With roots in the work of Jacob Moreno (1934), sociometric testing is a technique used to measure social relationships in a group. Although there are many variations, a typical sociometric test might ask students to identify the classmates with whom they would most like to associate in some particular activity. Results are used to diagnose problems in social interactions of individual students and the classroom in general and to help teachers design strategies to solve those problems. `

Why Use Sociometrics?
Why should the teacher take into consideration student's choices of who they would like to sit by or work with? Why, in fact, should a teacher not separate children who might disrupt the classroom by socializing with friends? According to Helen Jennings (1959):
  • Sociometric grouping offers teachers an opportunity to create a predisposition for active give and take. Teachers usually find such classrooms easy to work with and generally responsive. Work projects can be initiated at any point in the known network of pupils' affinities and readily extended to include other members of the class. Under such conditions children develop considerable appreciation of and interest in different viewpoints with surprising ease; their sympathies come into action with very little prompting. When their most immediate psychological needs are adequately met, they can - and usually do - start reaching out toward more varied and challenging contacts.

One way to create a comfortable, non-threatening environment in the classroom is to give students input into some of the classmates they will be working with. Alternatively, thwarting the warmth and security a child feels in being allowed contact with one s/he likes or admires, can generate behavior problems as students attempt to satisfy their social needs covertly or as they act out in frustration. Moreno "stressed that if the structure of the formal organization does not take informal relations into consideration and nurture them, then discord, strife, and conflict will appear at the formal level of organizational functioning" (Schmuck and Schmuck, 1979). In addition, grouping to thwart, rather than meet, the social needs of children is like trying to plane wood against the grain. Meeting children's social needs (planing with the grain) not only makes the job easier, in addition, "there is little risk that arrangements based on spontaneous choice will turn out to be unsound either educationally or psychologically, (even if) the children's wishes run counter to the teacher's judgement" (Jennings, 1959). Every child benefits from greater acceptance and an improved classroom climate. Isolates may be the most visible gainers but even the stars are better off in a class where there is not so much pressure on them to lead in every aspect of class activity.

Prerequisite Classroom Climate - Students may not willingly state their real classmate preferences if there is any reason for them to believe that their choices might be used against them. They must truly believe that their responses will be kept confidential. It is recommended that the teacher interested in using sociometrics consider some of the suggestions in the classroom climate section which describes some strategies to develop a classroom climate of trust and openness. According to Jennings (1959), "in groups where freedom of communication is a rule, when children feel secure in exchanging feelings and ideas, in giving and taking criticism, there is also maturity, clarity, and directedness in their sociometric choices and in their reason for their choices." Of course it is always possible to have children group themselves. In fact allowing students to organize their own groups recognizes the possibility that they may be the best judges of who they can work with and who can best satisfy their needs (Schmuck and Schmuck, 1979). However, given a free choice of arrangements, most classrooms will end up with some students who are unchosen, some who feel obliged to join a class star, etc. If there are any problems in the class's social patterns - for example, there are one or two small clusters of highly chosen students, with the rest of the class members being mere followers - the teacher may be advised to try another method like sociometric grouping (Schmuck et al., 1966).

Besides the meeting of their obligation to use sociometric results as promised, teachers also need to keep sociometric data strictly confidential. In order to keep the trust of students and enhance the chances of open, honest responses, neither the teacher nor the students may reveal their choices. Jennings (1959) recommends making a statement to students that, even if they are asked by their classmates, they must not tell who they chose because the friends that were not included might be hurt if they knew.

Types of Sociometric Tests - As teachers begin using sociometrics it is tempting to try to dig as much information as possible out of the test. There is no doubt that there is a great deal of data to be had: Who do students want to work with in a variety of settings? Who do they want to play with? Who do they want to invite home? Who don't they want to do these things with? How do they feel about every classmate on every one of these criteria? The best advice of the experts is to keep it simple. Only ask the question(s) that directly relate to the purpose for using sociometric testing. Asking students to fill out long questionnaires can taint the results. In addition, data from complex sociometric tests is difficult to evaluate. The general-purpose sociometric test can be as simple as asking students to name the three students with whom they would most like to work for the next six-week unit. This type of data is relatively easy to analyze while, at the same time, giving a good snapshot of the classroom social structure. Before discussing how sociometric data is analyzed, a list of some of the ways to vary (and complicate) the "basic" sociometric test:
  • Include Negative Responses - Negative responses are helpful in differentiating between rejected and neglected children, and identifying social polarizations from large numbers of rejections. A rating scale, by requiring a response for each child, may ease the discomfort teachers feel about having students single out those they dislike.
  • Bubble Art - Vacha and his colleagues (1979) recommend the use of the "bubble art" survey in which students are given a duplicated sheet of the names of classmates in bubble style and are instructed to color the letters of the three students they want most to work with one color, least another, their own a third, and the rest any other color. This procedure makes the test more interesting to students and tends to keep them too busy to try see the papers of others.
  • Weighted scores - Compute a status rating based on weighted scores on acceptance scale: first choice +1, acceptance +.5, indifference 0, unacceptance -.5, last choice - 1. Students with combined scores of zero and below may need help (Zeleny, 1960).
  • Multiple Criteria Tests - Multi-criteria sociometric instruments measure more breadth of choices and whether students discriminate between companion choices for various criteria like: eat lunch with, field trip with, do a school project with, play a game with, sit in class with, invite to a party, newest friend (Roberts, 1986). The Ohio Social Acceptance Scale has students group each of their classmates under one of six headings: My very, very best friends, my other friends, not friends but okay, don't know them, don't care for them, dislike them. Maddux and Maddux (1983) adapted the scale so that students rate each classmate by telling how much they agree with the statement: "I want to sit by (or work with) this person." Responses are given on a continuum of agree strongly, agree, neutral, disagree, disagree strongly.
  • Skills Matrix - Gresham (1982) suggests a complex process of having children rank their classmates using a 3, 2, 1, 0 continuum on each of thirteen skills: says nice things to others, says please and thank you, smiles at others, says hello to others, listens to others, helps others, shares with others, says excuse me, waits for turns, participates in school activities, fun to talk to, is liked by others, follows rules in games and class. (3 = a lot, 2 = sometimes, 1 = never, 0 = don't know). Analysis of a matrix of scores could reveal skill deficits of the group and of individuals. Whole-group, small-group or individual instruction can be planned to build those skills.

Whatever the complexity of the sociometric test, it is important to keep all steps in administering the test unstressful and to make them real to the students. In other words, "students must fully understand that the question is asked in order to be useful to them and that their personal reaction is important because they are to benefit from the consequences" (Jennings, 1959).

Using Sociometric Test Results - What Sociometric Test Results Can and Can't Tell You
Of course the analysis of test results will depend on the sociometric question or questions asked. Regardless of the structure of the test, however, there are three main goals for sociometric testing:
  • To measure classroom interaction patterns and their changes over time.
  • To help plan intervention to improve the overall climate of the classroom.
  • To help plan intervention for specific children seen as having status improvement needs. Define isolates (recessive, socially uninterested, or socially unskilled students).

While sociometric tests are very useful in providing information to achieve these goals, they are not a panacea. Because data from sociometric tests are self-reports from one moment in life of the class, they can reflect aberrations in addition to on-going patterns of classroom interactions. Best friends who just had a disagreement, an absent child, a new student, any number of short-term situations can taint the results of a single sociometric test. Therefore, drawing conclusions about a class from the initial sociometric test should be done with some skepticism. Standard practice is to give a second test, using the same criteria 6-8 weeks later, to get an idea of what patterns from the first test are relatively stable and which are more volatile. For example, a student who was new to the school when the earlier test was given would likely have a completely different pattern of choices (to and from) on the second. If the teacher has made a good-faith effort to honor student choices on previous sociometric tests, "students will be eager to get a continuing opportunity to exercise choice and to learn to act on their own behalf and to live by their decisions" (Jennings, 1959).

Even though the usefulness of the first sociometric test is somewhat limited, there are six things for which the teacher might look. And on subsequent tests teachers would look for improvement in the same things plus constancy and changes in individual student choice patterns.
  • Surprises - Do any student's scores surprise you? Perhaps you need to reassess either the child or classroom norms.
  • Stars and isolates - If a large percentage of the class's choices are concentrated on a few students and there are a number of students with no choices, it may indicate a highly stratified social structure in the classroom, with influence concentrated in a few individuals. It seems reasonable that a better structure might show "a narrow distribution of scores, few isolates, more reciprocals and less cleavage, and wider networks (Northway, 1952).
  • Patterns - Is the pattern of choices fragmented and unfocused or are their webs of choices that eventually include almost all students in the class.
  • Cleavage and cliques - If there are very few choices across racial, social class, or sex lines or if there are groups that only choose each other, effectively locking other students out, this may be an indication of hurdles in the path of a positive classroom climate.
  • Mutual choices - "The existence of mutual ties among members is a good indicator of classroom cohesion." (Vacha et al, 1979) A count of mutual choices and the change in the count over time can give a measure of how strong the bonds are among classmates. (Grouper calculates and tallies mutual choices.)
  • Cohesion - Vacha (1979) provides a formula for a "coefficient of cohesion" which uses mutual choices as a metric. This coefficient is calculated in Grouper and sets goals of an increasing coefficient over time or 6.00 or better for a single score.

There is no "normal" class. Each classroom will be different and only one who is familiar with the students can interpret the data. For example, the age of the students changes the results markedly. From kindergarten to about fifth grade, there is typically a steady decline in cross-sex choices, then the trend reverses itself in about eighth grade. In addition, mutual choices usually increase with grade level (Jennings, 1959). To get a better view of the social structure of the class a teacher can also conduct follow-up interviews which can provide more information about the "motives and values underlying the choices and rejections" (Jennings, 1959). For the teacher who has conducted multiple sociometric tests, Bonney and Hampleman (1962) list several comparative measures of improved classroom climate increased positive choosing in later tests, decreased negative choosing, increased mutual choices, increased mutualities among highly chosen students (top 25%), increased breadth of choices when the sociometric question asks for choices for more than one criteria, broadening of cliques, and increased cross-subgroup choosing. Such changes do not occur quickly, however. According to Jennings (1959), "whenever the same criterion is employed at different times, the results reveal that shifts in feeling between individuals are not rapid" but when the class has undergone skills training and practiced activities in cooperative groups, some individuals will typically experience fairly dramatic changes in status. In many cases the choice patterns become more diffuse (usually a good sign) as students who were originally overchosen based on surface criteria such as appearance or wealth lose some choices to students who were for one reason or another initially overlooked.
Indications of Students with Problems - For the most part, students with possible problems would be those who receive few choices from their classmates. Those with zero positive choices are of most concern. Two common ways of using the numbers to find children who might need attention are:
  • Establishing a Baseline - If students are asked to make three Dchoices, one could consider two to four choices the baseline. Help may be needed by children with fewer than two positive choices or more than four negative choices (Vacha and colleagues, 1979).
  • Calculating Ratios - Use a ratio of the number of nominations received from same sex classmates divided by the number of same sex classmates (Gresham, 1982). Because most choices are same sex, this formula allows comparison of students in classes that are heavier in one sex than the other.

Northway (1960a) puts students who get few choices - she calls them "outsiders" - into three categories: recessives (those showing total listlessness, no interests in or out of school), socially uninterested children, and socially ineffective children. Determining which students fit which category requires observational data as well as sociometric test results. According to Northway, it is important to consider the differences of the three categories because treatment varies with the type. Some other considerations Northway (1952) suggests for evaluating sociometric test results include:
  • Distinguish between isolates and other "outsiders". A student chosen by any one classmate is not an isolate and can be considered a student less in need than those with no choices.
  • Sociometric scores "which are statistically identical are rarely sociometrically equivalent." For example, three scores of nine could be nine third choices or three firsts, etc.
  • Sociometric scores do not measure mental health or well-being, rather they are "an index of the degree to which the individual attempts to conform to and abet the group's norms; in this sense it is a measure of his drive towards external social adjustment."
  • Make a conscious decision as to whether or not the child is really in need of help or "whether he is contented or discontented in his low sociometric status. Remember there is no law saying a person must be highly social. Has he genuine interests he is pursuing? Recognize these."
Before deciding whether an intervention plan is appropriate for an isolated student, the teacher should always get a more complete picture of the student by looking at student records, observation, home visits, and interviews with the student.

Sociometric Grouping - After evaluating the results of sociometric testing, the first order of business is to fulfill whatever promise was made to the class when the test was administered. Usually this means placing as many students as possible with at least one of their choices. To complicate matters, sociometric data should not be the only consideration in grouping. In addition to honoring student choices, groups should be heterogeneous with a mix of races, genders, abilities, age levels and backgrounds. Of course, every group is not going to be balanced but as long as most students are placed with someone with whom they share some bond as well as students with differences who are not friends but have the potential to be friends, the groups have a good chance of working out. When grouping children systematically it is important that the purpose for grouping is under constant consideration. If, for example, a class is racially split and the teacher's purpose is to nurture cross-racial friendships, the primary criteria for grouping would be race. In an evenly split class, each group would be close to half one race and half the other. If there is a minority of one race, it is sometimes good practice to have some 50-50 groups while others are 100% majority students in order to avoid isolating minority students. Sometimes sociometric test results can provide a seemingly obscure clue to how to begin eroding a classroom cleavage. In the above example, for instance, the teacher would look for any cross-racial choices, no matter how far down the list, and honor those choices (Jennings, 1959). Similarly, a class with a ruling clique can be grouped not so much to break up the clique (probably impossible anyway) as to expand it by honoring the choices of members who picked students outside the clique. Since sociometric data show social aspirations, not social facts, they can provide clues for bringing about more satisfying relationships to help solve social relationship problems targeted by the teacher (Bonney and Hampleman, 1962). The following list of ways to use student aspirations to group students was gleaned from the research. It gives an idea of the options available to the teacher:
  • "Identify the isolated students who are not chosen by any other classmates. Then build a group of skillful and supportive students around each isolated child" (Johnson et al., 1984).
  • "With sociometric test results the teacher can identify the actual or potential friends of low peer-status students. If the teacher is unable to locate any friend for a student, he might place him in a group with students receiving many positive choices, because such students are probably quite secure and able to support the rejected or neglected student. The neglected student also may be able to learn new behaviors by observing these socially effective #classmates" (Schmuck et al., 1966).
  • If the sociometric question asked students to name in general the classmates they would like to sit with or work with (rather than work with in a specific subject or on a specific project), that student's first choice is the one aimed most directly at meeting his needs. That may provide some insight into the student's aspirations (Schmuck and Schmuck, 1983).
  • "Children who reject each other should never be placed together" (Bonney and Hampleman, 1962).
  • Groups can be structured to contain a heterogeneous mix of high, medium and low status students (Schmuck et al., 1966).Ä
  • Plan extra-curricular activities to include groups built sociometrically around underchosen children. Design the extra-curricular activities to appeal to the interests and strengths of targeted underchosen students (Flynn, 1990).
  • Gronlund (1959) recommends these steps in forming groups:
  • Place isolates with stars,
  • Place no isolates in the same group,
  • Place students in ascending order of number of choices received,
  • Honor reciprocal choices if possible.
  • In the treatment of cleavages such as a classroom split along racial, socioeconomic or sex lines it may be helpful to ignore first choices, which tend to go to members of the chooser's subdivision, and look for cross-cleavage choices among lower level choices (Jennings, 1959). Data which records every student's rating of every other student is more helpful in this case than just the first three choices of a traditional sociometric instrument.
  • Vacha and his colleagues (1979) recommend children receiving no choices be placed with at least one, preferably two of their choices. Their system:
  • Decide on group size (groups of at least five allow most children to have one of their choices in the group).
  • Separate children who express dislike for each other.
  • Place those with no choices in separate groups.
  • Place one-choice students with their choices.
  • Repeat the process in ascending order of the number of choices.
  • Do not place "stars" with more than one mutual choice.
  • Avoid placing stars with more than two of their choices.
  • Diminish cleavages by race, sex, athletic ability, etc.
  • Try to "open up" rather than "break up" cliques.
  • There are at least two schools of thought on rules to follow in grouping: 1) painstakingly structured groups that assure "the best possible arrangement for each student from his point of view realizing that since the same consideration must be shown to all of his classmates, there will have to be some compromise" (Jennings, 1959), or 2) less carefully assigned groups with an emphasis on helping students develop the skills they need to work with a variety of children. Which to choose depends on the needs of the class.
Does grouping students with their choices cause disruption? While it is true that noise and talk do sometimes increase in the excitement of being placed with students of choice, this is a temporary condition and one that would be found any time groups or seating arrangements are altered. Getting down to business right away is a key in establishing the educational purpose of new groupings. "The social atmosphere (of the classroom) is very largely created and maintained by pupil interaction, and this can be constructively influenced by the tone the teacher sets and the grouping practices she uses (Jennings, 1959). Of course children placed in sociometric groups do not automatically get along with each other and work well together. They need training, guidance, methods of solving problems, and time. Sociometric results can be used to tailor skills training to the needs of the class by basing it on the needs of the class's isolates or, if the whole class is not in need of the training, individuals and small groups can be taught the skills. In addition, individual goal setting can be used to build social skills in unskilled students. Training will be discussed in detail in the next section of this paper. Finally, a reminder that a single sociometric test is less useful than multiple tests and each test requires regrouping. According to Maddux and Maddux (1983), "the beneficial effects of grouping rejected students with popular students has been found to be temporary. Therefore, assessment of peer status, followed by grouping and assignment of cooperative tasks, must be an ongoing process. Four to six assessments per year, with continual initiation of task oriented groups, should be sufficient."

Computer Assisted Sociometrics - Despite their value, sociometrics are not widely used in today's classrooms because their analysis is so time-consuming (Markus and Barasch, 1982). A glance at the dates of the citations above is evidence of this decline in use. Even the most simple form can take several hours to evaluate. And if some of the areas of analysis listed in the table below are considered, well, it just seems too much to most teachers. But each analysis type has the potential of providing invaluable information. It seems to me that computers offer a potential solution. Computers excel at rapid calculations and rapid reorganization of stored data. It occurred to me that a friendly computer program that allows easy entry of student choices and that translates the results of complex computations into a usable form would make sociometrics more accessible to busy teachers.

Systems of Sociometric Analysis

Computer Mapping
Markus and Barasch, 1982
Matrix Analysis
Gresham, 1982
Class Profile
Jennings, 1959; Northway, 1960
Cross-Group Ratios
Jennings, 1959; Roberts, 1986
Isolate Distribution
Katz, 1960
Reciprocation Analysis
Criswell, 1960
Jennings, 1959
Target Sociograms
Northway, 1952; Northway, 1960b
Frequency Distributions
Northway, 1952
Index of "Knittedness"
idea from Glasser, 1986
Clique Analysis
Edwards, 1960; Langeheine, 1978; Burt, 1976; Hansell, 1981;
Forsyth and Katz, 1960
Watson, 2008
Walsh's Sociometrics (for PCs only)
Network Analysis

Sociometrics are not only a way to group children but a handy tool in structuring the environment of the classroom to maximize the positive classroom climate.