We performed opinion-based sentiment analysis on 2,062 users who tweeted about the Selangor Menteri Besar crisis from August 9th – August 15th 2014. These users were randomly selected based on tweets we collected since the crisis began. Spammers, news agencies and accounts with automated tweets were not included in the sample.

Our goal initially was to gauge public opinion primarily on whether to keep or remove Tan Sri Khalid Ibrahim. However mixed opinions emerged that allowed us to gauge responses to other parties involved e.g. Pakatan Rakyat (PR), Wan Azizah and Anwar Ibrahim. We were also able to gauge support for the State Assembly to be dissolved and state elections called.

Based on this analysis we categorised users as belonging to one of the following categories:

  1. Keep Tan Sri Khalid
  2. Remove Tan Sri Khalid
  3. Uncategorised (No Opinion / Other Opinions)
  4. Mixed – Not Supportive of PR
  5. Mixed – Not Supportive of Wan Azizah/Anwar Ibrahim
  6. Mixed – Not Happy with All Parties/The Crisis
  7. Mixed – Call for State Election

Mixed opinions are from users who are not clearly supporting Tan Sri Khalid staying on or being removed. Some of the reasons given by the Mixed opinion groups are the same as the ‘Keep Tan Sri Khalid’ and ‘Remove Tan Sri Khalid’ groups, the difference being that the Mixed opinion groups had no clear stand.

Uncategorised opinions are from users who have no clear opinion that matched the criteria of the other stated categories. A previous version of this report ran on our Facebook Page with the label ‘No Opinion’ instead of Uncategorised (No Opinion / Other Opinions). This was corrected to avoid confusion.


What follows are the prominent reasons users gave for supporting the stated opinion.

1. Keep Tan Sri Khalid (445 users, 22%)

  • critical of suspected cronyism demands from PKR
  • hopeful that Tan Sri Khalid joins PAS and stays on no matter what
  • not happy with the reasons provided by PKR, or PKR’s actions
  • no proof he has done anything wrong, only allegations
  • don’t understand the rush to remove him
  • Tan Sri Khalid has done good work e.g. increased the cash reserves; developed the state
  • PR has always praised him until now
  • PR should respect Sultan’s decision
  • Tan Sri Khalid should stay on and clear his name
  • Tan Sri Khalid should serve out his term
  • Tan Sri Khalid can work better without political interference
  • Wan Azizah is not a better choice
  • Tan Sri Khalid should not give in to Anwar Ibrahim or PR’s grab for power
  • Praised PAS for standing by Tan Sri Khalid

2. Remove Tan Sri Khalid (209 users, 10%)

  • want him to step down because he is no longer a party member and does not hold majority support in the State Assembly
  • thanked him for his service but want him to step down
  • reluctant for him to step down – the user did not understand or was not happy with PKR’s decision but accepted it
  • majority tone of voice was neutral, grateful or sad
  • (few users) used a negative tone of voice e.g. curse words
  • negative users accused Tan Sri Khalid of betraying the rakyat by staying on
  • (few users) listed reasons given by PKR e.g. the Bank Islam loan settlement; not attending party meetings
  • (very few users) were initially supportive of Tan Sri Khalid but changed their mind after seeing his actions during the week
  • (very few users) asked Tan Sri Khalid to leave because he is too clean for politics

3. Uncategorised (No Opinion / Other Opinions) (988 users, 48%)

  • questioning what Tan Sri Khalid done to deserve this
  • asking questions to learn more, making jokes
  • pity Tan Sri Khalid for what happened to him
  • confused about the reasons for this crisis
  • retweeting news, repeating statements from the news
  • making observations about the crisis, with no clear opinion
  • want the crisis to end fast

4. Mixed – Not Supportive of PR (206 users, 10%)

  • not happy with the reasons provided by PKR, or PKR’s actions
  • questioning PR’s reasons for keeping Tan Sri Khalid as a candidate in GE13, given their admission of knowing he was a problem prior to that
  • not happy about UMNO/BN being blamed for this crisis
  • majority of the blame directed at PKR and Rafizi Ramli. Minority of users blamed PAS for supporting Tan Sri Khalid or delaying their decision
  • don’t understand why PR did not use a motion of no confidence to dismiss MB
  • don’t understand why PKR could not wait for PAS to have their meeting
  • don’t understand the rush to remove Tan Sri Khalid
  • other issues more important e.g. water supply; dengue outbreak

5. Mixed – Not Supportive of DS Wan Azizah as MB or DS Anwar Ibrahim (115 users, 5%)

  • evenly split between not supporting Wan Azizah or not supporting Anwar Ibrahim
  • reasons for not supporting Wan Azizah as MB Candidate:
    a) she is not a Selangorian (main complaint);
    b) perception that she is a proxy for Anwar;
    c) she is inexperienced or ‘not good enough’.
    d) (few users) stated that she was unsuitable because she was a woman
    e) (few users) Azmin Ali seen as a better choice
  • complained about lack of good reasons given by PR on why Wan Azizah is a better candidate than Tan Sri Khalid
  • critical of Anwar Ibrahim’s ambition
  • personally blame Anwar for the crisis and see it as a power grab

6. Mixed – Not Happy With All Parties/The Crisis (65 users, 3%)

  • not happy with actions of all parties concerned
  • water supply issue and dengue issue viewed as more important current issues
  • some users fed up with politics due to this crisis

7. Mixed – Call for State Election (34 users, 2%)

  • these are users who want the State Assembly to be dissolved and state elections called to resolve the situation
  • these users tended to be critical of politicians and wanted the rakyat to resolve it
  • (few users) wanted Tan Sri Khalid to call for state elections so they can punish PKR and DAP

8. Other Observations

The majority of users tweeting about the State Excos were:

  • supportive of TS Khalid’s move to remove them
  • critical of them for not resigning; for staying on after being sacked; or for saying their dismissal is illegal
  • (few users) praised Rodziah Ismail for resigning



As a result of this analysis we can say that the public perceive Tan Sri Khalid as a well-respected people’s representative and not a political party representative.

The reasons given by PKR for sacking him were not explained well or accepted. As a result Tan Sri Khalid’s sacking was seen as unfair and unreasonable.

For example, signing the MoU with the Federal Government on the Langat 2 Project was one of the reasons cited by PKR for his dismissal. However users that mentioned Langat 2 viewed it as a good move to end the water supply problem. The water supply issue was also seen as more important than a change in Menteri Besar.

Pakatan Rakyat’s (primarily PKR’s) actions during this whole crisis were widely criticised, even by users who accepted Tan Sri Khalid’s dismissal. PR was seen as taking action against Tan Sri Khalid in their own interest and not the people’s interest.

As the week progressed a pattern emerged in user timelines:

a) PR publicly speaking about past problems with Tan Sri Khalid served to make more users doubt PR’s leadership. This pushed users to support Tan Sri Khalid further, though some became fed up with politics.

b) When Tan Sri Khalid dismissed the State Excos following their refusal to meet him, users responded with praise for Tan Sri Khalid. They saw it as a bold move and decisive leadership.

c) When the State Excos returned to work the next day, PR earned criticism.

d) PAS was both praised and criticised by different groups of users for supporting Tan Sri Khalid.

It is clear that damage has been done to Pakatan Rakyat’s reputation. Based on the users’ opinions, a good way to undo some of the damage would be to:

a) resolve the crisis as soon as possible;

b) prove the allegations against Tan Sri Khalid;

c) prove that the new Menteri Besar is better than Tan Sri Khalid



To perform opinion-based sentiment analysis we used the following method:

  1. Collect all tweets from users on a given topic for a fixed length of time
  2. A human examines tweets in sequence, on a per-user basis
  3. Based on the examination, determine the final opinion of the user
  4.  Common reasons for support / opposing an issue are noted

The pros and cons of this method are:


  • More accurate measurement of sentiment than standard approaches
  • Offers details on why users oppose or support an issue
  • Not influenced by large volume of tweets because it is based on users


  • Time-consuming to prepare
  • Requires researchers familiar with the language and the issue