Skip to main content

Settling on a Big 5 type

Inconsistent Results

Choosing a SLOAN type is a complicated issue. Due to the inadequacy of the PDB SLOAN system and the lack of consensus regarding the best version of SLOAN it is not obvious how one should convert big 5 scores into a SLOAN code. Before you can even reach this issue however you need to be confident about what your big 5 scores are. Unlike other personality systems, determining your Big 5/SLOAN is not a matter of reading through profiles and deciding which one describes you best. Rather it is a matter of discovering your precise scores (typically represented by percentages) and then converting those scores into a SLOAN code. So how do you work out your scores in the first place?

As stated in previous pages, Big 5 tests are generally presumed to be reliable. The system is built upon test results and the assumption that those test results are meaningful. The reliability of the tests and their superior performance over most other personality tests has also been objectively proven by statistical analysis. But just because the big 5 is a reliable system it doesn't mean that there is a universal standard for how all of these various tests are scored. If two big 5 tests use slightly different questions or grade on a different scale then it is entirely plausible that someone might consistently get a low score for Extraversion on test A and a high score on test B. Big 5 tests should be similar enough that dramatic differences (e.g. below 10% on one test and above 90% on another) will not occur. But there will still be some variation. You will never get exactly the same scores on two different tests. 

If a big five test asks more questions about assertiveness than about thrill seeking then an assertive person can be expected to get a higher score for extraversion on that test than on a test that weights both facets equally. Some tests present their results as percentiles. Rather than giving you your raw scores (e.g. where a person who gives a 5/5 answer on all 24 questions relating to a trait and so scores 120pts for that trait (24 x 5)) these tests take those scores and tell you where you score in relation to a population. So if your score for a trait is 20/100 but this score is lower than 90% of other test takers but higher than 10% of others then the score that this test gives you will be 10% not 20%. Although you scored 20% on the raw scale, you scored in the 10th percentile for that trait. This raises the question of the source of the sample against which your scores are being measured. Which country are they from? What is the sex balance? Is it a representation of a general population or just a representation of the highly self-selecting group of individuals who like to fill out personality tests? The answers to these questions are usually not supplied on the test website and it is probably best, all things considered, to simply take the results at face value. If however you have taken numerous tests and find that one site gives you a radically different score for a trait than all the other tests then it might be reasonable to assume that that test is unreliable.

As well as the tests measuring you differently, there also exists the issue of giving different responses on different days. If a person is in a particularly apathetic mood when they take a test they may opt for lots of middling, "Neither agree nor disagree" answers, whereas on another day they may feel more conviction.

Calculating Averages

So how do we adjust for these variations and inconsistencies in our test results? We may be tempted to figure out which test is the most reliable and just use that result but this is not easy to do and still does not solve the problem of the same person giving different answers on different days. The simplest answer is to take lots of tests and work out the average scores.

To work out a mean average you simply add up all the numbers and then divide the total by the number of data points. If you have taken 4 different tests and your scores for extraversion were 24, 26, 31, 40 then your calculation would look like this.
24+26+31+40 = 121
121/4 = 30.25
So although you had a range of results you now know that the average is about 30%. This figure is based on an array of results so it should soften the effect of tests that are scored in an unorthodox manner or a subject giving inconsistent answers.

A median average is another, possibly simpler calculation. If you have an odd number of data points, then you simply pick the one in the middle and that is your median average. If your scores for openness from 5 tests were 55, 57, 60, 66, 82 then the median is 60. If we look at the previous example, which used an even number of data points (24, 26, 31, 40) here we would take the two middle numbers (26 and 31) and use the figure that is exactly halfway between the two. 
31-26 = 5
5/2 = 2.5
26+2.5 = 28.5

So for that first set of numbers (Extraversion) the mean is 30.25 and the median is 28.5.
For the second set (Openness) the mean is 64 and the median is 60.
There are advantages and disadvantages to using one type of average over the other, which I will leave for the reader to ponder.

To reduce the effect of different tests being scored differently, take several different tests and work out the average. To reduce the effect of giving inconsistent answers to these tests, take the same test 2 or 3 times (on different days) and work out the average.

Different Scales

This method shows you how to work out the average across a set of tests that all use the same scoring scale (0-100) but not every test uses the same scale. Before you can work out your averages you first have to ensure that all the data points are on the same scale. BigFive-Test.com scores each trait on a scale of 24-120. On this scale the mid point is 72. To average out results from other tests (that use a 0-100 scale) with a 72 from BigFive-Test.com would be an error as it would give the impression that your score for that trait is higher than it really is. To convert your score for a site like this to a percentage you need to know the highest and lowest possible score on that site. Subtract the lowest possible score from both your score and the highest possible score and then divide the smaller number by the bigger number:
72-24 = 48.
120-24 = 96.
48/96 = 0.5 or 50%.
This score of 50 can now be added to your other figures without skewing the results.

Another issue to watch out for is that some tests score some traits on an inverse scale. For example, if a test describes Neuroticism as "Emotional Stability" this means that the figure is the inverse of the neuroticism scale. Emotional stability is the opposite of neuroticism so a high score for that "trait" would mean a low score for neuroticism. A 90% result for "Emotional Stability" should be recorded as a 10% result for Neuroticism.

Why This is Worth Doing

The reason for all of this effort is that it will eventually result in an extremely stable set of figures that won't fluctuate much with new test results. For example, let's go back to our results for Openness: 55, 57, 60, 66, 82. This suggests a moderately high result for this trait. Now let's imagine we've just taken a test that has returned a surprisingly low result for openness: 32. Let's look at how adding this new figure to our averages has changed our presumed overall score for this trait:
Before: Mean = 64. Median = 60.
After: Mean = 58.7. Median = 58.5.
The averages have changed a bit but not much. The Median has barely changed at all. Because the number of data points increased from 5 to 6, the proportion of this increase is minimal and so too is the change that occurs in our average. If you had simply taken 2 tests, one that returned the result of 82 and one that returned the result of 32 you would be left feeling that you had virtually no idea what your correct Openness score might be. Because you have been accumulating results and working out the average the change is small and so your score and your overall self perception may have shifted slightly but this shift is in no way equivalent to the kind of identity crisis that occurs when MBTI users worry that they might have totally mis-typed themselves. Although it may seem arduous and nerdy, this approach to collecting and calculating big five scores is worthwhile because it results in the sense that you can be confident about your personality profile; that you aren't way off the mark. This feeling of confidence in one's typology seems to be a surprisingly rare phenomenon in other systems.

None of this solves the problem of how best to represent your scores with a SLOAN code, but at least you have your scores. These scores after all are the real big 5 "types" and will tell you much of what you need to know in terms of your general personality.

Written and maintained by PDB users for PDB users.
No Comments
Back to top