WIXLIB

Table 2. Major Sources of Variability on theGEN, VER, MATH, and HSPER Meansfor the MBTI Scales and SEX.VERaMATHaHSPERaFemales (42.6)Males (45.7)Females (.71)Males (.59)VariableGENaSEXFemales (2.42)Males (2.31)J-PJs (2.45)Ps (2.26)E-IIs (2.50)Es (2.30)Is (41.5)Es (39.8)Is (45.3)Es (43.1)Is (.69)Es (.64)SoNNs (2.46)Ss (2.32)Ns (43.9)Ss (38.0)Ns (46.0)Ss (42.5)Ns (.68)Ss (.64)T-FMale F (2.38)Male T (2.27)(No differencefor Females.)Js (.68)Ps (.63)Male F (.62)Male T (.56)(No differencefor Females.)aMeans are shown in parentheses. Data have been treated as in Table 1.In fact, if the J-P scale was examined by itselfinstead of in conjunction with the otherscales, the VER mean of the Ps was higherthan the mean of the Js.Relationship Between GEN and VER, MATH,HSPER, the MBTI, and SEX. A breakdown ofthe deviations of the GEN means from theoverall GEN mean (GENDEV) for the 16 maleand female MBTI types is shown in Table 3.The 16 types are grouped by combinations ofE-I and SoN: ES, IS, EN, and IN. These groupings reflect the variation in the academicpreparation of the different MBTI groups,where the ES students were least wellprepared and the IN students were bestprepared.The unique portions (identified throughregression analyses) of the GENDEV variation that could be attributed to the VER andMATH scores (SAT), HSPER, and the MBTIcharacteristics (PER) also are shown in Table3. The SAT portion was a computed sum froma VER weight of .017 and a MATH weight of.016; the standardized weights were .192 and.210, respectively. The HSPER weight was1.268, a standardized weight of .367. Thesestandardized weights indicate that the unique contributions of MATH and VER inpredicting GEN were nearly equal, while theunique contribution of HSPER wassomewhat larger.The values for PER reflect the unique contributions due to the MBTI characteristicsthat could not be attributed to VER, MATH,and HSPER. Statistically significant contributions of the MBTI characteristics in accounting for GEN were (a) among the ES, IS,EN, and IN groups (F (3, 2697) 3.49, p . 01)and (b) between the Js and Ps for each of thefour groups: ES (t 3.10, p . OO ), IS (t 5.23p . OO ), EN (t 3.21, p . OO ) and IN(t 2.82, p . OO ). All four the J-P differences indicated that the Js achieved morethan the Ps, but the J-P difference was largerfor the IS group than for the other threegroups (t 2.11, p . 04). Unquestionably, themost consistent difference in achievementwas between the Js and Ps.Differences between the Ts and Fs alsowere included in computing the PER values.The difference between the Ts and Fs wasneither significant for the total group norsignificant for any of the ES, IS, EN, or INgroups. However, Ts tended to achieve morethan the Fs for the ES, IS, and EN groups, butless than the Fs for the IN group (t -2.23,p . 03). This difference, in conjunction withthe Keirsey and Bates assertion about thedifferential achievement of Ts and Fs in dif- Journal of Psychological Type Page 29

Table 3. Deviations of GEN from the Overall Mean (GENDEV) and the Portionsof the Deviations Attributable to VER, MATH, (ACH),HSPER and the MBTI (PER)TypeGENDEVbMale8ACHSAT HSPERPERGENDEVbFemale8ACHSAT 80068344422181819122114180809-091801ferent subject matter areas, made it advisable to retain the T-F scale when achievement in the different subject matter areaswas examined.The data from Table 3 are summarized inTable 4. Essentially, this summary confirmswhat would be expected from the Simplecomparisons of the means that were shownin Table 2.1. Females achieved more than males, andmost of this difference could be predictedfrom females achieving more in high schoolthan males; that is, differences in SAT scoresand/or MBTI characteristics were not factorsin their differential achievement.2. The largest portion of the variation in theachievement of the ES, IS, EN, and IN groupscould be attributed to corresponding differences in these groups' SAT and HSPERscores. The SoN difference was the majorMBTI characteristic associated with the SATaccounting for the differences in achievement, while the E-I difference was the majorMBTI characteristic associated with HSPERaccounting for the differences in achievement. The difference in the HSPER of the Esand Is was particularly interesting, as the IsPage 30Journal of Psychological Typealso achieved more than Es in college, evenafter their difference in the HSPER was considered. That is, whatever behaviors of the Isthat led to their superior performance in highschool also led to a superior performance incollege.3. The largest independent contribution(magnitude of an achievement difference) ofan MBTI characteristic in accounting for college achievement was due to Js achievingmore than Ps. Js also achieved more than Psin high school, but the SAT scores were nothigher for the Js than the Ps. Thus, as it wasobserved for the superior college performance of the Is, whatever behaviors of the Jsthat led to their superior high school performance also led to a superior performance incollege. However, a difference between thecontributions of the E-I and J-Pcharacteristics (other than their relativemagnitudes) is that the SAT scores of the Iswere higher than the Es, while the SATscores of the Js were not higher than the Ps.Therefore, it appears that the J-P scale is indicative of the personality characteristic thatis most uniquely associated with college instructors' evaluation of achievement.

Table 4. Summary of the SAT (VER and MATH) HSPER, and PER (MBTI)Influences on the Overall Achievement 127350005E-I/S-N:J-P, 103-04170608IN1817130009aValues have been rounded to two places and multiplied by 100.4. There were no consistent differences forthe comparison of the Ts and Fs on either theSAT or HSPER. As previously noted, the Fstended to achieve more than the Ts for theacademically best prepared IN group, whileTs tended to achieve more than Fs for theother three groups.Overall, these data indicate that combinations of the E-I and SoN scales (ES, IS, EN,and IN) were the MBTI characteristics thatwere most associated with achievement thatcould be predicted from the SAT and HSPER,while combinations of the E-I and J-P scales(EP, IP, EJ, and IJ) were the MBTIcharacteristics that were most associatedwith achievement that could not bepredicted from the SAT and HSPER. On theaverage, IJs achieved about .11 greater thanwhat would be expected from the SAT andHSPER, and the EPs achieved about .11 lessJournal of Psychological TypePage 31

than expected.The difference of .22 in the achievement ofthe IJs and EPs ( .11 versus -.11) corresponds to a difference of about 130 pOintson SAT scores (where the theoretical mean is500 with a standard deviation of 100) oneither the VER or MATH tests, or to a difference in HSPER of about 18 percentilepoints. Of course, such a difference on onemeasure and not the other two measureswould be unlikely because of the positivecorrelations among VER, MATH, and HSPER.The .22 difference in the achievement of theIJs and EPs more reasonably corresponds todifferences of about one-half of a standarddeviation on all three measures: 40 points onthe VER and MATH tests, and 10 percentilepoints on HSPER.Finding the IJ achievement to be thehighest and the EP achievement to be thelowest (after controlling for VER, MATH, andHSPER) is not surprising. Essentially, thefinding concurs with the common sense notion that students who live in a planned,systematic manner and who relate easily tothe inner world of ideas (and perhaps, areless distracted by others) achieve more thanstudents who live in a spontaneous mannerand who relate better to (and enjoy) interacting with others. However, it is enlightening that the MBTI identifies thesecharacteristics, particularly after the othermajor contributors to (or predictors of)achievement have been controlled.It might be inappropriate to extend the interpretation of the data beyond what hasalready been summarized, but there was oneinteresting trend that should be noted for theTs and Fs. The tendency of the Fs to achievemore than the Ts for the INs, as opposed toTs achieving more than Fs for the other threegroups, was due mostly to the males. Essentially, the male INTs, and particularly themale INTPs, had the SAT scores to predictabove average achievement in both highschool and college, but relative to the scoresthese males did not achieve well in either.This suggests that there may be an incongruence between the male INT personality and the formal education process, whichmay indicate that these males represent apotentially underdeveloped studentresource. The variety of subject matter areasand instructional methods involved in thegeneral studies program precludes a definition of "formal education," but the descripPage 32Journal of Psychological Typetions of the 16 types (Myers, 1980) suggestwhy the INT males might not do as well asthe INF males: INTs are skeptical and independent, and they have sharply defined interests, while the INFs, who are equallybright, are full of enthusiasm and put theirbest efforts into their work. SUCCinctlystated, both the male INTs and INFs shouldexperience little difficulty in courses; but theINTs may do very well only in areas that interest them, while the INFs may do well in agreater number of different areas.Subject Matter Areas. The analysis of thesubject areas determined if the (a) VER,MATH, and HSPER regression weights, and(b) differences attributable to the J-P and T-Fscales, SEX, and ES, IS, EN, and IN groupings that were obtained in the analysis ofGEN, could account for the relationships ofthese variables with achievement in the individual subject areas. The courses used inthe subject areas are shown in Table 5. Themean grade of the courses included in anarea was used as the indicator of achievement. While there were several differencesdue to SEX and MBTI characteristics, differences in the VER, MATH, and HSPERweights accounted for the most variabilityamong the areas.SEX and MBTI Characteristics There wereseven subject areas in which there was a difference between the achievement of malesand females. After controlling for VER,MATH, and HSPER, the males achieved morein business (BUS), biology (BIO), astronomyphysics (ASTPHY), political science (POLS),and history (HIST) courses, while femalesachieved more in the English (ENG) andphilosophy-religion (PHILREL) courses.These differences appear to conform to thegender vocational stereotypes, where malesare relatively more represented in practical,concrete and/or scientific vocations, whilefemales are more represented in aesthetictype vocations.There were four subject areas in which therelationship between achievement and theES, IS, EN, and IN groupings differed fromthe relationship that was observed for GEN.In interpreting these data, as well as assessing possible trends that included other areas,the focus was the relative achievement of theESs and the INs; these represented, respectively, students who apparently were least

well prepared for academic achievement andstudents who apparently were best prepared.The ESs did relatively better than the INs inthe (a) music, art, and theatre history (ARTS),(b) business (BUS), and (c) economics(ECON) courses, while the INs did relativelybetter in the anthropology (ANTH) courses. Apossible implication of these findings is thatESs may experience more success incourses that are more factual, practical, concrete and/or applied than in courses that areabstract and theoretical, while the reversemight be true for the INs. Support for such aninterpretation was provided by the tendencyof the ESs to do relatively better in thespeech and journalism (COM) and thepolitical science (POLS) courses, while INsdid relatively better in the astronomy-physics(ASTPHY), psychological science (PSY),biology (BIO), and natural resources (NR)courses.There were three subject areas in whichthe differences between the Js and Ps variedfrom the relationships that were observed forGEN: the classical culture (CC), history(HIST), and mathematics (MATH) courses.However, Ps achieved more than Js for onlythe (a) ESs in the mathematics courses, (b)INs in the classical culture courses, and (c)ENs and INs in the history courses. In addition to these four differences that favored thePs, there were only four of the remaining 68J-P comparisons that favored the Ps. Thus,although many of the differences were quitesmall, the finding for GEN in which Js achieved more than the Ps was very consistent forthe different subject areas.Astronomy-physics (ASTPHY) was the onlysubject area in which the relationship for theT-F scale differed from those observed forGEN: Ts achieved more than the Fs for theISs and INs. Additionally, there were no apparent trends in that data with regard to theT-F scale. Thus, in conjunction with findingonly non-significant relationships betweenthe T-F scale and GEN, it appears that the T-Fpreference is not an important function instudents' achievement-at least, not inthose survey type courses that are found inthe general studies curriculum.Course Difficulty and Correlations. The sub-ject area difficulty (deviation of the subjectarea mean from the GEN), and the VER,MATH, and HSPER correlations and standardized regression weights are shown inTable 5. The order of presentation of theareas, from top to bottom, reflects the uniqueverbal (VER) and quantitative (MATH) aptitude that might be inherent to achieving inthe areas. The uniqueness is indicated by therelative magnitudes of the standardized VERand MATH weights. In the extreme cases forPHILREL and MATHGRD, VER containsunique information about the PHILRELachievement (.23), while MATH containsunique information about achievement inmathematics courses (.43). Negative andsmall positive weights should not be interpreted as indicating either a negative relationship or a trivial, positive relationship.Clearly, the simple correlations indicatemoderate positive relationships of VER,MATH, and HSPER with achievement in allareas. The regression weights indicate aunique characteristic beyond the commonaptitude-achievement component whichunderlies the scores on all three of the VER,MATH, and HSPER measures.Some observations about the data contained in Table 5 are as follows:1. Too much reliance on the VER score-as apredictor of students' achievement mayneglect other readily available informationabout a student's academic potential. Whilethe VER simple correlations are larger thanthe MATH correlations in certain areas(PHILREL through ANT H), there is virtuallyno difference in other areas (COM throughCC), and in some areas VER is not as good apredictor as MATH (GEO through MATH). Infact, VER does not predict GEN as well asMATH (.46 versus .49). Additionally, HSPERwas a better predictor than VER in 13 of the18 subject areas that are listed.2. HSPER, which is sometimes treated assupplementary information about a student,is the most consistent predictor of thestudent's achievement. In 12 of the 18 subject areas, the HSPER correlation equalled orwas larger than both of the VER and MATHcorrelations. In 14 of the areas, the HSPERregression weight was larger than the VERand MATH weights. There also was lessvariation among the HSPER weights thanamong either the VER or MATH weights.Thus, independent of the common aptitudeachievement component underlying the VER,MATH, and HSPER scores and the uniqueVER and MATH components, HSPER provides additional information about theachievement of students. A cognitiveJournal of Psychological TypePage 33

Table 5. Subject Matter Area Difficulties, Correlations,Regression Weights, and Coursesa .VERDeviationsof 01940aValues have been rounded to two decimal places and multiplied by 100.bMultiple correlation coefficients.cThese values were computed without consideration of SEX and the MBTI.psychologist might interpret HSPER as being an indicator of motivation, of whichachievement has been and will be an outcome. A behavioral psychologist might interpret the result for HSPER as meaning thatthe best predictor of future behavior is thepast behavior. Certainly, a degree ofbackground knowledge in an area is required,but HSPER generally adds to that measure ofknowledge in predicting subsequentachievement.3. There were no consistent trends in thedata regarding subject area difficulties, but itis quite -clear that there was considerablevariation in the difficulty of the subject areas.However, an interesting question is why thePage 34Journal of Psychological Typethree areas associated with the most uniqueMATH aptitude (BUS, ECON, and MATHGRD)were also the three most difficult areas. Conversely, three of the four areas associatedwith the most unique VER aptitude(PHILREL, HSC, and ARTS) were among theleast difficult areas.Ideally, data such as presented in thisreport might be used to advise studentsregarding individual courses and majors.However, while it is important for individualsworking with students to be aware of thereported trends, there is sufficient "noise" inthe data to advise one to use the datajudiciously. Additionally, certain results arenearly paradoxical. For example, ESs are less

well prepared academically, but relative toother courses they do well in businesscourses, which happen to be among themost difficult courses-that is, they may dorelatively well but still obtain a low grade.One source of the "noise" is that differences in factors such as class size, instructional methods, and instructor gradingsurely influenced the results. A secondsource is that there was considerablevariability in the results for sections of thesame course-for example, averagegrades-as well as among courses that weregrouped into subject areas. Thirdly, resultsfor subject areas might be different incourses for majors than they were in thesurvey courses. But, even in light of this"noise," there were conceptually meaningfultrends that have practical implications. Forexample we would not advise a low-VER, ESP (Extraverted, Sensing, Perceiving) studentto take both PSY and ANTH during the samequarter. At the other extreme, we would notoverload even the best academicallyprepared student-for example, enrolling ahigh-MATH INJ (Introverted, Intuitive, Judging) student in the three most difficultcourses, MATH, ECON, and BUS.DiscussionSome cautions must be expres

The Myers-Briggs Type Indicator and First-Year College Achievement: A Look Beyond Aptitude Test Results K. Terry Schurr Virgil E. Ruble Ball State University A thorough study of an entire entering col/ege class shows that Is and Ns tend to have both higher