High and Low SAT Scores: The Difference

While at the University of Georgia in 1984, I was asked by the Georgia State Department of Education to examine student performance in the aftermath of state-wide school integration.  Specifically, I was to consider how well Georgia’s students performed when identified by race, among other issues.  Because of the sensitive, if not volatile nature of this topic at the time, the results of this study were not expected to be widely disseminated.

To assist with the request, I was provided a data tape that included the SAT scores for all of Georgia’s high school seniors, a subset of 33,027 students.  The tape was stored at the University’s campus-wide computer center for use in data processing.  All analysis was performed using the University’s SAS program, and was completed independently of Georgia’s State Department of Education.  The final report of the study was submitted to the State Department of Education in May of 1985.

At the time of the study, most of Georgia’s high school student population consisted of two primary racial groups, white and black, in about a two to one ratio respectively.  Hispanic and Asian students were scattered throughout the state, but were present at that time in much smaller numbers compared with today’s student population.

Table 1 shows the average group SAT scores when separated by race.  Georgia’s average SAT score is represented by 0 on the scale.  Variation above and below this average is in standard deviation units of 200 points.  Positive numbers represent average group SAT scores above the state-wide mean, while negative numbers show group performance below the state’s average.  With this scale two thirds of all scores will fall within plus and minus 200 points of the mean.

Table 1
Georgia SAT Scores by Race (N=33,027)

  As may be seen, race produced above average performance for white and Asian student groups.  For Hispanic, Native American, and black student groups the average scores are below the state’s mean.  Comparing the two primary racial groups in Georgia, the SAT scores of black students, on average, is over a full standard deviation, 200 points below that of white students.  These averages reflect the best and the brightest within each group, those who may choose to attend college.  Statistically this performance discrepancy is equivalent to a black hole in outer space.  It deserves a reasoned explanation.

While the supreme court concluded in Brown vs. Board of Education (1954) that separate schools are not equal, this state-wide study thirty years later suggests that together is not equal.  It seems clear that the supreme court’s focus upon race is a serious diversion of emphasis from choices that might make a difference, to choices that makes no difference at all.  If integration was intended to narrow the gap in performance between white and black students, the first thirty years of the experiment was a dismal failure.  Unfortunately the experiment is continuing much as before with blind faith that a miracle will appear to rescue us all.

In the present case it is clear that the schools in Georgia were certainly failing with a majority of black students who, as a group, have seriously fallen out of the ranks of well educated students.  An appropriate explanation is probably that the black culture does not value, endorse, nor support doing well in school.  At its worst a portion of the black culture actively discourages its members from succeeding in school.  This is a group generalization, but one that does not apply to all individuals within the black student body.

Average scores, as displayed in Table 1, are excellent devices for representing a group, when in fact they obscure the minority report that individual students, like cream, rise to the top in every human endeavor.  As Connant observed, “Each honest calling, each walk of life has its own elite, its own aristocracy based upon excellence of performance”.  This is true of all students, regardless of their racial origin, although rising above a host of negative influences may require unusual character and an iron will that must persist throughout twelve years of schooling.

In an effort to explain the scattered SAT performance, an additional analysis was undertaken to identify what some of the positive and negative influences might be.  At the time of testing, each individual is given a standard questionnaire which elicits information about each student’s circumstances.  Sixty-three biographical questions were asked on the 1984 questionnaire.  From these questions, a number were retained either because of their known educational influence, or because they should be considered as part of an educational environment.   The responses to these additional questions were then included as potential predictors of SAT combined scores within a statistical regression procedure.

Thirteen variables were selected for determining relevance to SAT scores.  They are as follows:
1)        Gender (male / female)
2)        Public / non-public schools
3)        Type of school program with which the student has been associated:  academic, general, career, or other
4)        Size of the student’s high school class
5)        Number of hours of part-time work, if working
6)        Participating in community or church organizations
7)        Level and degree of participation in athletics
8)        Level and degree of activity in school clubs and offices
9)        Race or national origin (American Indian or Alaska Native, Black, Hispanic, Asian, White, Other)
10)      Education completed by father or male guardian
11)       Education completed by mother or female guardian
12)       Number of dependents within the family unit
13)       Approximate income of the parents or family unit

Because of known gender differences in predicting success in school, males and females were processed separately.  To avoid getting lost within the complexities of statistical analysis, the following results are presented based upon significance testing, which compares each individual factor as differentiated from sampling variation.  Table 2 shows the six questions which were related most closely to higher SAT scores among Georgia’s seniors.  The F value is the significance level for each variable, while the Positive or Negative shows whether the variable is directly or inversely related to SAT scores.

Table 2
Male predictors of SAT scores

Variable name F-value Pos or Neg
Academics 596 Positive
Athletics 229 Negative
Father's Educ 140 Positive
Caucasian 106 Positive
Working 77 Negative
Mother's Educ 65 Positive

Table 3
Female predictors of SAT scores

Variable Name F Value Pos or Neg
Academics 723 Positive
Father's Educ 155 Positive
Mother's Educ 113 Positive
Caucasian 69 Positive
Family Income 68 Positive
Working 56 Negative

The Academic Program

        For both males and females, the primary predictor of higher SAT scores is participating in the academic programs in high school.  The academic program, as compared with the general or career tracks, is the one that is chosen by students who expect to continue their education beyond high school.  Rather than taking fundamental courses, which provide minimum requirements for graduation and little else, students in the academic program, particularly in the larger high schools, take courses that provide both depth and detail in many areas.  It may include, for example, classes in chemistry, physics, solid geometry, trigonometry, or a fourth year of English.  Fifty-seven percent of Georgia’s students taking the SAT reported being in their school’s academic track.  The 43 percent who were in a different track did not do so well on the SAT.

This important choice is often obscured by the fact that many courses are sequential, and before a student may take solid geometry, for example, they must complete plane geometry satisfactorily.  Algebra must be taken before trigonometry, while a general mathematics class is rarely taken by students in the academic program.   Another distinction may be word on the street that certain courses are easy grades for which one does not need to apply oneself diligently.  Avoiding actual school work or studying may be the key to selecting high school courses that satisfy the minimum compulsory school graduation requirements rather than participating in the academic program.

This choice is the primary factor in doing well in school for both males and females.  It is the most important single factor that relates to obtaining higher SAT scores among high school seniors.  It is a choice that is made by default early within ones school career, and often is expressed as a keen, positive interest in school subjects.  This choice may be predetermined by the record each student establishes during prior years in school.  The academic program is not alone in its importance in producing good students with higher SAT scores.

High School Athletics

        The second factor within the male regression is participating in athletics.  It is a negative predictor.  This means that athletic participation detracts from doing well on the SAT.  The interpretation may be obvious.   For all males in school, the time spent practicing, preparing for and playing in sports is time not spent preparing for school classes.  Given the attraction sports careers present for black students, the hours spent in athletic activities is time taken away from math, English, and other academic pursuits.

As practice makes perfect, the time devoted to one activity is time taken away from other endeavors.  This time conflict explains some of the result for black males and SAT scores, but it applies for all male students who choose to participate in high school athletics.  It is also a negative factor for female students, but its significance is much farther down the list of influences.  Thirty-nine percent of those taking the SAT reported participating in school athletics.

Education of Father and Mother

        For female and male students the second and third most significant predictor is the educational level of the father.  As a positive influence, the greater the educational achievement of the father, the better the students did on the SAT.  This would seem to identify a climate or cultural factor within the student’s household or out-of-school environment.  Fathers who went farther in school themselves encourage their children to be good students.  While one does not choose ones parents, the household influence becomes an ever present, daily exposure.  Well educated parents offer encouragement and support for their children to do as they did.

For female students, the educational levels of both the father and the mother contribute positively to their daughters’ doing well in school.  These are the second and third most significant predictors for all female students.  For black students in Georgia’s schools in 1984, this predictor was often half missing, as a large number of households had no fathers at all, and the influence of the single mothers did little to compensate for the missing father’s influence.

As both fathers and mothers contribute positively to their children doing well in school, this speaks volumes to broken and single parent homes, and the importance of all parents to support the schools’ programs for their offspring to do their best.

The Culture of Race

        Race is a predictor for both males and females.  Whether it is positive or negative depends upon the race, and whether that race performs above the average or below the average as shown in Table 1.  Minority status does not seem to interfere with Asian student achievement within integrated schools.  Being black is clearly a substantial negative influence for earning high SAT scores.  This factor should not be elevated above the more substantial influences discussed above.  Choosing to compete in the academic program is by far the single most important factor in doing well on the SAT, and any racial component, with the proper family or cultural influences, can be overcome.

Working Part Time

        Part time work is fifth and sixth in the male and female regressions respectively, and is also a negative predictor.  Like athletics, time spent working in the evenings or weekends is time not spent studying, and interferes with doing well in school.  This applies across all races.  If a student works, and also participates in athletic activities, the academic work at school is very likely to suffer.  The greater the number of hours spent in these distractions, the greater will be the damage to doing those things that good students do.  Somewhat surprisingly, 46% of Georgia’s seniors who took the SAT reported working outside the home.

Family Income: The Tail of the Dog

        Finally, estimated family income is a positive predictor, of the six most significant predictors, for female students only.  As a minor factor, it may contribute to the female sensitivity to dress style, grooming, and the personal image presented at school, much of which is supported by the family’s income.  It may fail to rise as a factor for male students because they simply do not place as much credibility upon appearance.  It appears that family income, coming from a wealthy neighborhood, or ones socioeconomic status as a powerful influence simply carries little weight for good students.  Good students may rise above their family’s status in life, and dedication to performing well in school is tied to being poor much as the tail is tied to the dog.  It may have little influence over the choices the dog makes.

The Real Predictors: The Dog

        All of these predictors, singly and in combination, explain only a third of doing well in school, as measured by SAT scores in 1984.  There are no reasons to presume that any of these influences are different today than they were two decades ago, except possibly for the major restrictions on the support of single welfare mothers with children.  In many respects, the primary predictors are not even included within the current study.  What remains, then, is to identify what the other factors may be,  and how they may be managed to improve the performance of students in the schools.

The primary contributors to doing well on the SAT are not included directly as variables in the present study.  Every capable teacher, administrator, and most parents know good students when they see them.  Intelligence is one of the predictors that is not included directly in the above.  Most educators would include the SAT itself as a measure of intelligence, as it represents the cumulative work-product of ones academic experience over a 12 year period.  As such, the interpretation speaks for itself about doing well in school.

One does not choose to participate in the academic programs in high school for remedial training, as the competition among the students is keen.  Honors courses are not intended for the weak of heart, and those students without the proper background quickly find themselves in deep water and unable to swim.  Intelligence is present indirectly in the student regressions, as the genetic link with both parents is a primary factor.  The genetic link remains in the child regardless of the physical presence of the parent in the home, while parents who are absent from the household are powerless to have much influence on a daily basis.

A second major predictor of success is accepting schooling as a cumulative experience.  This is a painful notion when viewed day-by-day for twelve years.  What a student learns on one day often becomes the background for what happens the next day.  Like building blocks in any complex construction, the foundation comes first and much of the rest depends upon it.  When the foundation is faulty or missing the time spent catching up or in remedial work is time not spent on task.  Remedial work is time going back to fill in the blanks that other students acquired earlier.

When new learning is based upon previous learning that is faulty, the entire structure becomes fragile.  Like Humpty Dumpty, fragile students eventually fall off the wall.   Humpty Dumpty is history in a cumulative schooling experience.

Unlike Humpty Dumpty, most students do not drop out of school suddenly, but are required to suffer remedial work over and over until they get it right, or don’t get it at all.  After a few or repeated failures with school coursework, every normal response is to find a different road.  Compulsory education laws prevent realistic alternatives that are not endorsed by the schools, at least until the student has put in twelve years.

Such inflexibility in programming produces students who thoroughly detest school in all its forms, and see no hope in any other direction in their lives.  This unpleasantness in ones personal experience often radiates to the students around them, and the climate for others to learn is likewise degraded.  When such students drop out of school physically their disruptive influence allows those remaining to return to the task at hand.  Looking carefully at the miserable attendance records of students who care little about school should be a wake-up call for all administrators to develop programs in which academic dropouts can achieve new life through totally new, non-academic learning.

The world can be an exciting place, and a great deal more of the real world should be a part of each and every school for those students who show little interest in academic subjects.  It would appear that the pervasive anti-school culture within the black community has not been improved through school integration into a single academic mold.

Brink Lindsay writing in the Wall Street Journal (The Culture Gap, July 9, 2007) captures the cultural influence within the failing school community precisely when he states:
“The problem is not lack of opportunity….. The problem is a lack of elementary self-discipline: failing to stay in school, failing to live within the law, failing to get and stay married to the mother or father of your children.  The prevalence of all these pathologies reflects a dysfunctional culture that fails to invest in human capital…..  At the upper end of the socioeconomic scale…. parents intensively oversee kids schoolwork, and stuff their after school hours and weekends with organized enrichment activities.”

The current study provides the supporting data that was learned over 20 years ago, and due to racial interference and cultural enabling, the courts continue to divert attention within the schools from those choices that can make a difference to the racial divide in student performance that has very little to do with failing schools.  What is described as the failure of America’s schools is most clearly the direct byproduct of a pervasive cultural dysfunction within the lower socioeconomic classes as exemplified by the school performance of most black students.

 Gaming the SAT System

A final observation is essential regarding a pattern of test behavior among the student respondents in 1984.  Almost one out of six, over 5,500 students who took the SAT appear to have been gaming the system.  At the time, Georgia colleges required taking the SAT as a condition of admission, while at the same time the admission of black students with very low SAT scores was almost assured.  Most all of these students did not bother to answer the 64 questions providing biographical information.  The SAT scores of these 5,500 students was at the same level as the Native American group.  A startling observation was also the hundreds of students who were assigned minimum scores for their effort. Such scores can be achieved simply by signing one’s name, or some other person’s name without answering a single question right.

Under pressure to admit black students to college campuses to document compliance with racial guidelines, actually answering the questions on the SAT was widely known to be wasted effort.  Pay the fee, sign your name on the form, and leave testing.  It really didn’t matter to a large number of students who were smart enough to know it wouldn’t influence their college admission.  This system of admission is called racial preferences.

After fifty years of togetherness, the Supreme Court finally decided in 2007 that race, by itself, was an insufficient and illegal basis for assigning students to programs within the schools.  It took them over 50 years to get it right, but that was by a five to four decision.
*  *  *  *  *  *  *  *  *

Bob Settles was on the faculty in the Counselor Education Department of the College of Education, University of Georgia from 1966 till 1993.  For three years he ran the Research Processing Unit of the College.

Leave a Reply