research, so we used the preliminary data included in the Demographic Survey of Cambodia, (1996),[3] as we had done for the Baseline Survey.
Representative quota samples were constructed as far as possible on the basis of the Demographic Survey, using categories of gender, age and rural /urban based on the composition of the total population, and applied to the sample in each district selected for study. The sample was obtained from a randomly selected commune by random household calls in rural areas as well as from various sampling points in public urban space, especially markets. Following this methodology, a sample of N=546 was collected in August to October 1999.
As detailed Census information on CD becomes available from the National Institute of Statistics and when commune boundaries are settled after the commune elections, random sampling at khum, phum and krom levels will be more firmly based. It should also be possible to combine survey data with GIS mapping data to produce spatial analyses in studies like these. This a refinement of findings that could be especially useful to future voter education campaigns.
Survey execution was carried out by three teams of five researchers each. The rainy weather and poor road conditions made travel difficult in our 4wd vehicles. This reduced our rate of questionnaire completion, but did not deter us from visiting the often remote communes that had been selected for the sample.
In conducting the Impact Survey, our teams did not face the difficulty we had experienced in the Baseline Survey, in which respondents regarded our questions with suspicion that we were operatives of a political party. The open questions elicited brief but frequent responses from informants, and are summarized in the following report.
The last question of the questionnaire involved a test of recognition of voter education print material. We showed the informant a packet of laminated photographs (5 x 7 inches) of posters, brochures and booklets that had been distributed widely before the elections. We had made a collection of these materials from voter education sources. Each photograph was numbered. We asked respondents whether they had seen the original item during the pre-registration and pre-election period. We found respondents were very conscientious in examining each of the forty-two photographs to recollect if they had seen the item or not. The photographs often attracted a group around the informant and the so the response elicited may reflect a more general awareness of the print material than that of an individual respondent. But we felt that the response given could still be very useful to ascertain how widely the posters and other materials had circulated among the electorate.
The responses to these photographs enabled us to generate a “voter education awareness” variable. We counted the number of items that each respondent recognized in this set of images. We ranked the respondents from fewest to most items recognized and then divided this ranked list into three approximately equal groups. The groups were then identified as “low,” “medium” and “high” levels of voter education awareness. This variable was then used, along with the demographic variables, to develop interesting cross-tabulations with knowledge, attitude and awareness variables in the survey.
We did not develop an efficient way to test voter recognition of live theatrical, audio or video voter education materials. A high capacity portable computer taken to the field with multimedia recordings of such performance and broadcast materials should make this test possible in the future.
Analysis
We used SPSS for Windows to process the questionnaire data and produce the chart output, as we had done for the Baseline Survey. This made the comparisons of Baseline and Impact findings very straightforward.
To assess the statistical significance of cross tabulated variables we used Pearson chi-square tests. Significance is conventionally cited at the p=0.05 level, p=0.01 level and p=0.001 level. This means that the probability, p, of the association between the variables being due to sampling error is at this level or less. To put it another way, our confidence that the association observed between two variables is not due to sampling error is 95%, 99% and 99.9%, respectively. In this study we note significant associations at the 95% level of confidence and better.
I. Background
Demographics
These background variables are presented to show how our closely our sample mirrors the Cambodian population as a whole. We also use several of these demographic variables to test for association with variables concerning voter knowledge and awareness. This analysis can help provide specific recommendations for future voter education campaigns.
A.
Rural/Urban
We decided not to use government designations of jurisdictions as the basis for our distinction of rural and urban because the boundaries of the jurisdictions were often not easy to ascertain. Following a procedure we used in our Baseline Survey, our researchers noted whether the community in which respondents were sampled was a rural or urban locale, depending on their consensus.
For the purposes of refining our cross tabulations, where a rural/urban distinction might be interesting and informative, we created a variable derived from a question on whether a respondent were a rice (srae) farmer, garden (chamkar) farmer or neither. The first two represent “rural” respondents or “farmer” respondents and the third represents “urban” or “non-farmer” respondents, for the purposes of this study.
According to the
estimates of the Demographic Survey(1996),
B. Gender
The proportion of
men to women of voting age (³18) which
can be derived from 1998 Census tables is 46.6% males to 52.2% females. Our sample of 43% males to 57% females
over-represents women slightly. One
reason for this is that during the day in village
C. Age
According to Census tables, Cambodians of voting age (³18) comprise 49.4% of the total population. The age distribution of our sample, which was designed to obtain quotas for each age category, conforms fairly well to the distribution of ages of voting age persons that can be derived from Census tables. Middle aged voters are slightly over-represented and the youngest and oldest voters are slightly under-represented in our sample.
|
Age cat. |
18-24 |
25-29 |
30-34 |
35-39 |
40-44 |
45-49 |
50-54 |
55-59 |
60-64 |
65+ |
|
Census |
21.5% |
15.8% |
13.6% |
12.3% |
8.7% |
7.3% |
5.5% |
4.5% |
3.6% |
7.1% |
|
Survey |
18.3% |
15.8% |
12.3% |
13.6% |
8.8% |
10.6% |
8.4% |
4.9% |
4.0% |
3.3% |
D. Education
Our survey included people of voting age (³18) and included both literate and illiterate respondents. We asked informants the highest level school they had ever attended, regardless of whether they had completed that level or not. The 1998 Census discusses education only for literate persons and only considers persons ³25 years of age. Literacy was not at all a criterion for participation in our survey, and that is reflected in the comparatively high rate of no schooling and low rate of primary schooling in our sample, compared to Census figures.
|
Ed. Level |
None |
Primary (completed
and not) |
Middle Secondary |
Higher Secondary |
Higher |
|
Census ³25 |
2.1% |
81.3% |
11.8% |
4.0% |
0.8% |
|
Survey ³18 |
16.5% |
42.3% |
27.7% |
9.9% |
3.1% |
F. Marital
Status
Our survey was aimed at voting age persons (³18), whereas the Census gives marital status figures for persons ³15 years of age. This accounts for the comparatively lower figures for “unmarried” and higher figures for “married” in our sample than in the Census figures. Interviewing in a village is also likely to focus respondents in households, who would likely to be married persons. Young unmarried women of voting age are unlikely to agree to speak with strangers conducting a survey. Young unmarried men are not likely to be at home during the day. Our survey followed colloquial Khmer usage, which does not distinguish divorced and separated categories.
|
|
Never |
Married |
Widowed |
Divorced |
Separated |
|
Census ³15 |
29.4% |
61.3% |
6.5% |
2.4% |
0.4% |
|
Survey ³18 |
18.3% |
70.1% |
7.9% |
2.6% |
|
G. Ethnicity
The recent
national Census completely avoids the issue of ethnicity, which reflects the
controversy surrounding the status of some minorities in
Self-identified “Chinese” respondents are very highly associated with the urban category. “Khmer” respondents are highly likely to be rural farmers. Respondents identifying themselves as “Sino-khmer” are not significantly associated with one category, rural or urban, over the other in our sample, but they are highly associated with the occupation “trader.”
G.
Socio-economic Level
We provided several questions in the survey to find a reliable indicator for socioeconomic level that might associate in an informative way with voter attitudes and knowledge.
1. Radio, TV ownership; Income
Ownership of a radio was not significantly associated particularly with rural or urban respondents. In contrast, respondents who owned a television, or whose reported income was ³$50 are very highly associated with the “urban” category. Absence of TV in the household and income <$50 are highly associated with the “rural” category, as would be expected. Accordingly, radio broadcasts are more likely to reach both rural and urban audiences, while TV broadcasts would be more likely to reach wealthier, urban audiences.
2.
Land Ownership, Land Holdings, Occupation
Our respondents,
rural or urban, generally consider themselves to be owners of their land,
regardless of the actual uncertainties and controversies over land tenure in
the law of
In an attempt to distinguish socio-economic levels within the farming community we asked about the extent of land cultivated. The “not a cultivator” category in this variable is virtually identical with the “urban” category of our Rural/Urban variable.
The distribution of respondents in occupations indicates that many of those who identify themselves as farmers also engage in trade or business (farmer= 79%, non farmer= 21% in our sample).
H. Migration
Our findings on whether respondents moved into the present locale from elsewhere are comparable to the Census figures. The 1998 Census finds that 68.5% of the population had always lived in the place of enumeration. Our sample, which excludes children under voting age (<18), shows 60% of respondents had been born in the present locale (khum, commune). “Liberation” in the chart below is the way the defeat of Pol Pot Democratic Kampuchea forces by Khmers backed by the Vietnamese Army is commonly referred to by the present regime.
II. Participation in the 1998 National Elections
Our sample mirrors the voter turnout in the National Elections, which was 93.74% according to the National Elections Commission.[5]
Although the proportion is small, we were particularly interested to learn what reasons the non-voting respondents had for failing to participate in the elections. The main reason was that the respondent had been too young at the time.
III. Knowledge of Registration and Balloting
In our Baseline Survey we found that nearly a quarter of respondents did not know that they had to register to vote, even if they had voted in the UNTAC elections of 1993. In this Impact Survey we asked respondents to mention, in order of importance, the sources by which they had been informed of the need to register to vote. Likewise, we asked which sources were most informative in explaining the registration requirements and process and in explaining the voting process.
The striking
finding of our survey is that the village chief, mephum, is by far the most important source of information on these
civic affairs for our largely rural respondents. (The mekrom is the head of a “group” of about 10-20 households. The mekhum
is the commune chief). Given the great
importance of the village chief in informing his constituents, it is noteworthy
that current designs for decentralization of governance in
It is clear that the elected commune councils will be the most important target for good governance capacity building efforts at the grassroots and the natural focus for voter and civic education campaigns. But the future role of village and group leader in facilitating elections and other civic activities remains unclear.
IV. Experience of Difficulties During the Process
from Registration to Elections.
A. Experience of
the Registration Process
The registration process, which was heavily supported by the European Union, appears from the responses to our Impact Survey to have been conducted with few major problems for the electorate. This finding is reflected in very low frequencies for the potential problems we listed. We specifically asked for any personal encounters with difficulties rather than comments about registration irregularities about which respondents may have been aware in the form of rumor.
B.
Perceptions of Danger and Fear during the Elections
In both the Baseline Survey, before the elections, and in the Impact Survey, after the elections, we asked respondents if they felt the contest and campaigning associated with elections created unsafe or dangerous conditions for them. Our findings naturally show a sharp decrease in “don’t know” responses from before the elections. But we note a doubling of both the “yes” and the “no” responses after the elections.
|
Danger |
Yes |
No |
Don’t
know 1998 (No
reply 1999) |
|
Baseline (1998) |
14.9% |
34.8% |
50.2% |
|
Impact (1999) |
29.3% |
69.6% |
1.1% |
The “yes” response, a positive perception of danger, is very significantly associated with urban, non-farmer, respondents (p=.002). But there is a similarly strong association of this perception of danger in connection with the elections among very poor, landless respondents (p=.002). Men tend to be associated with a positive response (p=.046) to this question. No other statistically significant associations were found.
At one end of the socio-economic spectrum, urban respondents evidently feared a recurrence of factional fighting in the city and the possibility of danger to themselves or their property. At the other end of the spectrum, vulnerable landless peasants are probably prone to feel the potential for danger in any political events.
We asked another related question in both Baseline and Impact Survey, if respondents felt fearful about going to cast their votes.
|
Fearfulness |
Fearful |
Not fearful |
Don’t
know 1998 (No
reply 1999) |
|
Baseline 1998 |
20% |
62% |
18% |
|
Impact 1999 |
7% |
86% |
(6%) |
The dramatic rise in the feeling of “not fearful” about voting is reflected in the very high voter turn-out in the 1998 National Elections. In other questions in the Impact Survey, we tested for some possible connections between fearfulness and sense of danger and the activities of the voter education and election monitoring community.
C.
Perceptions of Monitoring During the Election
We asked respondents if they noticed any independent monitoring of the polling process. Unfortunately we failed to ask respondents to identify any insignia or logo in order to verify that they had accurately perceived a monitor from an independent organization was present or not.
The “yes” response is significantly associated with urban respondents, and the “no” response with rural respondents (p=.027). The 25% of our respondents who said “no” or gave no reply may not have recognized the independent monitors who were there.
This finding suggests that all independent monitors should agree on a standard identification device, a logo on a tee shirt or a badge, and that this identification be a prominent part of future voter education campaigns.
We also asked how the presence of monitors made the respondent feel. The responses were: “The monitors did not affect me, they were irrelevant to me.” “The presence of monitors made me feel safer about voting.” “The presence of monitors made me feel fearful about voting.” The bar chart below shows the distribution.
If we correct for the respondents who did not notice monitoring, the frequencies are:
|
Irrelevant |
Safer |
Fearful |
No reply |
|
33.4% |
60.9% |
5.7% |
9.6% |
The occupational group “traders” is very highly associated with the response “monitors made me feel fearful” (p=.004). Those in the “high”group for recognition of voter education materials are very strongly associated with the response “monitors made me feel safer”(p=.006). While those in the “low” group for recognition of voter education material are associated with the “monitors were irrelevant to me” response.
Voter education materials had a demonstrable impact on making people aware of the function of independent monitors and their organizations in contributing to a safe experience at the polling place. However, our findings also highlight the fact that certain important sub-groups, like traders and market sellers, require special targeting with voter education messages to convince them of the positive value of monitoring to assure safe polling
D.
Experience of Difficulties during the Polling
We asked respondents to note which difficulties they or anyone in their household had noticed or encountered at the polling station. The sum of percents is greater than 100% because a few respondents ticked two or more items.
Two difficulties stand out in the chart. Respondents were very strongly aware of armed persons within 200m of the polling place (44.7%) and noticed the presence of local authorities in or near the balloting station (59.9%). Both represent possible sources of intimidation and both are specifically prohibited in the Election Law.
The widespread presence of armed persons and civil authority at polling stations, indicated by these findings, provides a powerful argument for the importance of independent monitors in assuring free and fair elections. The presence of monitors is clearly needed to counter intimidation and to promote a sense of safety for the electorate while they vote.
V. Influences on
Voting Behavior
A.
Secrecy of the ballot
In the matter of secrecy of the vote, a comparison of Baseline survey results from before the National Elections, reflecting expectations of respondents, with Impact survey results of respondents’ experience, is very instructive.
The “wait and see” attitude registered by 52% of respondents before the election, who said they did not know if the elections would be secret or not, was largely converted to a positive “secret” response, after the experience of balloting. The presence of international and national monitors no doubt contributed to the predominant experience of secrecy and safety of the ballot, unresolved complaints about the process notwithstanding.
B. Need for
advice on how to vote
In both Baseline Survey and Impact Survey, we asked respondents if women should seek advice from men on how to vote.
|
Seek advice from
men |
Yes |
No |
Don’t know |
|
Baseline 1998 |
47% |
45% |
6% |
|
Impact 1999 |
36% |
58% |
5% |
The context for the question is the widespread party propaganda that husbands and wives should vote the same way. This, in turn, draws on the stereotype that the man should take care of matters outside the house. The results of our comparison show impressive changes from near equal frequencies of “yes” and “no” responses before the election to decreased “yes” responses and increased “no” responses in our post-election survey. Nevertheless, there is still a sizeable segment of the electorate who believe that men should instruct women on these political issues.
The responses to this question, as one would expect, are significantly associated with gender. Men are significantly associated with the “yes” response and women with the “no” response (p=.000). Of those who replied “yes” in our Impact Survey, 57.1% were men, and 42.9% were women. Conversely, of those who replied “no,” 35.3% were men and 64.7% were women.
We found a significant association of education with responses to this question. Those with higher education level (high school and above) were associated with “yes” responses (p=.039). There is also a significant association of greater wealth with the “yes” response (p=.041).
These surprising findings suggest that relatively sophisticated, richer, more educated respondents, especially men, are the important target to reach with messages about women’s right to an independent vote. The corollary message to women in these same categories might be that their vote is secret and has nothing to do with their loyalty to their husband.
C.
Intimidation
In the Baseline survey, we asked if respondents expected any pressure or force to vote for any party. In the Impact Survey, we asked if respondents had experienced any pressure or force to vote in a particular way.
|
|
Yes |
No |
Don’t know |
|
Pressure |
|
|
|
|
Baseline (expectation) |
7% |
77% |
15% |
|
Impact (experience) |
6% |
91% |
0 |
|
Force |
|
|
|
|
Baseline (expectation) |
6% |
77% |
17% |
|
Impact (experience) |
5% |
93% |
0 |
The “don’t know” responses, indicating a “wait and see” attitude from before the elections, gave way to strong responses reflecting the voters’experience of no pressure and no force to vote in a particular way.
The strong perception of ballot secrecy and safety, enhanced by the presence of independent monitors, probably contributed to voter resistance to extreme forms of intimidation.
D. Vote buying
We asked Impact Survey respondents about attempts to influence their vote with gifts by asking two questions. The first question, “Did you or any one in your household receive any gift to vote for the party of the gift giver” got a 40% “yes” reply and 60% “no” or “don’t know” reply. However, when we asked, “What gifts did you receive?” 78% of respondents named a gift and only 22% gave a “no reply” response.
The tendency to
use gifts to buy votes, very noticeable in other countries in the region, is
already quite prevalent in
These results suggest a strategy in voter education to stress messages emphasizing the right to vote one’s conscience regardless of the gifts offered to vote a particular way, in a political climate where vote buying will likely increase.
E. Obligations
to vote in a particular way
In both Baseline and Impact Survey we asked respondents if they were obliged to vote for the party of which they were a member, or for the party to which they had made a pledge.
|
Membership |
Yes |
No |
Don’t know |
|
Baseline 1998 |
28% |
24% |
48% |
|
Impact 1999 |
31% |
26% |
43% |
|
Pledge |
Yes |
No |
Don’t know |
|
Baseline 1998 |
35% |
38% |
27% |
|
Impact 1999 |
18% |
39% |
43% |
In the case of obligation entailed by party membership, the “don’t know” responses dropped in 1999 from the pre-election 1998 levels, but both “yes” and “no” responses increased in the survey conducted after the election.
On the other hand, in the case of obligations entailed by party pledging, the “don’t know” response increased dramatically from 1998 levels while the “yes” response decreased and the “no” response stayed virtually the same.
In our analysis of perceptions of obligation linked to party membership, we find a significant association of “high” voter education awareness with the “no” response and of “low” voter education awareness with the “don’t know” response (p=.028). Otherwise, we found no significant association with gender, age, education, or socio-economic level.
These results show a demonstrable impact of voter education on loosening the bonds of party membership, in a context where aggressive recruiting techniques are often used to swell party membership. The results suggest intensified voter education emphasis should be placed on voting one’s conscience in the privacy and safety of the polling booth.
In our analysis of perceptions of obligation entailed by party pledges, again, we found a significant association between “high” voter education awareness and the “no” response and between “low” voter education awareness and the “don’t know” response (p=.009). Of the respondents replying “don’t know,” 40% were in the low voter awareness category, although that category comprised only 32% of total respondents.
The only other significant association we could find was to age. Younger voters, £39, are associated with the “yes” response (p=.038).
These findings suggest that continued voter education efforts be directed at undecided young voters. These messages should emphasize the freedom and secrecy of the ballot booth in which the voter’s real convictions and preferences can be expressed, regardless of any commitments made outside the polling place.
VI. Voting Rights
In the Baseline
and Impact survey, we asked the same questions about whether certain non-Khmer
ethnic groups in
|
Non-Buddhists |
Yes |
No |
Don’t know |
|
Baseline
1998 |
61% |
7% |
32% |
|
Impact
1999 |
58% |
7% |
36% |
|
Minorities |
Yes |
No |
Don’t know |
|
Baseline
1998 |
58% |
6% |
37% |
|
Impact
1999 |
56% |
5% |
39% |
|
Immigrants |
Yes |
No |
Don’t know |
|
Baseline
1998 |
12% |
24% |
63% |
|
Impact
1999 |
36% |
19% |
45% |
There is little difference between Baseline results and Impact results in the findings on voting rights of non-Buddhists and tribal minorities. However, there are dramatic changes in perceptions about voting rights of immigrants between the 1998 pre-election survey findings and our findings from 1999, after the elections.
Using chi-square tests (with significance level 0.05 or better) in our analysis, we find for voting rights of non-Buddhists that women, poor (holding £ 1ha of land) and those with low education (<primary school) were highly associated with the “don’t know” response. Young voters (£ 39) and those with lower levels of education were associated with the “no” response.
For the voting rights of minorities, women, younger voters, voters with lower education and poor (those not owning a radio, or TV) were associated with the “don’t know” response. Richer voters were associated with the “yes” response.
For the voting rights of immigrants, women, older voters (>40) and less educated voters were associated with the “don’t know” response. Men were associated with the “no” response. Younger voters (£39) and more highly educated voters were associated with the “yes” response.
For all three questions on voting rights, there is a strong association between “high” voter education awareness and “yes” responses, and a corresponding association between “low” voter education awareness and “no” and “don’t know” responses (p=.000 for these associations).
These findings suggest
that voter education campaigns directed at undecided and younger voters can
have a demonstrable impact on perceptions of the rights of non-Khmer minorities
in
VII. Consequences of
the Vote
A. Knowledge
of the elected representative
In order to test voter knowledge of the results of the election, in the Impact Survey we asked respondents to name any member of the National Assembly who was elected from the respondents’ province. We had asked the same question in the Baseline Survey before the elections to provide a point of reference for comparison.
|
MP’s name |
Correct |
Incorrect |
Don’t
know |
|
Baseline 1998 |
1% |
4% |
94% |
|
Impact 1999 |
13% |
3% |
83% |
The sharp increase in correct responses in 1999 over the Baseline results of 1998, from 1% to 13% correct, is expectable from the fact of the recent election. However, the profound lack of knowledge in the electorate of the identity of any MP elected from their province, expressed in the huge “don’t know” response, indicates a fundamental problem of communication of election results to the electorate.
The finding that 83% of our respondents do not know the name of any of the “representatives” they just elected suggests possible problems in the accountability of MPs to a constituency of voters in Cambodia.
These findings indicate the importance of civic education campaigns directed at elected officials in order to raise the question with them about their concept of the relationship of elected officials to a constituency that votes them into office.
The importance of this issue of accountability is heightened by the prospect of commune elections in the near future. Vigorous voter education and civic education efforts are clearly needed at the grassroots to avoid the gulf in communication between elected officials and the electorate at the commune level that seems to exist at the national level. Only with such education efforts to increase effective demand for accountability will the process of democratization and decentralization in Cambodia be successful.
In our Impact Survey, we asked what the respondents felt the MP should do for the electorate that elected him/her. A remarkably consistent and strong reply from nearly every province we surveyed was that the principal duty of the MP was to improve the livelihood of the people, particularly in regard to water control and rice cultivation.
The most often cited needs were irrigation, dams, canals, bridges and irrigation pumps. Other infrastructure improvements mentioned were to build pagodas, schools, hospitals and roads. The second most often cited duty of elected MPs was to improve the livelihood of the electorate by distributing rice seed, fertilizer, farming tools, ox carts, cattle and buffalo and mosquito nets.
In a few provinces, Battambang, Kampot, Kampong Thom, Siem Reap and Kampong Cham, respondents called for MPs to seek improvements in livelihood by the construction of factories to provide employment for the people.
Respondents from Phnom Penh stressed the duty of MPs to repair streets, drains, sewers and water supply and to reduce water and electricity rates. Phnom Penh respondents also cited the duty of MPs to improve security against gangsters, theives, robbers and kidnappers and to control anarchic building and to arrange housing for homeless people.
Respondents from several provinces called on MPs to protect the environment and prevent the destruction of forests, to increase the price of farm products and decrease the prices of shop goods.
A number of respondents, scattered over all provinces we surveyed, mentioned that MPs broke their promises and did nothing for the village, or that the MP should visit villages to become acquainted with what people need and so that people can know their face.
B.
Awareness of benefits to the community from the election
In the Impact Survey after the election we asked if respondents were aware of any benefits the community had received as a result of the election. The 11% of respondents who answered “yes” answered a second question by naming the benefits of which they were aware.
The connection of local development projects to the election of representatives who will advocate on behalf of their constituency is still relatively weak in Cambodia. One of the main reasons for this is that development is often perceived to be facilitated by International Organizations. Traditional forms of labor mobilization for local infrastructure improvements, which were so misused during the Khmer Rouge regime, are slowly reappearing, often in connection with the Buddhist pagoda.
The election of commune councils with responsibility for development and with the power of taxation may bring significant changes in the demand for accountability and effectiveness in elected officials to fulfil community needs for public goods.
VIII. Commune level
elections
A. Voter
awareness
We used the opportunity of the Impact Survey to look forward to commune elections, which are being planned for the near future. We first asked if respondents had heard that commune level elections would be held soon.
The strong “yes” response is significantly associated with male respondents (p=.000). Of respondents answering “no,” they were unaware of upcoming commune elections, 26.5% were men and 73.5% were women respondents. Wealthier respondents and those with higher education levels were also strongly associated with positive awareness of forthcoming commune elections (p=.000 for various indicators).
Women and the poor should be especially targeted for voter education and civic education efforts to bring up their understanding of the importance of these upcoming local level elections
B.
Sources of information
We asked respondents to mention the sources of information by which they had learned of commune elections. It is noteworthy that, in comparison to the importance of the village chief as a source of information about National Elections, the government authorities at the local level at the time of our survey were not yet mobilized to inform the public of impending changes in governance at the local level.
These findings suggest an opportune opening for civil society voter education organizations to reach the electorate with civic education that will assure the success of decentralization plans that are underway. Radio broadcasts with special programming to raise grassroots civic consciousness could be developed and presented even before the draft laws on commune council formation are finalized.
C. Secrecy
We asked respondents to the Impact Survey whether they thought the elections for commune council would be secret. The results can be compared with Baseline Survey expectations for secrecy in the 1998 National Elections, and with the experience of secrecy in the balloting that we recorded in this Impact Survey after the election (discussed above in section V).
|
Secrecy |
Secret |
Not
secret |
Don’t
know |
|
Baseline (1998) expectations of National Elections |
47% |
1% |
52% |
|
Impact (1999) experience of National Elections |
86% |
2% |
12% |
|
Impact (1999) expectations of Commune Elections |
41% |
9% |
49% |
The chart below shows the comparison between our Impact Survey results about voter experience of secrecy of the ballot during recent National Elections and the expectations of our respondents about the secrecy of forthcoming Commune Elections.
Our analysis shows that respondents in the “high” category for voter education awareness are strongly associated with the affirmative view that the commune elections will be secret. Those in the “low” category for voter education awareness are significantly associated with the “don’t know” response (p=.000). Younger voters (£39) are more likely to say “yes” the elections will be secret, while older voters are likely to give the “don’t know” response (p=.048). Those with a higher level of education are likely to give the negative response that the commune elections will not be secret (p=.022).
These findings present a
clear challenge to the voter education community and election monitoring
organizations. They must try to overcome
the deep and abiding suspicion in the electorate that the secrecy, and thus the
free and fair nature of elections planned for the commune level, may not meet
the standards already set by recent National Elections.
IX. Candidates for
the Commune Council elections.
A.
Political party candidates
We asked respondents whether they thought political parties should put up their candidates for commune elections. Respondents with a higher educational level are significantly associated with the “yes” response, while those with lower education are associated with the “don’t know” response (p=.003). Respondents with “high” voter education awareness are associated with the “yes” response, while those with a “low” voter education awareness are associated with the “don’t know” response (p=.005).
The dramatically divided response to this question suggests that respondents with less education and information are especially likely to withhold judgement about the involvement of political parties in local level elections. A massive civic education effort will be needed to assure that the electorate understands whatever commune election system is adopted in law. That grassroots civic education will be necessary to overcome the doubt and misgivings associated with the history of party conflict and violence in Cambodia.
B. Political
party members elected to commune level office
We asked if respondents thought political party members should be elected to commune level office. On this question, the only significant association our analysis could uncover was between wealthier respondents (those owning a TV) and a “no” response (p=.029).
We asked respondents to explain their answer to this question, if they wished. The main positive views respondents gave for their answers to this question were that “they have a right like other people in a democracy,” or “they should be elected if they have the knowledge/ability, if they are good/honest and if the people like them” (all provinces). The main negative remarks given by respondents are that “they are partial to their party and nepotistic,” or that “we do not know their mind/heart well” (all provinces).
Some other illustrative responses include the value of competition and cooperation between parties. “The more people from all parties who participate, the more likely we are to get good people” (Kandal). “The competition among many parties will be an opportunity to show the mistakes they have made” (Kampong Cham). “We have a coalition of parties in parliament; it is the same in the commune” (Kampong Thom). “Everyone is a member of one party or another” (Kampot). “It is better to share ideas and so parties should cooperate; party members should not kill one another” (Phnom Penh).
C.
Candidate lists
We asked if there were an election for commune council, if members should be elected from lists prepared by Parties, groups and individuals in the commune. Such a closed list system is familiar to Cambodian voters from the recent National Elections. This involves a mechanism of proportional representation, which is favored in the draft law on commune councils that was available at the time the Impact Survey was designed.
Those in the “low” category of voter education awareness are highly associated with the “don’t know” response to this question (p=.000). Women are also significantly associated with the “don’t know” response, and men with the “yes” response (p=.013).
D. Candidates
from villages
We asked if there were an election for commune council, if the members should be representatives elected from each village of the commune. This more direct form of representation was being widely discussed in civil society as an alternative to the party list option favored by the government, at the time the Impact Survey was designed.
Our findings show extremely strong support for village representation and weaker support for the party list alternative. This suggests voter preference for direct election of individual village representatives to the Commune Council rather than the system of proportional representation based on lists.
Our respondents clearly perceive that direct elections produce officials who feel accountable to a village constituency of voters rather than to a political party. It is essential that the voter education campaign organized in connection with the new commune councils acknowledge the evident importance of the village and village identity as the focus for local issues, whatever election system may be adopted in the law.
E.
Rights of women and poor to be candidates
We asked if women and poor people should be elected to commune level office and why. The results are overwhelmingly positive and mirror the very strong sentiment in favor of voting rights of women and the poor that we observed in the Baseline Survey.
The most frequent explanations for the positive response for women and poor as elected officials at the commune level were, “they have equal rights as citizens,” and “they should be elected if they have the knowledge and ability” (all provinces). An additional frequent response concerning poor people as elected officials was, “they have experienced the bitterness of poverty so they can know the needs of other poor people” (all provinces).
Other illustrative positive responses concerning women as elected officials include, “women can work better than men and do not drink wine” (Siem Reap, Kampong Thom). “Women are good at finances of the family so they would be good working with commune finances” (Kandal). Negative responses for women as elected officials included, “there never was a woman mekhum in the past” (Kampot, Prey Veng). “Women could not work at night” (Takeo).
Illustrative positive responses for poor persons elected to commune office include, “the poor are less corrupt than the rich and do not take advantage of others” (Kampot, Kampong Thom). “The poor can also lead, they have the same right as the rich and they would not look down on poor people” (Prey Veng). Negative responses for the poor as elected officials included, “people will not listen to a poor person who is a leader; the rich will not listen to a poor person” (Kandal, Kampot, Takeo). “If a poor person becomes mekhum he will become corrupt because he thinks about his living condition” (Kampot, Kampong Thom, Battambang). “Poor people should not be elected to commune office because they have no knowledge and no money to bribe anyone” (Siem Reap).
X. Expectations for
Commune Councils
A. Community
needs
We asked what the priority needs were that respondents thought elected commune level officials should address. One main category was to attend to local infrastructure development, roads, schools, health centers. Another category was to take care of the people and to conduct peaceful, fair resolution of disputes.
We tried to approach the issue of expectations about elected commune officials by asking respondents to mention any advantages they saw in electing local officials. The responses to the open question can be summarized under three categories, leadership, dispute resolution and development.
Under leadership, we note that respondents often used the idiom of the elected commune official as a shade that would shelter villagers and provide security. “They should work for the people, improve security, peace, stability” (Prey Veng, Kampot, Takeo, Kampong Thom). “They can help solve problems when there is an emergency”(Kandal, Kampong Thom, Kampot). “They should eliminate thievery, make good order and security in society” (Kampong Speu). “They can bring information from the government to the people and convey the peoples’ ideas to higher ranking levels” (Kandal, Takeo, Svay Rieng).
Under dispute resolution, the most frequent reference was to land disputes. “The elected mekhum should be able to solve disputes about roads and land” (Kampong Thom, Banteay Meanchey). “He should resolve conflicts over land and family disputes and give fair judgments” (Pursat). “The new elected mekhum must solve land problems with a good attitude and without corruption” (Kandal, Svay Rieng).
Close after land disputes among our respondents’ remarks was the commune leader’s responsibility to solve general social problems. “He has to mediate disputes, provide justice, settle problems related to quarrels and theft” (Prey Veng, Kampot, Kandal, Takeo, Svay Rieng).
The third main area of commune leader responsibility was to facilitate development. “The new mekhum will build infrastructure, roads, schools, wells, ponds” (Kampong Chhnang, Kandal, Kampong Speu, Siem Reap). “He will obtain low cost loans from organizations when we need to borrow money” (Kampot). “He will call organizations to help the people” (Pursat). “He will enlarge the market, it is too cramped” (Kampong Som). “They will help people gain access to fishing” (Kampong Cham).
B. Limits on the Commune Council
We asked respondents to mention any duties or powers that elected commune officials should not be given. The most frequent response was that the new local officials should not oppress or threaten the people or exploit them in other ways by corruption or expropriation. Nearly 80% of respondents voiced a response of this kind. This may indicate the background of complaints about local authority that has developed since most of these officials were appointed during the Vietnamese-backed PRK regime. This is the context in which the forthcoming commune council elections will be held.
C.
Accountability of elected commune officials
We asked two questions to probe respondents’ sense of the significance of elections for local officials in contrast to the pattern of appointed officials which has prevailed up to now.
We asked respondents to mention who they thought the new elected commune officials should listen to when making their decisions.
The very strong response in favor of consultation between elected officials and the villager electorate suggests that the respondents to this survey understand the concept of a constituency. Both voter and civic education efforts aimed at the general population and capacity building efforts aimed at the new elected officials should build on this understanding of a linkage of responsibility between an elected official and the voters who put him/her into office.
Finally, we asked respondents if they thought they would have the right to use their vote to remove from office any elected local official who was found unacceptable. The responses given to the question about powers the newly elected commune leaders should not have probably indicates what respondents might mean by “unacceptable.” At the same time, the responses given to the question about the most important duties of elected local officials probably indicates the expectations voters have for their local officials.
Unacceptable or unsatisfactory conduct of the newly elected commune level officials evidently means oppressing, threatening and stealing from villagers. Voters evidently expect their commune council members to provide good leadership, fair dispute resolution and development activities. Failure in these areas will likely have an effect on a commune official’s chance for reelection.
These results indicate that the Cambodian electorate overwhelmingly grasps the fundamental tenet of democracy: that elected leaders are supposed to be accountable to their constituency and can be replaced by the voters if they are found to be unsatisfactory. These extremely significant findings suggest that Cambodian voters will be expecting to hold their elected commune and sangkat leaders accountable, whatever shape or form the commune/sangkat council takes in the law.
This is the theme that
should be stressed in voter and civic education efforts in connection with the
commune council elections. The concept
of accountability to a constituency is a resource already present in political
thinking at the grassroots. This
resource should be strengthened and consolidated to assure the success of the
process of decentralization and democratization in Cambodia.
XI. Voter Education
A.
Broadcast media
In the Impact Survey, we asked what media programs from the period before the elections were most memorable to respondents. The strongest response volunteered was “stories” (26%) followed by “other” (12%) and “songs” (10%). The narrative vignettes and dramatizations with voter education messages were clearly most effective in making a lasting impression on listeners and viewers.
We also asked what message in the media was most prominent and memorable to respondents. The strongest response volunteered was the message urging the audience to “vote in the elections” (18%). The second most memorable message was “voting instructions” (9%) and the third most memorable message was the urging to “register to vote” (7%).
Our findings indicate that narrative dramatizations are the most effective programming to reach the electorate through the broadcast media.
B. Print media
We showed respondents photographs of posters and brochures that had been used in the voter education campaign to carry messages about registration and voting. We asked respondents if they recognized the material from the pre-registration and pre-election period or not. Scanned images of these print media for registration material and for voting material are arranged below in descending order of percent recognition. (The number before the name of the organization that produced the poster is the number of the item in our questionnaire.)
We included two items that are virtually identical, except for the title and indication of the sponsor of the message. These items were placed far apart, near the beginning and near the end of the packet of photographs shown to respondents. Our aim was to test the reliability of this kind of test of voter recognition of photographs of poster material. The two items are #149 and #172. They were recognized at rates close to one another (60.6% and 54%) and are found ranked next to one another in the following presentation of images. The test verifies the usefulness and reliability of this kind of question in a survey in Cambodia.
We also used the responses to this poster recognition test to develop a variable that we called “voter education awareness.” We counted the number of items that each respondent recognized in this set of images. We ranked the respondents from fewest to most items recognized and then divided this ranked list into three approximately equal groups. The groups were then identified as “low,” “medium” and “high” levels of voter education awareness. The variable was then used to develop interesting crosstabulations with other variables in the survey.
This voter education awareness variable was crosstabulated with a number of demographic variables to compose a profile of respondents with “low” and “high” voter awareness that might be useful for future voter education campaigns.
We find that the “low” level of voter education awareness is significantly associated with female respondents and the “high” level with male respondents (p=.000). Older respondents (³35 years) are associated with the “low” level of voter education awareness, while younger respondents (18-29 years) are associated with the “high” level (p=.016). Respondents with low education level (primary and less) are associated with the “low” level of voter education awareness, while those with higher education level (middle school and higher) are associated with the “high” level of voter education awareness (p=.000)
There seems to be no significant associaton to most of the socio-economic level indicators we used (radio, TV ownership, size of land holding, income). However, there is an interesting pattern for urban/rural respondents.
When we analyzed the voter education awareness level based on the Registration Materials, we found a significant difference. Rural respondents were associated with the “low” level of voter education awareness and urban respondents with the “high” level of awareness (p=.000). In contrast, when we analyzed voter education awareness based on Election Materials, we found the association with urban/rural respondents was not significant at the levels we had adopted for this study (p£ .050).
This finding suggests that registration related material was prodominantly distributed and recognized in urban locales. While election related material was distributed and recognized more evenly in urban and rural locales.
The finding clearly reflects the improvement in distribution techniques from the initial registration education phase of the election process, when EU support was strong, to the later balloting education phase of the voter education process, and demonstrates the impact of the Cambodian NGOs involved in developing and distributing voter education materials nationwide..
We did a cross-tabulation of poster recognition and province and identified 14 posters where significant associations existed (#136, 138, 139, 148, 149, 151, 152, 153, 154, 156, 158, 160, 170, 171). We were then able to rank provinces by the rate of greater than expected frequency of poster recognition. Phnom Penh was first (with all 14 posters recognized at a rate greater than expected), followed by Kampong Thom (11), Kratie (8) and Siem Reap (7). Then came Kampong Cham (5), Takeo (5), Banteay Meanchey (5), and Battambang (4), followed by Prey Veng (1), Pursat (1) and Kampong Speu (1).
Without a
knowledge of the size of printing runs for posters and the distribution
strategies of the poster producers we are unable to determine whether our
findings are due to a greater saturation of poster coverage in the high scoring
provinces, or due to a greater than usual sensitivity of respondents in those
provinces to political information.
ANNEX
Abbreviations
|
CAS |
Center for Advanced Study |
|
CIHR |
Cambodian Institute of Human Rights |
|
COFFEL |
Coalition for Free and Fair Elections |
|
COMFREL |
Committee for Free and Fair Elections |
|
CSD |
Center for Social Development |
|
EU |
European Union |
|
KWVC |
Khmer Women’s Voice Center |
|
NEC |
National Elections Commission |
|
NICFEC |
Neutral and Impartial Committee for Free Elections in Cambodia |
|
UNDP |
United Nations Development Programme |
|
WMC |
Women’s Media Center of Cambodia |
Provinces
and Districts Selected for the Impact Survey
|
Kompong Spueu |
Srok Samraong Tong Srok Chbar Mon |
|
Kandal |
Srok Kandal Stueng Srok Mukh Kampul Srok S’ang |
|
Kompong Chhnang |
Srok Rolea B’ier Srok Kompong Tralach |
|
Takaev |
Srok Bati Srok Tram Kak Srok Somraong Srok Treang |
|
Svay Rieng |
Srok Svay Rieng Srok Svay Teab Srok Svay Chrum |
|
Prey Veaeng |
Srok Kampong Trabaek Srok Preah Sdach Srok Peam Ro Srok Prey Veaeng |
|
Phnom Penh |
Khan Doun Penh Khan Prampir Meakkakra Khan Toul Kouk Khan Chamkar Mon |
|
Pousat |
Srok Krakor Srok Kandieng Srok Sampov Meas Srok Bakan |
|
Bat Dambang |
Srok Svay Pao Srok Banan Srok Aek Phnum |
|
Banteay Mean Chey |
Srok Serei Saophoan Srok Ou Chrov Srok Monkol Borei |
|
Kampot |
Srok Kampot Srok Dang Tong Srok Chhuk |
|
Krong Preah Sihanouk |
Khan Prey Nob Khan Mittakpheap |
|
Kaoh Kong |
Srok Srae Ambel |
|
Kompong Cham |
Srok Kang Meas Srok Prey Chhor Srok Tboung Khmum Srok Kompong Cham Srok Batheay Srok Kompong Siem Srok Cheung Prey |
|
Kracheh |
Srok Sambour Srok Kracheh |
|
Kompong Thum |
Srok Baray Srok Stueng Saen Srok Santuk |
|
Siem Reab |
Srok Chi Kraeng Srok Soutr Nikom Srok Prasat Bakong Srok Puok Srok Angkor Thum Srok Banteay Srei |