Opinion polls must be conducted properly in order to yield results that reflect reality, and this is achieved mainly by ensuring that the people chosen to participate are a microcosm that is representative of the general population.
Merdeka Center programme director Ibrahim Suffian said this after five media organisations commissioned a survey on public attitudes towards Prime Minister Anwar Ibrahim’s administration.
The survey, conducted by O2 Research Malaysia Sdn Bhd, was criticised for allegedly being methodologically flawed and only reflected the opinions of the media groups’ readership rather than the Malaysian electorate.
Among other issues, the survey had 35,077 respondents of whom 81 percent were male, compared to the census figure of 52.3 percent.
Ibrahim said if a survey is done correctly, even a survey of just 1,200 respondents can represent Malaysia’s 21 million voter base.
“Online surveys based on individuals answering questions without being selected through proper methods are usually unscientific and their results cannot represent the population of a country.
“In summary, conducting surveys is not an easy task. Recently, even surveys conducted by independent bodies in developed countries sometimes miss their mark and fail to predict outcomes such as election results.
“This usually revolves around the accuracy of obtaining respondents that represent all groups.
“In Malaysia, we face another challenge where many citizens do not want to share which party they will vote for because they were slow in arriving at their decision or they distrust the researchers.
“Therefore, it is important for pollsters to carry out surveys that meet scientific standards so that the results obtained reflect the actual situation.
“This not only promotes a culture of constructive criticism in the context of democracy in Malaysia, but it also helps to increase public trust in the country’s survey industry,” he said yesterday.
Ibrahim’s statement is translated from Malay in full below:
Recently, five local media organisations have conducted a survey to examine public opinion on the first 100 days of the unity government led by Anwar Ibrahim.
This is an excellent effort by the media in studying and subsequently reporting on how the people evaluate the progress and implementation of policies by the newly formed government after the 15th general election in November last year.
The media group's aim, which includes audiences in various languages, in launching the survey can provide a positive value addition to the news reporting process and provide context to the public's views, attitudes, and level of knowledge on current issues and their opinion on the administration of the country.
It also provides additional quantitative (measurable) information to supplement and provide a basis for any news or reports produced by the media.
However, the data that forms the report and news must be produced correctly and used responsibly.
At present, there are various surveys conducted to highlight issues or needs, usually, these surveys are conducted online, on social media platforms, or at shopping centres.
The main thing that distinguishes whether a survey is "scientific" or not is how the researchers choose who will answer the survey questions.
In a "non-scientific" survey, individuals answer the questionnaire without being selected, and their responses are received without being filtered to match the demography of the population being studied.
There are several characteristics that form the basis of a scientific survey, with the three main ones being how respondents are chosen, where they are from, and who answers the survey.
Selection of respondents
Every survey must balance the need for enough responses to provide high accuracy - accuracy being dependent on the number of respondents selected in a way that can represent voters in this country - and do so at a reasonable cost.
In a scientific survey, respondents are chosen using statistical methods. Respondents’ selection must be random, where each individual has the same probability of being chosen.
This means that each respondent must represent a group or segment in the country, which corresponds to the distribution of sex, age, place of residence (whether urban or rural), household income, and employment sector.
If only individuals who visit certain websites have the opportunity to answer the questions, this will bias the study results because those who don't use the website do not have the opportunity to answer the questionnaire.
The extent to which the sample of respondents represents the population is what affects the accuracy of a survey. If selected correctly, a survey with only 1,200 respondents can represent 21 million voters in Malaysia.
Composition and location of respondents
If the respondents answering the questions consist of customers of a particular product or media portal, this means that the survey findings represent customers of that product or portal - it cannot be interpreted as representing the entire population of a country.
It is also necessary to consider where the respondents are located, whether they live in urban or rural areas. This is because there are differences in local needs and issues that can influence their views on the administration of the country.
Interviewing respondents
There are three main methods of interviewing a respondent: face-to-face, via a telephone call, or online.
All these methods have their own strengths if carried out scientifically - that is, through a random and representative sampling of the population.
Online surveys based on individuals answering questions without being selected through proper methods are usually unscientific and their results cannot represent the population of a country.
In summary, conducting surveys is not an easy task.
Recently, even surveys conducted by independent bodies in developed countries sometimes miss their mark and fail to predict outcomes such as election results. This usually revolves around the accuracy of obtaining respondents that represent all groups.
In Malaysia, we face another challenge where many citizens do not want to share which party they will vote for because they were slow in arriving at their decision or they distrust the researchers.
Therefore, it is important for pollsters to carry out surveys that meet scientific standards so that the results obtained reflect the actual situation.
This not only promotes a culture of constructive criticism in the context of democracy in Malaysia, but it also helps to increase public trust in the country’s survey industry in the country. - Mkini
No comments:
Post a Comment
Note: Only a member of this blog may post a comment.