EMGC Selangor (34 Malay Majority DUN Seats) Survey

THE broad conclusion to analyses of the recent EMGC survey carried out in conjunction with its media partners was that there was going to be fierce competition to win Selangor state and that at this point, it was PH’s to lose. That is to say that while it is widely agreed that PN can win a majority of Malay votes in Selangor, it is unlikely that non-Malay turnout drops to much less than Malay turnout and that because of this PH is likely to retain Selangor although PN will make even more significant inroads compared to the 2022 GE.

However, there are some voices that do not differentiate between reporting and analyses. EMGC as an independent pollster who has worked with major media organisations as a survey data partner for this PRN would like to briefly explain the issues based on this principle we have always followed.

With regards to the data, while we did transparently report the seemingly extraordinary increase in Malay support for PH from an estimated 24% in GE15 (November 2022) to a reported 52% in the survey (July 2023), EMGC did not go on to conclude that this would indeed be the case on voting day. As most pollsters know, the reported data has to be interpreted and put into context by independent media as well as by pundits and political analysts.

EMGC has a responsibility as a pollster to be transparent with regards to its survey processes as well as its survey reporting and analysis, i.e. EMGC reported 52% of Malays said they would vote PH in the upcoming PRN as this was what was recorded from Selangor Malay voters randomly included as part of the survey sample.

When moving from direct reporting to analysis, we would state that like any poll, this survey had a margin of error (Malay sample: 3.6%). There is also incumbency bias i.e. when some Malaysian respondents (10% - 15%) in any survey provide responses to political choice questions more as what they feel they should answer rather than what they want to answer.

When margin of error and incumbency bias is considered in addition to the fact that Malay respondents felt PN would win a much improved 20 out of 56 Selangor DUN Seats in the PRN plus that this survey ended before the official PN Selangor campaign began, a reasonable analysis of reported EMGC survey results data could argue that it is PN that would, in reality, command the majority of Malay votes in Selangor come voting day. Note that this conclusion would constitute an analysis of the poll results and would not just be a basic reporting of the poll results.

Furthermore, on an issue of method and sampling bias, EMGC conducts stratified random polls. This means that it split the survey into 34 DUNs and subsequently split these into demographic strata within each DUN area. We ensured that field agents randomly selected respondents within strata all to achieve the objective of ensuring that the sample is representative of voter demography in that particular DUN.

EMGC survey was indeed able to illustrate a very high degree of committed responses to questions relating to political choice. However, the fact that EMGC managed to achieve this was not because it was better or worse than certain other surveys that could not achieve this, but rather because EMGC’s approach differs from some other pollsters.

Our interviewers do accept non-committal responses on political choice (don’t know, have not decided, confidential etc.) but are trained well to explain and highlight all committed options equally. This approach translates to most survey respondents committing to one political preference or another. In this, it should be noted that EMGC did report that 5% - 6% of respondents stated their political choice was confidential. Finally, as was carried by our media partners, it should be noted that the survey was completed before nomination day. This means that the data reflects the support level before the full-swing formal campaign period began, i.e. before the onslaught of political messaging and campaigning began in full swing.

Selan Rajendan

Director of Research & Analysis, Endeavour – MGC