IPS post election survey 2015

I refer to Institute of Policy Studies’ post election conference slides: Post election survey 2015 (S1_GK_POPS-8_GE2015_291015_Web1.pdf).

Response rate

The survey response rate of 24.5% (page 10) is rather low. This may compromise the randomness of the survey.

Percentage importance VS mean score

IPS ranks “need for efficient government” as the top issue of 2015 because it bears the highest mean score (page 21).

1 - need for efficient government

However, there is no change to the mean score for “need for efficient government” between 2011 and 2015, so it can’t explain the sharp change in voter sentiments between 2011 and 2015. To explain the sharp change in voter sentiments, we need to look for larger change in numbers.

Take the issue of “fairness of government policy” for example. The mean score rose marginally from 4.1 to 4.3 between 2011 and 2015, not substantial enough to show up on the radar screen. But the percentage of voters who viewed this issue as important or very important rose sharply from 81% to 94%. This sticks out more prominently.

2 - fairness of government policy

It is important to know what to look for in the data that we have.

Issues with larger increases in percentage respondents viewing it as important

The issues that have seen much greater increase in number of respondents viewing it as important or very important compared to 2011 were (1) fairness of government policy, (2) work of former MP, (3) foreigners & immigration, (4) neighbourhood facilities and (5) upgrading:

3 - work of former MP
4 - upgrading

Amongst these five issues, it can be presumed that the issue of foreigners and immigration worked against the PAP while the other four issues worked for it. Thus while the opposition had rightly capitalised on the issue of foreigners and immigration, this issue alone could not overturn the combined influence of the other four issues that the PAP was able to bring about by the sheer strength of its financial muscle.

The Pioneer Package may have contributed to the sharp increase in the percentage of respondents giving importance to fairness of government policy. This together with sharp increases in percentage of respondents who placed importance in upgrading and neighbourhood facilities suggests that ultimately it was generous providence that bought PAP success.

Comparing 2006 with 2011

We can similarly compare 2006 and 2011. The issues that experienced sharper increases going from 2006 to 2011 were cost of living, party manifestos, job situation and upgrading, which were roughly the more prominent issues back in 2011.

Candidate characteristics

IPS concluded that the electorate placed greater emphasis on a GE2015 candidate’s honesty, fairness, efficiency and empathy (page 25). However, because many GE2015 candidates were new faces, there would not be sufficient track record to judge them on these attributes. These would therefore be rather useless attributes in explaining electoral choice.

Communication channel

The percentage of voters influenced by TV, internet, grassroots workers, door-to-door visits, friends/family/colleagues increased in GE2015 so all these may have contributed to the election outcome.

Little separated the top three channels of TV, newspapers and the internet. What’s interesting to note is that amongst internet channels, Facebook stood head and shoulders above the rest.

5 - top 5 internet channels

There may be an urgent need for the opposition to expand its Facebook reach towards swing voters.

Party credibility

6 - credibility

All three most credible parties experienced sizeable increases in the percentage of voters viewing them as credible in 2015 but that did not translate to the same electoral success. Perhaps when it comes to credibility, being the second or third most credible party means nothing when compared to the most credible party.

It is unlikely for the opposition to ever match PAP in credibility because it can never be in a position to show that it too can deliver. While a lot of good can come from speeches or manifestos, in the context of practical minded Singaporeans, talk or sales brochures seldom trump action and actual physical providence.

Election system

There is a marked increase in the percentage of voters who perceived election fairness as being important in 2015 and this may have worked to PAP’s favour. PAP may have learnt not to be too blatant in its abuse of the election process to gain acceptance by the masses.

7 - election system is fair

Cluster analysis

IPS categorises voters into three groups:

1) Conservatives (those who support the status quo)
2) Pluralists (those who support political pluralism and change in electoral system)
3) Swing voters (those whose views are mixed)

IPS doesn’t give details on how it decides who supports the status quo and who doesn’t but from its brief description, it’s possible that IPS based its decision on respondents’ views on some mixture of the following issues:

1) Need checks and balances in parliament
2) Need for different views in parliament
3) Important to have elected opposition party members in parliament
4) No need to change election system

The problem with the above criteria is that they don’t specify how much check and balance, how many different views in parliament or how many elected opposition members of parliament. For example in the case of having elected opposition members, it’s possible that a respondent might give a “yes” answer even though in his mind, he is thinking of no more than just a handful of opposition members. From the given answer, IPS would conclude that this respondent is a pluralist (if other answers do not contradict) but in actual fact, this respondent is closer to a conservative than to a pluralist.

It would have been far better if IPS had simply asked respondents if they had switched votes this time. That would be more straight forward and surer in identifying the swing voters.


While IPS flagged differences between various voting groups’ mean scores to various issues as being significantly different, most of them hardly differ by more than 10%. It therefore seems that on average, voters of different groups by age, household income, education, ethnicity, housing type, gender, citizenship status do not differ too greatly from one another in voting pattern. That is, even new citizens do not vote too differently from citizens at birth.

However, as noted at the beginning, mean scores may not as meaningful as percentage respondents in deciphering voting patterns.


The percentage of respondents who viewed an issue as important or very important may be more meaningful than the issue’s mean score in deciphering voter sentiments.

Based on issues that experienced greater increase in number of respondents viewing them as important or very important, the key to PAP’s 2015 electoral success may have been its tremendous power of providence.

Facebook emerged as the runaway champion amongst various internet modes of communication.

It may not be of much use to be the second or third most credible party when up against the most credible party. Opposition parties must start to move decisively beyond talk and manifesto towards physical actualisation of its providence ability to match up to PAP’s credibility.

IPS’ cluster analysis’ categorisation of various voter groups is not very convincing. A simple question of whether the respondent switched votes may have been better.


Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Google+ photo

You are commenting using your Google+ account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )


Connecting to %s

%d bloggers like this: