Oftentimes, researchers opt for brief scales to minimize participant burden / to save time. However, increasing the length of a questionnaire also increases the consistency (reliability) of the measurement. Therefore, I would always try to use the full version of the test, especially if the Big Five are of focal interest to you.

Oftentimes, researchers opt for brief scales to minimize participant burden / to save time. However, increasing the length of a questionnaire also increases the consistency (reliability) of the measurement. Therefore, I would always try to use the full version of the test, especially if the Big Five are of focal interest to you.

I would also opt for the longer version. I didn't have the opportunity to use the 44-Item scale, but I have some experience with BFI-10 and the reliability coefficients were really bad for two/three of the five factors. For Agreeableness there was even a negative correlation between the items, therefore one cannot interpret the results meaningfully. I have also seen similar problems with BFI-10 in other studies. The only factor with a good coefficient was Extraversion (Cronbach's Alpha > .80)

Dear Omaya, notice that John Oliver (director of the Berkely Personality Lab), does not reccomend to use the 10-item version unless the assessment is conducted with dire timings or other circumstanves. You may  find additiional information here: http://www.ocf.berkeley.edu/~johnlab/bfi.htm

Somehow in depends on the purpose of the study. For academic exercise, high reliability score is crucial and most of the cases the 44-item would produce more reliability score due to the more variability of responses. Some how, for the HRM practical exercise the shorter version is more likeable by the employees and we can expect more genuine response.

I wouldn't be so sure about the 44-item BFI either. It depends on what you want to achieve. We have used a "home-brewed" 60-item questionnaire called "Short Five" (see attached publication) - which now has a 30-item version too (not yet published but we're currently writing it up) - and is available in several languages. Because the items are longer, the 30-item S5 probably takes a bit longer to answer than the 44-item BFI but there are other criteria where I would bet S5 clearly wins: for example, comparability of the scores with well established questionnaires like NEO PI-R, and availability of subscale scores. The very short (10-item) BFI has its uses -- if your respondents have only one minute or less -- but if they have 5 minutes, I'd recommend against it. And the short (44-item) BFI is still much shorter than the "gold standard" NEO PI-R (240 items) or its "freeware" analog based on IPIP (300 items).

Article The ‘Short Five’ (S5): Measuring personality traits using co...

Although the longer versions are more reliable (or at least can estimate reliability better), I think it depends on the purpose of your study - if personality is the main point, use more than the short BFI. However, if it a part of a larger study, with the aim of controlling for these personality variables, among others, and you do not predict that they will affect the relationship between the main variables in your study (those that are more important to your research), then I believe it is acceptable to use the 10-item version. In this situation, the participants' time and energy should be directed towards the real questions that interest you...

Honestly, even a 44-item measure is quite brief if it includes 5 scales - that's an average of just under 9 items per scale. The 10-item test, if it's useful at all, would be appropriate only for very large-scale studies (samples in the thousands, not hundreds) where lots of different research teams are contributing bits and pieces. If you have a huge sample, you can get away with a shoddy measure. If you have a highly precise measure, you can get away with a small sample.

I'd also recommend looking closely at item content on all short measures that claim high reliabilities. This can be achieved by making the items really similar. But if you are assessing a broad personality trait, that's a problem. Sometimes I have my doubts about the IPIP scales for this reason. As a presented (I forget who) once said at one of the first conferences I attended, it's easy to produce a 5-item depression scale with reliability of .95 or better. Just make every item read "I am depressed."

I have used the BFI 44 item measure in my research. The results regarding reliability and factor analysis were not so encouraging. I definitely warn you to be cautious when using this scale. 

Take a look at the correlation table in my master thesis, "The Influence of Dark Personality on Counterproductive Work Behavior," several scales in the BFI-10 were not reliable. Fortunately, I just had the BFI-10 in my survey to balance positive and negative-worded items since I was studying dark personality, which tends to have mostly negatively-worded items in current measurement.