According to latest opinion polls, 27 to 42% of voters have not yet chosen what (the) candidate (s) they would vote April 22. If you are in this case, here is a small inventory (alphabetically) tools available on the Internet that can help you determine which candidate you are closest in terms of its program.
Test World : graphically successful, at least for the rest .
Graphically very successful, but not the most convincing in its methodology. For each of the 19 themes, one must choose between three or four proposals (including sometimes but not always a box does not know).
My vote for me : complete and fairly solid. Moved by
sitoyen.fr and Sciences-Po Lyon from the application VoteMatch developed by the Netherlands Institute for Political Participation. You must answer 35 questions with "I agree", "I disagree" or "I do not know," which helps you determine your affinity with candidates.
Note: ability to define the issues we consider most important (thereby increasing their impact on the final result).
Polimètre : the most rigorous methodology, but a little long
Designed by two scholars, Paul Antoine Chevalier (ENS Cachan) and Lionel Page (University of Westminster, London) in collaboration with a team of RTL and Debate site 2007.fr. You must answer 30 questions in five terms (from strongly agree to strongly disagree) and we can choose the more important issue in the final result. This is by far the most methodologically rigorous even if we can challenge some options or the formulation of questions (see the methodology note and some discussions around Polimètre HERE).
Politest : to be on the left-right scale.
Directed by two former sciences-po, it offers 12 themes with 3-5 proposals each time.
can then choose the axis that gives priority (economic, social and identity of France). The test indicates the party of which one is closest based on the comparison between the position of the party and the respondent on a scale to 9 degrees.
To who to vote : the simplest, but a little too simple.
In version 1, the answers to questions gave 5 stars to candidates and whoever got the most stars was the candidate.
In the current version, three sets of 5 questions help determine a candidate who could match our expectations. We can then meet 10 other questions measuring the degree of proximity.
What candidat.com : not bad but too eclectic.
In partnership with the Dauphiné Libéré, the answer 25 questions (plus 6 issues tastes and values). Then obtained scores (in%) similarity with each of the 12 candidates.
Vote More : the funniest is the comparator
proposed by Canal Plus. By answering 10 questions, you will know if you are fairly close to Ségolène Bovancenot Sarkopen or François, or another crétaure unlikely. Superb graphics. Too serious not to.
How does it work?
calculated results to these tests? Two types of methods can be used:
- The tests points: they are similar to quizzes or tests found in magazines (those which crosses add triangles or rounds to determine his psychological profile). Each answer corresponds to a candidate (s) and is worth one point and we add the points to determine which candidate you are closest. Refine by giving it little more or less points to responses based on their importance.
- Tests based on the calculation of the distance : there is a little more complicated and it joins a stream of analysis of the vote, known as spatial analysis of voting.
Basically, this is how it works. It defines a political space using n dimensions (each dimension being a problem or an issue of the election). It is then each candidate in this political space with a point, which depends on its position on each dimension deductions (items that can be represented by a ladder, for example from 1 to 5). Thus, on a two-dimensional space, if the candidate's position on the first 3 and 5 on the second, he will be in space at coordinates (3.5). Of course, it becomes more realistic (but also much more complex view) when considering a dozen dimensions.
Similarly, we can define the position of each voter in the same space policy according to their preferences on each dimension (or problems). Finally, using an algorithm, calculates the distance between the voter and each candidate and determine which candidate the voter is the less remote.
Problems these tests
To design a good test, we must first identify issues of concern to many voters, then collect the candidates' positions on these issues and finally write good questions. Now each of these steps is problematic.
- It is not easy to find the right mix of questions. A test can give too much importance to social issues and not enough on economic issues. However, some tests used to weight the questions according to topics of most concern to respondents, and so give them a greater impact on the outcome.
- Programs candidates are not necessarily all available (remember that Nicolas Sarkozy has been made public until late March) and do not always know the candidate's position on a particular issue. Ideally, we should ask each candidate to complete the questionnaire is being proposed to voters. But the teams of candidates are solicited so that this is the tour de force. ( For the record, in 2001, we achieved with a team of students from Sciences-Po comparator programs of candidates for municipal elections in Paris. But we had great trouble getting their programs even playing of all our relationships. )
- Another difficulty of writing "right" questions, ie questions understandable by all, neither too long nor ambiguous. In this regard, several tests inventoried here are not entirely satisfactory. Questions can be sometimes be understood in several ways, some are sometimes formulated in a negative form that blurs the question as one of the tests, the questions are too long (making it difficult to understand the choices available).
Moreover, there is the method of treatment responses. Systems that add points (see previous section) are fairly rudimentary because they assume that proximity to a candidate depends on the number of issues where we agree with him. Now this of course depends on the type and number of questions that were posed.
Methodologically, the tests based on the calculation of the distance are more satisfactory, especially if they can take into account the intensity of preferences for a particular response. But technically, they are more complex to design and manage.
One regret to complete: some tests have clearly and completely above their methodology.
What's the point?
These tests were primarily a heuristic function. They may make you think about the candidates and maybe even your own values. We can thus think of left and noticed by a test that is close to right-wing candidates (or vice versa). These tests are sometimes invite to Enquiries of the games on a candidate that is unfamiliar or being neglected.
These tests do not say of course that you MUST vote, but who you should vote if you behave like a rational voter comparing all programs of candidates based on personal preferences.
Except we're not necessarily very rational decisions in our elections. We also respond to family or social influences. Our vote does not only depend on the political projects of the candidates but also their image. Our vote also results from emotional reflexes and emotional responses. And this dimension, rather difficult to measure, is absent from tests presented here.
FYI: testing the most viewed by the readers of this blog from April 15 to 22
My vote for me: 739 clicks
Polimètre: 539 clicks
More Vote: 463 clicks
Test World: 446 Who clicks
vote: 368 clicks
Politest: 288 clicks
What candidat.com: 276 clicks
0 comments:
Post a Comment