SPATIAL MARKERS STUDY: A REPLICATION AND EXTENSION Thesis for the Degree of M. A. MICHIGAN STATE UNIVERSITY GEORGEANNA M. TRYBAN 1975 SPATIAL MARKERS STUDY: A REPLICATION AND EXTENSION BY \e) Georgeanna M9 Tryban A THESIS Submitted to Michigan State University in partial fulfillment of the requirements for the degree of MASTER OF ARTS Department of Sociology 1975 TABLE OF CONTENTS INTRODUCTION .. . . . . . . . . '. Statement of the Problem . . . ME THO m LOGY . . . O O O O C C . Subjects . . Setting . . Procedure .. . Coding . . . Density . . . RESULTS 0 O O O O O O O O O 0 Comparison with Becker's Findings Additional Findings . . . . . Density . . . . . . . Direction Orientation Behavior Analysis of Marking Behavior The Influence of Occupant's Sex Analysis of Time Effects on Occupancy Effects of Sex-Composition of Occupancy-Pairs . . . . DISCUSSION . . . . . . . . . . BI BLI OGMPHY O O O O O O O O O 0 ii "U 9! L0 (D \DxlU'le N H I-" |-‘ O NNHHHHH wommmmo bww umo Table l. 11. 12. 13. LIST OF TABLES Position of Seat Taken by Second Occupant According to Sex of Second Occupant . . . Average Time in Minutes Before Empty, Marked and Singly Occupied Tables are Occupied . Average Time in Minutes Empty, Marked and Occupied Tables are Occupied by a Second Person . . . . . . . . . . Seating Orientation of Males and Females . . Proportion of Singles and Withs Using Markers Average Number of Seats Marked by Singles and Withs that Marked . . . . . . . Sex of the Second Occupant by Sex of the First Occupant . . . . . . . . . . Position of Second Occupant by Sex of First and Second Occupants . . . . . . . . Summary of Time and Sex Data . . . . . . Average Time in Minutes (t1) Between Arrival of First Occupant and Arrival of Second Occupant, Noted by Sex Composition of Resulting Pair . . . . . . . q . . First Occupant to Leave Table According to Same-Sex Composition of Table . . . . . First Occupant to Leave Table According to Cross-Sex Composition of Table . . . . Average Time in Minutes (t2) Between Arrival of Second Occupant and Departure of One of the Occupants by Sex Composition of Occupying Pair . . . . . . . . . . iii Page 12 13 15 19 21 21 22 24 26 28 29 31 34 LIST OF FIGURES Figure Page 1. Average Density in Books Checked Out Per Hour Over the Observation Period . . . . . 17 iv INTRODUCTION Statement of the Problem The study reported here is a replication and exten- sion of an earlier study by Becker (1973). His study of spatial markers focused on the invasion of one person's "territory" by another. Becker tested whether persons maximize the distance between themself and others in choos- ing a place at an occupied table. He also looked at the selection preference in taking a seat and the duration of occupancy in a seat when these conditions applied: empty table, singly marked table or singly occupied table. He further noted whether original occupants defended their table when "invaded." Becker spent three months studying different areas of a library, noting whether people sat at tables that were empty or that were marked with notebooks and/or paperback books. Becker selected areas for maximum density and arrived before anyone else did each morning. He found that persons sitting at already occupied tables position them— selves so as to put maximum distance between themselves and the already seated person. Becker also reports that the preference for occupancy, measured by how soon and how long a table is occupied, showed that people chose to sit at and stay at first empty, then marked and finally singly occupied tables. He also reports no occasions of the first occupant asking the second occupant of a table to leave. My study involved not only a replication of Becker's research, but also included an examination of the relation— ship between occupants in regard to positioning at tables along divisions by sex, which Becker recorded but did not report. In addition I made an analysis of duration of occupancy at two person tables, on a time-continuum. Like Becker, my study was conducted in a library, although my sample size was approximately four times as large as his. I gathered data for 10 days out of a two-week period (Becker took three months) for a total of 100 hours of observation (Becker does not report hours of observation). METHODOLOGY Subjects The subjects were students at Michigan State Univer- sity who frequented the library during winter term 1975. There were 410 students, 223 males and 187 females, included in the analysis. Other students moved in and out of the region of observation but were not subject to analysis unless they sat at a library table. Students were assumed to range from freshmen to graduate students. No direct contact was made with any of the subjects nor were there any indications that these persons were aware that they were being observed. Setting The experiment took place at the Michigan State University library on the second floor, in the reserved. reading section. Students desiring a particular book that has been reserved by the professor teaching a course can come to a counter and request it. These books are loaned to the student, who must show identification of student status, for a two-hour renewable period. Since these books cannot be taken from the library, students generally take the book to one of the approximately 25 tables on the south side of the reserved counter and read there for the two hours. It was decided to study 16 of these tables in an attempt to replicate Becker's study (see map on following page). These tables stood between book shelves (not con— taining reserved readings) and pillars between which were cubicals for individual readers. This caused these parti- cular tables to be isolated from the others and grouped together. The length of the lOan period causes these tables to be characterized by high turnover. Thus a large number of seating choices could be efficiently observed in this area. L '8' BEII t--—-Back Wall of Library Windows along all of wall Other tables were arranged in this area F— Cubicals ———' seAIaqs xooa IIEI (a) 5 .5 U) [‘1 .5 I Usual Point of Entry to Area Reserved Reading Desk Line to Take Out Reserved Books 4— MAP OF LIBRARY A Representation of Observed Area--Not Drawn to Scale J seAqus xooa PSI? eorjgo pue Item areas Procedure The tables observed were wooden, four-person tables. They were arranged in two rows of eight. I observed the first half of each of the two rows and an assistant took the second half. The assistant was a black male who knew only the general nature of the study. We arrived at the library each day, from Monday through Thursday, at 1:00 p.m. A random order of seating predetermined which seat among the 32 being observed would be occupied by the observer and which would be marked. If the indicated seat was already occupied the next seat on the random seating list was taken. We entered the area and occupied (for one to five minutes) the seat that was to be marked. We opened a notebook and began reading a book, then turned the book over, open with the title visible, on the already open note- book and walked into the rows of library books standing next to the tables. We walked around the library until we were able to enter the area again, as if for the first time. At this point each of us sat at the table selected for us to observe from and stayed seated until 5:00 p.m. The experi- menters then picked up their own books, and the books that had served as markers and left the area. .This procedure was designed to create three distinbt seating configurations: (1) empty table; (2) marked table; and (3) singly occupied table. During the observations it was expected that students would enter the area and react to these conditions through their choice of a place to sit. Since this area was never entirely devoid of occupants it was not always possible to begin an observation session with all three of these conditions present. At times there were no empty tables and the marker and experimenter were then located at singly occupied tables. In each case the experimenter noted the time at which tables equipped for the three conditions became empty. There were also cases in which students' choices of seats created conditions similar to experimentally induced ones. If, for example, a student approached an empty table, put some books and a notebook on it and left the area for awhile, condition two was produced. When these things occurred they were treated in the same way an experimentally-produced condition was. The record of students' movements was kept in a time-progression format. When any person sat down at or left a table within the area of observation it was recorded, according to the time of occurrence, sex of person, type of behavior exhibited and location of behavior. Two types of behavior were focused on. The first was simply whether the subject was beginning or terminating occupancy of a given area. The second type of behavior noted was marking. Anything that an occupant placed on the table or chair was considered a marker. Marker positionings were of two kinds, those placed on or directly in front of a chair by a subject who then left for a period of time; and the other kind were those things placed in any part of the desk area, or accompanying chairs, that were not within the boundaries of the quarter of the table directly in front of the sub- ject's chair mentioned by Becker (1973, p. 443) as seeming to belong to the subject. Subjects generally used coats or books as markers. Since it was January when the study was conducted, winter coats were worn by all subjects who came to the library and were often draped over the table or an adjacent chair. The subject's position was recorded by assigning each table a letter, A through P, and numbering Ithe position of the chairs 1 through 4. Arrows were used to indicate beginnings and endings of occupancy at a table. For example: if at 2:45 a female sits at the first table on the right in the seat to the right, facing the south and keeping her books within the quarter of the desk directly in front of her, it would be recorded thus: 2:45 F I I2--where F designates female, I designates being seated, I indicates the table at which she was seating herself, and 2 the position or chair she took at the table. Coding The observations recorded were coded and tallied in the following ways: a count was made of the number of subjects of each sex sitting at empty tables facing either north or south. The number of minutes a table was left empty, and the number of times this occurred was recorded, with no special conditions necessary to defining an empty table other than that it be occupied. It was considered no longer empty when a person chose it, either by marking it (at which time it became classified as an unoccupied marked table) or by sitting at it. In this same way the coding was done for a marked table which was left empty. In this coding it was therefore possible to count a table as empty until it was marked, then as marked-empty until it was occupied, and then singly, doubly or triply occupied, or occupied and marked after that. The amount of time and number of times a person occupied a previously empty table without another person also occupying the table was coded. The same scheme was used in coding the amount of time a person spent at a marked table that was not occupied by any other person during the original occupant's tenure. I The number of singles (persons who arrived and sat alone) was coded along with the number of times each marked a table and the number of markers used. In this condition the table had to be empty before the arrival of the singles. Markers were counted when a person's possessions exceeded the one-fourth of the desk space (including the chair) directly in front of the subject. There were three other quarters available for marking to a single person sitting at an otherwise empty table, and "markers used" indicated I how many, if any, of those three were marked. Some coding was done for "with's,"1 and the same definitions were applied. I also coded the occasion when a person sat at an unoccupied table and was joined by one and only one other occupant. The sex and time the original occupant sat down and the sex of the second occupant, and the time and place of occupation at the table in relation to the original occupant was noted. There were three possibilities in location, the second occupant could sit directly across, diagonally across or adjacent to the original occupant. Also which occupant left the table first and the time at which this happened was noted on the code sheet. This condition was recorded only when the two people at the table were judged not to be a with. If a third person arrived at the table the case no longer could be counted since it was the second occupant's choice of a seat and the two persons subsequent behavior that I wished to study. Density A count was kept of the number of books being checked out at the reserve reading desks. This was kept on an hourly basis during the ten days we conducted the experiment by the student librarians who were checking 1"With's" were considered to be groups of two or more persons who gave some indication of being together. Erving Goffman, Relations in Public (New York: Harper & Row, 1971), p. 19. 10' out the books. A counter is used to register books being checked out and the number was recorded for the experi- menter's later use. This was used as an index of density since it was felt that the twofhour loan period constrained most people to read in that general area, thus the number of books checked out should give a rough indication of the number of persons in the observational area. The exact time of the hourly readings occasionally varied according to how busy the librarian was. At times the unavailable hourly numbers were statistically approximated. RESULTS Comparison with Becker's Findings In examining his first hypothesis, Becker tests whether "persons who sit at a marked or occupied table sit in a position that maximizes the distance between them- selves and either the marker or other persons" (1973, p. 441). His hypothesis was reported as being confirmed (p < .02, x2 = 4.6, df = l) with 83 percent sitting diago- nally at occupied tables and 51 percent doing so at marked tables. In this replication it is clear that this finding was substantiated. There were, out of a total of 131 subjects, 75.6 percent sitting at diagonally marked tables. Fourteen and one-half percent sat across from each other and 9.9 percent sat adjacently, as compared to Becker's report of 83 percent who sat diagonally (he does not report ll percentages separately for the adjacent or across cate- gories). This can be seen by looking at the bottom marginals of Table 1. Looking at a further breakdown by sex this same table demonstrates that regardless of the sex of the first occupant slightly more men sit diagonally while a few more women chose adjacent seating. In a test of significance these relationships didn't prove to be significant however (.20 < p < .30, x2 = 2.5493, df = 2). Becker's second hypothesis concerns table prefer- ence. His hypothesis which was reported confirmed (p < .05, t = 1.81, df = 39) stated that marked tables would be occupied before already occupied tables, implying that empty tables would be taken first. Another way to conceptualize his hypothesis is that table preference goes from empty + marked + occupied. In looking at my data a somewhat different pattern emerges. The average period of time before a table was occupied was used to calculate its desirability.2 I found empty tables were occupied after marked ones, contrary to Becker's finding that empty tables were the first to be occupied. According to Table 2, an empty table was occupied after an average of 16.3 minutes of emptiness, while marked tables were occupied after an 2Actually Becker doesn't consider the amount of time before an empty table is occupied. He only reports the amount of time before (singly) occupied and marked tables are occu- pied, an average of 127.1 and 105.5 minutes, respectively. 12 Table 1. Position of Seat Taken by Second Occupant According to Sex of Second Occupant.* Diagonal Across Adjacent .792 .143 .065 Male (61) (ll) ( 5) .585 Sex of Second .704 .148 .148 Occupant Female (38) ( 8) ( 8) ‘415 .756 .145 .099 (99) (17) (13) 1'00 *Data given in proportions of persons making particular seating choice. Raw data given in parentheses, propor- tions taken of right hand marginal. 13 Table 2. Average Time in Minutes Before Empty, Marked and Singly Occupied Tables are Occupied. Empty . 16.3 Marked 13.7 Type of Table Occupied (by 1 person) 19.1 14 average of only 13.7 minutes. The already occupied table however was the last choice for occupancy since it was occupied by one person for an average of 19.1 minutes before being chosen by a second person. Sequential t-tests were done to determine significance and showed that marked tables were occupied significantly earlier than were empty tables. Also both empty and marked tables were occupied significantly sooner than singly occupied tables.3 Becker's third hypothesis was similar to the second. It dealt with how long a table was occupied. Becker found that persons would sit for the longest amount of time at a previously empty table, their next longest duration would be at an otherwise singly marked table and they would spend the shortest amount of time at a formerly singly occupied table. (Again empty + marked + occupied.) He reports this was confirmed (p < .01) with the averages being 77.1, 53.2 and 32.4 minutes at otherwise empty, marked and singly occupied tables respectively. I found a somewhat similar though hardly as definitive a trend (refer to Table 3). The empty table was occupied an average of 46.40 minutes, a marked table for 46.46 minutes and an occupied table for 39.70 minutes. These averages of 81 cases showed no 3For empty vs. marked tables p < .0005, (t = 4.377, df = 177); for marked vs. occupied tables p < .025, (t = 2.173, df = 263); and for empty vs. occupied tables p < .005 (t = 2.908, df = 252). 15 Table 3. Average Time in Minutes Empty, Marked and Occupied Tables are Occupied by a Second Person. Average Time a Table Is Occupied 1 Empty 46.40 Condition 2 of Table Marked 46.46 Occupied3 39.70 1No one else is at the table when the subject sits down, during occupation or at time of departure. 2No one but subject sits at table (as in l) but a marker is at the table, before,during and after occupation. 3Only 1 other person occupies this table and is sitting there before, during and after second occupant (whose duration of occupancy is reported above). 16 significant difference between duration times at empty, marked and occupied tables.4 Additional Findings Density The average number of books taken out each day was 321.8, the standard deviation was 44.07. This small a deviation shows that although there were some times of greater activity, in general the density remained relatively constant. The changes in density are too small to conclude that the setting itself changes enough to have vastly differing dynamics operating at different times of the week. The average number of books taken out during each day varied sharply with the time of day (see Table 4). The early afternoon hours show much higher density, which drOps at both 3:00 o'clock and 4:00 o'clock. The later density being half that of the first two hours. Direction Orientation Behavior In addition to the replication of Becker's work that has already been detailed, I explored a number of other hypotheses and questions. The first was, do people arrange themselves in any consistent way facing either the 4For empty vs. marked tables p < .1, (t = .004, df = 51); for marked vs. occupied tables p < .l, (t = .472, df = 39); and for empty vs. occupied tables p < .1, (t = .855, df = 66). 110‘“ 100.. 90+P 80 «lb 60‘" 50 «- 40-% Number of Books in Circulation 304' 20;} lo'j' 17 104.56 103.Qq 87.11 52.22 I I‘ lp.m. 2p.m. 3p.m. 4p.m. 5p.m.~ Hour of the Day Figure 1. Average Density in Books Checked Out Per Hour Over the Observation Period. 18 windows or the area of entry? (See the map following page 3). This subject interested me because of the findings of Sommer. In a paper and pencil study he found that students preferred to sit with their backs to the door of a room. When the opportunity to examine male and female choice behavior presented itself, I decided to explore the area to satisfy my own curiosity. Table 4 shows the direction both males and females sat, either facing the area from which they entered, or facing toward the windows. In the bottom marginals of Table 4, it can be seen that 43.2 percent sat facing away from the area of entry to the desk grouping, and 56.8 sat facing it. The body of the table shows that males and females were highly similar in their preference for a location facing the entrance. Analysis of Marking Behavior I then sought to answer the questions, how many singles mark tables compared to with's and how many places at the table do the two types mark? As already mentioned, the word "with" denotes any group of more than one person and in this study four or less, since library tables only seat four. Without exception with's emerged as being groups of two persons. Calculations were made of with's and singles who sat at empty tables and it was noted how often and how many areas (on a chair or the table directly in front of it) were marked. In this analysis the with and the single were considered individual social units. Table 4. Seating Orientation of Males and Females. l9 Facing Entrance Back to Entrance Male Sex of Fe le Occupant ma .570 .430 .534 (45) (34) (79) .565 .435 .466 (39) (30) (69) .568 .432 1.00 (84) (64) (148) (.95 < p < .98, x 2 = .0028, df = 1) 20 Table 5 shows the singles and with's that used markers. A greater proportion of withs than singles (.375 as opposed to .306) used markers. Table 6, which shows the number of seats marked by singles and withs, indicates that both social units mark very nearly the same number of seats. .Note, however, that singles who marked one space are creating a ratio of one person to 1.068 seats marked in terms of the additional space they commandeer. With's, being made up of two persons, have a ratio of l to .5 in terms of the amount of extra space gagh person has acquired through the marking process. To summarize, although with's considered as a single social unit mark more often than do singles, on a per-person basis with's mark fewer seats than singles. The Influence of Occupant's Sex I believed that the sex of a person might influence the direction in which the person faced when seated. The results already mentioned in Table 4 show that it's unlikely that sex influences which direction is faced. The general preference is clearly for seats that face in the direction from which persons enter the area. Once the first person is sitting at a table I wished to see if their sex influenced the sex of the second person choosing to sit at the table. Table 7 shows the relationship between the first occupant and the person 21 Table 5. Proportion of Singles and Withs Using Markers Proportion of Social Unit Using Markers Type of Singles .306 Soc1a1 Unit Withs .375 (p < .001, z = 3.0) Table 6. Average Number of Seats Marked by Singles and Withs that Marked. Average Number of Seats Marked Type of Singles 1.068 S°°1al unlt Withs 1.000 22 Table 7. Sex of the Second Occupant by Sex of the First Occupant. Sex of Second Occupant Male Female Male .519 .500 .511 Sex of the (40) (27) (67) First Occupant Female .481 .500 .489 (37) (27) (64) .588 .412 1.00 (77) (S4) (131) (.80 < p < .90, x2 = .04816, df = 1) 23 choosing to become the second occupant according to the subjects' sex. When a male or female was sitting alone at a table they were as likely to have a member of the same sex as they were to have a member of the opposite sex become the second occupant. Once the choice of a table has been made a further choice takes place: where at the table will the person sit in relation to the one person who already occupies a seat. It has been shown in Table l the overwhelming preference is for a seat diagonally from the person already seated. Table 8 shows a greater prOportion of males, regardless of whether they sat with other males or with females, sat diagonally. Female second occupants, in con- trast, exhibited less of a tendency toward diagonal seating, both with males and females. The remaining cells in Table 8 will be considered in the following section which deals with the vacating of mutually occupied tables. Analysis of Time Effects on Occupancy In looking at a person's behavior in the library, the following behaviors were focused on: finding a table and a position at that table, sitting there and later leaving. 24 in u we .meom.m u «x .om. v d v om.e 2.. V 3 V 83 S e 3 V 32 who. mmH. Hen. mos. moa. ems. masses usemnooo ocoomm mo xmm 3 v 8 v 33 3 v In I 3.2 «mm. HHH. sow. mno. mud. com. mam: uneconod. mmouod HmcomMAQ ucmomflod mmouofi accomMHo cowuflmom madame cowuflwom was: .mucmmsooo oncomm pom umuwm mo xmm an ucmmsooo Ugoomm mo GOAHHmom .w OHQMB 25 In pictoral form this is the flow of behavior that was examined: occupant selects another occupant one of the an empty table selects a seat two occu- at that same pants leaves table + + + t1 t2 The symbol t was used to denote the time, in 1 minutes, between the arrival of the first occupant and the arrival of the second occupant. t2 represents the time between the arrival of the second occupant and the depar- ture of either the first or second occupant. It was expected that some factor or factors other than sex or position might determine how long people sat together at a table and which occupant left first. First I will analyze the effect of the initial period that an occupant sits at a table before being joined by a second (in other words the t1) upon who leaves the table first. It was found that the first and second occu- pants leave the table first at an approximately equal rate (27 as opposed to 28 times as shown in Table 9). Thus, being the initial occupant or the second person at the table seems not to influence who leaves the mutually occupied 26 ~v.ov wo.m¢ mm.va mn.m~ mm mm mm.wm mv.mm mm.ha Hm.mm h h mamamm was: an.>m mm.mm mv.va mm.mm n a mam: mHmEmm 3.2. 8.3 ~22 3.3 m m menses 38808 vv.mm ha.mv mm.wH mm.mm m m mam: was: umuflm umuwh umuflm umnflm umuwm umuwm ucmmsooo ucmmsooo mm>mmq mo>mmq mw>moq mobmmq mm>mmq mw>mmq wcoomm umuflh ucmmsooo ucmmoooo ucmmsooo ucmmsooo pcwmsooo ucmmsooo mo xmm mo xmm pcoumm “mush ccoowm umuwm vcooom uwnwm H nowaz CH :03; u :053 u mmocmunooo mo Hmnfidz .mumo xmm was made no Sunfifidm .m manna 27 table first. In looking at the time period that the initial occupant spends as the sole occupant (t1) an interesting finding emerges. Viewing the process in retrospect, when the first occupant is also the first person to leave the table, the time spent as sole occupant is an average of 28.75 minutes, while that time only averaged 14.98 minutes when it was the second occupant who left first (see Table 11). The amount of time the two occupants sat at a table together (t2) is virtually equivalent, regardless of which occupant leaves first. A first occupant is as likely to leave as is a second. Both occupants are equally likely to leave first and they do so after spending about the same average number of minutes at the table together, 42.04 and 40.42 minutes. The crucial factor in who leaves first is not who was the first or second to arrive, or how long the two occupants have been sitting at a table together. What seems to influence one of the occupying pair to vacate the table is the amount of time the first occupant occupies the table, which is shown in the t1 scores. Viewed retro- spectively we see that the first occupant who leaves has been sole occupant of the table almost twice as long as when it is the second occupant who leaves first (see marginals in Table 11). 28 Table 10. Average Time in Minutes (t ) Between Arrival of First Occupant and Arrival of Second Occupant, Noted by Sex Composition of Resulting Pair. Average time between arrival of first and second occupants When First When Second Occupant Occupant Leaves First Leaves First Male-Male 28.83 16.88 Sex Composition Female-Female 20.40 11.22 of Pair Female-Male 29.56 14.43 Male-Female 36.21 17.38 28.75 14.98 (.95 < p < .98, x2 = .2855, df = 3) 29 Table 11. First Occupant to Leave Table According to Same- Sex Composition of Table. Occupant That Leaves First fiv— First Second Occupant Occupant Leaves Leaves First First Male-Male 24g? 253? (fig? Composition of Table .500 .500 .400 Female Female ( 5) ( 5) (10) ~440 .560 1.00 (11) (14) (25) (.50 < p < .70, x2 = .2435, df = 1) 30 Effects of Sex-Composition of Occupancy-Pairs IV v' If the sex composition of occupancy-pairs is examined a slight male-female difference of behavior seems to emerge. Table 9 shows that there is a slightly greater proportion of first occupants leaving first in same-sex tables (.44 compared to .56). This difference is not sig- nificant, actually female first and second occupants are equally likely to leave, it is in the case of the same sex male tables that a greater proportion of the first occupants leave second (.40 compared to .60). Table 10 presents the prOportions of occupants who are first to leave from tables having cross-sex occupancy pairs. When the male is joined at the table by a female they are each equally likely to be the first to leave. When the female is there first and a male sits with her, there is a slight tendency for her to be the first to leave. When the same-sex and cross-sex pairs behavior are compared (Tables 11 and 12) females who are joined by females do not vacate in any greater proportion than do the females who joined them. However, this proportion changes when the female is joined by a male. The female first occupant is then more likely to leave. When she joins a male however this tendency is no longer seen and the probability of who will leave the occupancy-pair first resumes a random status. Summarizing male behavior; when the male is occupying the table and is joined by a male it is more likely than would 31 Table 12. First Occupant to Leave Table According to Cross-Sex Composition of Table. Occupant that Leaves.First ‘— First Second Occupant Occupant Leaves Leaves First First Male-Female 253? 153? 2:2; Composition of Table .438 .563 .533 Female Male ( 7) ( 9) (16) .467 .533 1.00 (14) (16) (30) (.70 < p < .80, x2 = .1173, df = 1) 32 be expected by chance that the male who joins him will be the first to leave. When joined by a female however the proportion of those who are first to leave is random. To summarize female behavior; when a female is joined by a female, or takes the initiative of sitting with a male the prOportions of which will leave first are equal. When she is joined by a male she's more likely than chance to be the first to leave. When a male is joined by another male the likelihood is that the second male will leave before the first, but he's as likely to leave as the second occupant when that second occupant is a woman. When the male takes the initiative of joining the female, unlike when she joins the male and chance pervails, the original female occupant is more likely to leave first. When time is taken into consideration one of the first possibilities is that one of the two sexes may be sitting alone longer before being joined (the t1 may be larger for a particular sex-composition of an occupancy pair) thus challenging the sex difference in regard to who leaves first reported above. The findings show a slight difference in the average times before a single table became doubly occupied. The average minutes were 23.2 before a second male joined the first and 18.5 before a second female joined the first. Cross-sex times were very similar. Females were joined by males after an average of 18.75 minutes and males were joined by females after 20.5 minutes. Thus no outstanding differences in tl's for different sex- 33 pair combinations exists. It should also be noted that males joined males 15 times, females joined females 10 times, males joined females 16 times and females joined males 14 times. This second indicator of preference again shows that there seems to be no sexual discrimination or strong preferences in forming the different combinations of occupancy-pairs. Table 11 shows that when the tl‘s are either excep- tionally long or short there is a difference in which.person leaves the jointly occupied table first. The marginals show that if the first occupant sits at the table for a considerable time before being joined by the second occu- pant he or she will be much.more likely to be the first to leave. When the first occupant has only been there a very short period of time before being joined it is less likely that he or she will be the first to leave regardless of the sex composition of the pair. Thus when the average t1 is either quite long or quite short it modifies the prev viously mentioned results with regard to sex differences summarized on page 26. Considering the amount of time a pair remains at a table, Table 13 shows that time spent sitting together (t2) is significant in determining who leaves the table first. An analysis of sex differences shows that different sex composition of occupancy pairs results in differing amounts of time spent sitting together at a table. Which member of a pair is more likely to leave first has already been 34 Table 13. Average Time in Minutes (t2) Between Arrival of Second Occupant and Departure of One of the Occupants by Sex Composition of Occupy’ ing Pair. Average Time Pair is Together at Table When First When Second Occupant Occupant Leaves First Leaves First Male-Male Female-Female Sex Composition Female-Male of Pair Male-Female 48.17 33.44 66.00 43.60 28.56 57.71 25.43 26.93 42.04 40.42