"The meaning of three socionic criteria, extroversion/introversion, logic/ethic, sensation/intuition is almost the same as in MBTI for E/I, T/F and S/N criteria. There are only slight (but sometimes important) differences. For example, socionics does not associate extroversion with sociability but rather with initiative. Neither socionics associates introversion with intellectual independence. Another difference is that MBTI does not associate control of personal space and imposing of own wishes and will on others with sensing, while socionics does.
The criterion of rationality/irrationality deserves a special remark.
The meaning of "rationality—irrationality" in colloquial language is somewhat different from the original Jung's (and socionic) understanding of this term. The colloquial meaning of this word pair should be attributed rather to another dichotomy, logic-ethic (or T/F in the Myers-Briggs theory).
Although the J/P criterion in the Myers-Briggs theory resembles rationality-irrationality by its description, it's not the same. As it was proved by Gregory Shulman, J/P is not dichotomy (i.e. it does not split the 16 socionic types into 2 equal parts). It is a pseudo-dichotomy that represents two poles with multiple transitional options; in other words, one or two types in socionics may be called "super-J", and another one or two types "super-P", while other types tend more or less to one of these poles. This explains, for example, why ENFP or ENFJ in socionics may both get scored ENFP according to MBTI, etc.
One more remark: according to I.Myers' hypothesis, rationals correcpond to EJ and IP, while irrationals to EP and IJ. However, Isabel Myers herself wrote that her hypothesis somewhat contradicts to the result of her test, i.e., the J/P description does not work well for introverts. On the other hand, statistical researches of parallel testing of people by socionic methods and by Keirsey Temperament Sorter (online version) does not agree with her hypothesis and speaks rather in favor of identification of J/P with rationality/irrationality (although not 100%, as it was said above).
Such nature of J/P criterion sometimes causes funny mismatches. For example, the descriptions of ESFP and ESFJ according to D.Keirsey match well with the corresponding socionic types, while their descriptions according to Kroeger and Thuesen match rather reversely: i.e., Kroeger's ESFP correlates better with the socionic ESFJ, and Kroeger's ESFJ with the socionic ESFP.
If you find any contradiction between the descriptions of socionic types and/or those proposed by MBTI authors, do not hesitate to discuss them in one of the forums (see the top left menu).
Another problem is that MBTI adherents rely too much on verbal testing and only in recent years began to develop methods of objective verification of types. By contrast, while socionists developed a lot of methods of determining types without tests (by interviewing, observation etc.), only in recent years they began to develop good tests (among them—the Multifactor test you can find at the start page of the English section). Twenty and even ten years ago socionists used very primitive tests (like this one), just because they did not believe in tests. It was difficult to believe in tests in the former Soviet Union—living under the ideological pressure, people used to lie or at least hide their real thoughts or to speak about them vaguely, and this can partly explain why good tests appeared so late in socionics. This may also explain the difference between the socionic and the MBTI type descriptions: MBTI describes types through their behavior, while socionics describes types from the viewpoint of their appearance, motivation and style of thinking."