Loading...
HomeMy WebLinkAboutParticipant Data Development 1967 . BRUCE T. WILKINS PARTI CIPJlNT DATA IIVELffi''ENT. TOO (f SOUTHOLD VepaMmellt 06 COn6€//.va.U.an Nw YOIlR sta,te CoUege 06 AgJt.i.cuUwte A Statutolly CoUege 06 the S.ta:te Un-i.v eJL6U:y , a:t COllneU Un-i.veJL6U:y, IthlLC.ll, New YOIlR JIJ-lE 1967 . TABLE OF CONTENTS METHODOLOGY 1 Development of Participant Days - - - - - Discussion of Techniques Used -- Discussion of Participant Interviewing - 1 7 12 LITERATURE CITED - - - - - - - - - - - - - - - - - - - - - - 15 APPENDIX - - - - - - - - - - - - - - - - - - - - - - - - 16 . i . . METHODOLOGY Primary concerns in a study reported elsewhere (Wilkins, 1967) were the determination of problems arising in estimating total participant days in various outdoor recreation activities, and the determination of problems arising in developing data on socio-economic and attitudinal attributes of participants in those activities. Details of the development of the data presented in that study (Wilkins, ibid.), discussion of problems recognized, and suggestions on possible improvements in techniques are included in this portion of the appendices. Development ~ Participant ~ Instantaneous counts from land of persons or of conveyances engaged in specified activities were made from 34 observation points. Binoculars and mechanical counteres were used to assist in these counts. The obser- vation points were apportioned into three segments (A, B, and C, Figure 1), each requiring approximately two hours to drive. From observation points 55, 70 and 40 percent of the shoreline of areas A, B, and C respectively were visible. Selection of dates to cover a segment were stratified with one stratum for weekdays and another for weekend days and holidays collectively (henceforth referred to as "weekend days"). In 1964 observations were originally scheduled on all days worked. Whether observations were to occur on a particular day was determined by a coin toss in 1965. The time of observation of segments was initially based upon a pre- determined schedule providing coverage of all segments between 6 a.m. and 9 p.m. on both weekdays and weekend days during the "summer." "Summer" is interpreted here as June 27 - September 7, 1964 and June 26 - September 6, 1965. Extreme dates in both years were chosen as inclusive of all summer periods in which schools (local and in nearby metropolitan areas) were on vacation. . To obtain a measure of participant days using instantaneous counts, the portion of the total number of participants which an instantaneous count represents, must be determined. To determine this proportion observation points at three public beaches were manned from 8 a.m. _ 8 p.m. for a total of six days inclusive of weekdays, weekend days, fair and inclement weather. The total number of participants in all activities considered, the total number of boats observed, and instan- taneous counts of these persons or boats (at 15-minute intervals) were recorded on these days. 1 Long liliaI'd Sound A Pe con Ie Bay Figure 1. Town of Southold study area. . '" Gardlnerll Bay o 2 4 Scale in miles -3- Running of segments prior to 1 p.m. or after 5 p.m. resulted in observing only minor fractions of the total number of daily participants and this fraction varied widely. The ratio of counts made from 2 to 4 p.m. to counts at other times also varied widely. S\!~LlIDerS observed on fair days between 11 a.m. and 1 p.m. averaged 39 percent of the swi~ers observed on that segment from 2 to 4 p.m. on the same day. This varied 148 percent, however, ranging from 23 percent of the 2 to 4 p.m. total observed in the earlier time neriod in one instance, to 57 percent in another. W~ekend days seemeC"to provide a higher pro?ortion of the 2 to 4 p.m. counts (on the average) in the ea..,..,Ucr time pej,i.ads, but weekend days provi,led both of the ext:.: U1'es noted c:::.ove. Two i:tstanc"s of inclement weather provided much "'.d.er c',itferences between early "'ld lat.er counts-- the 11 a.m. to 1 p.m. counts representing less than 10 percent of the 2 to 4 p.m. count in both cases. Observations run between 1 and 3 p.m. yield~d an averace of 55 percent of the maximum nULlber of swimmers observable at any instantaneous count. Proportions of participants in other activities showed similar discrepancies between counts at early and later time per;.ods. The rather low perc',!nt of the maxiwum represented, together with what I consider adequate counts between 2 to 5 p.m., resulted in the 1 to 3 p.m. counts also being discarded. Only obser-lations run between 2 and 5 p.m. are used in this study to estimate the num-::>or of participants. Observations f'rom 2 to 5 p.m. included the largest propcrtion of participants in most activities. The proportion of' total d~ily use this represented varied less than observations at other time periods. Observations on inclement weather days in 1964 and 1965 indicated major differences in participation rates for certain activities, in comparison to "good" weather days. In 1965 random selection with non- replacement was used to select segments for observation on "inclement weather days" durL1g each calendar period. This latter schedule was used if on a day scheduled for ohservation: rain occurred; maximum temperature did not reach 70 degrees Fahrenheit; winds exceeded 19 mph, or if the weather was "windy," and the temperature maximum fell between 69 and 75 degrees Fahrenheit. Participation on fair and inclement weather days for both weekdays and weekend days were computed as noted in Chapter II. Inability to observe participants from observation points was another potential source of error and could result in erroneously low estimates of participation. Observation on five aerial flights indicate that the only mainland participants substantially unobservable by land observations as used in this study were swimmers at some private beaches. The starting point for observing a segment on a given observation day was chosen haphazardly. At each point, participants in activities under study and m""ber of' conveyances (cars without occupants and boats with occupants) were recorded. -4- Combining the summers of ""964 and 1965, segments A, B, and C were observed between 2 and 5 p.m. 18, 13, and 14 times respectively. Data presented combines the data for both 1964 and 1965. . Swimmers. Figure 2 (based on data from 8 a.m. to 8 p.m. observations described above) indicates the proportion swimmers recorded at 15-minute intervals were of the maximum number of swimmers observed at anyone of those interval counts. The percentages were computed using an aggregate percent method whereby the proportion present at each 15-minute interval for each of the days on which continuous coverage was maintained, were averaged to give the percent figures shown. Maximum and minimum percentage recorded on any day are also shown in Figure 9. Maximum numbers of swimmers were on beaches between 2:30 and 4:15 p.m. In five of the six cases a maximum number was recorded between 3:15 and 3:45 p.m. In one instance the same (maximum) number of persons was present at 3:15 and again at 4:15 p.m. At the point in time when (on the average) a maximum number of swimmers were on the beaches (3:45 in Figure 9) 42 percent (extremes of 33 percent and 53 percent) of the total number of swimmers using the beach that day were present. The percent of the maximum instantaneous count that would be observed between 2 and 4 p.m. was determined as 77 percent (average of percentages within this period). The maximum instantaneous count represented 42 percent of all persons swimming at the beach that day. Therefore, numbers of swimmers observed on segments between 2 and 4 p.m. were increased by a factor of 3.1* to approximate total usage. The averages of the percent of the maximum instantaneous count observed between 3 and 5 p.m. represented 74 percent of the maximum count and the 3 to 5 p.m. counts were therefore increased by a factor of 3.2. Averaging five aerial flights, swimmers on private beaches of segments A, B, and C were indicated to be underrepresented by 50, 53, and 53 percent respectively. The number of private beach swimmer participants (otherwise determined as described above) were increased accordingly. Boaters. Estimating the number of boat trips (to compute boater days) posed a unique problem. On two days when boat numbers were recorded from 8 a.m. to 8 p.m. peak numbers were noted at 3:30 and 3:45 p.m. (equaled at 12:00 noon on the latter date). A third day gave a mode at 12:15 p.m., while on a fourth day of recording boats a sailboat race yielded peak numbers at 8:30 a.m. An average of 11 percent of all observable boats would have been observed in any two-hour period between 2 and 5 p.m. on the four days when day-long counts were recorded. This figure might then be used as a factor for all boat observations. . * 3.1 = 1 .77 x .42 . 100, Percent of Maximum . 5 -~~'''''''''M.... ..~. ..-. " ... ...-. .. ..._,. . --. .':-.",. .. '," .-.','_.',-. .",' ....... ._-,., ".. ..- ....... ...'".. ""'".. ...... ..-.... "....... ..-.. .,-,-, -,.._-. "... . ...-..... ...-.. .-..... ...-.. , ....-.. .._,., . ..._,... ....-. . .....-. ... ............ ...........:,.-.......... .... .. ". . ,..--. -,...... ..'.-.....'. ," ","-".-.",' . ...._,. ....-... . "-". ........ . .-.... --... >):):..~/<~? .:: ':<<'::'::-::::". 90 Maximum 80 ....-...... ......._-,. .---,....-. j ....... . ...... . .... ,. .... ". 70 60 .'.-.. . ..... . ..... .. ..... . ...... .. ..... .. ....... .. "". .. ....". . ..... .-. "'.", .. .._,.. ". ..... "' ....... ... ..... .. ....... ... ..7.... ". ,",. . .... ... ... ". ..-. ". .. ... ... ... .. .... .... .. ... .,. ... -,. .,... ..- ,... .........~............... .,,-.- ... "...,. ,.. ....... .. ...... '''. ....... ... ""... .... - ...... .... ""... ... .....-..-.. ...... ..... '" ..... ..... '''.. "". ...... ..... .~..... .... ...... ..... ...... ..... ...... .... ....... .._" ....... .... ....... .." ....... .... .".... .... ....... .... "'''". ..... ........ ". ....... ).............. "" .. ..... ... '''..... ..... ........ ..... ...".. ~_................... ........ ........ '''''''''. ....... .......... "." "....... ..... .......... .". ......... '" .......... ... '''''"... ". ................. ....... . ...... .. ""'.... .. .......... ... ......... ". ."...... ... .......... ". ....."... .. .......... ... ........ ... .."...... ... .......... ... ......... ... .......... ... .......... ... .......... ... ........... ... .......... ... .......... .. .......... ... .......... ". ........... ... .......... ... .......... .. lfA............. ........... ''''''..... . ........... .. "....... . ........... .- "'''''.... .. ""...... ." ........... ... ........... ... ""-""'" .... ''''''''". ... "......." ............................. .. . ... .... '''''''''.... .....'....,...-.. ..-.-...,...-,.. ..,-......._,.... .,-...,....,..... ..,.,-...,...... ....-...,......,. ...".....".. .....,......,... ....._,..,.,-... ..........,..... J............ ............. ..........". ". .........".. ".. ........"... .... ......"....- ..... ............. '''. .........".. '".. ............ ..... ............. "". ............. ..... . ....,.,........ ....... ... ....... ..... ........................... . ............ "'.... . .."....". ........ .. ........... ........ .. .....".... ........ ... .......... '''..... .. ...... ......... I>..>> ...... ...............". ...... ................. ...... ...,........ ,"'" ..-........ .....................-:-:_...... ..... ................. :::::::::. .:::::::::::::::::::::::::::::::::.., }}' .::>;:::::>;:::::)~:::::::::: . ................ ::::::..,:::::::::::::::::::<:::::::::::::::: n...:.......,...:...................... .....1, 11 AM 12 M 50 :-:.:.....,... '.".::'.:':':" ..."... ,,'''''''''' ....... ............. ...... ............ ....." ...,,-..---. .'.,....,. ......,..... ...... ............ ......... ............ ........ ............ ........ ........... ........ .,,--,,_.... ........ ........... ......... ............ ........ ............ ......... .......... ......." ..---,---_. ......... .......... ..-...... ........ ...."". . ......... ........ .......... ........ ......... ..."" .......... ..... ....... ... .......... .... ......... .. ......".. . .:-. .......... ...... .......-_.. ........ ........... ........ ........... ...." .......... ........ ..."".... ... ........... ....... ........... ....... ......"... "..... ........... ....... ...... ...... ........... ...... ........... ...... .......".. ..." ............ ...... ,---..---- ..... ............ ..... ........... ...... ..."..-..- ..--.. ............ ...... ... .........." ............ ..... ............ .... ........... ..... ............ .... ............ .... ............ .... ........... .... ....- ----.- . ............ .... ............ .... ............ .." ............ ... ....----.. -, ........... .... ........... ... ............ ... ............ ... ..._---..... ... ............. ... ............ ... ............ ... ............ ... ~............... ............ ... .................... .. ..... ......... ............ .. ............. -, ............ .. ............ .. ............. . ............ .. ............ . ............. .. ............. . ".......... . ............. . ............. . ............. ............. . ......"..... ............. ............ .........-... .............. ............. .............. ............. .....-....... .............. ..,........... .-........... ........... 40 ;~ ....-:.:e 30 Minimum 20 I " . , 10 Mean 5 PM , -6PM o 1 PM 2 PM 3 PM 4 PM Hour Figure 2. Proportion of the maximum number of swimmers observed, by IS-minute intervals (6 observations, Town of Southold, 1964 and 1965). -6- . This would be satisfactory provided all boats were stationary, i.~., would have been observable from only one point. Of all participants and conveyances recorded, boats violated this premise most severely. Most boats recorded were moving, thus though duplicate counts of a given craft during one observational route count was avoided, expansion from a day's continuous count data would yield much higher estimates of boating trips than actually occurred for a far greater area was "considered" by counting at a ste.tionary point. This is an unknown area but potentially is as great as tha.t covered by a boat in a twelve-hour period of travel. In addition, boats leaving and then returning would be observed as two crafts. For these reasons a factor of only 5 was applied to all boat observational counts. This may still be excessive. To derive boater participant days, the resulting boat trip figures were increased by 3.56 (average number of persons per boat as revealed by personal interviews). Boats using town waters and all other participants are underrepresented by the proportion using town areas from Plum Island eastward. In the case of' boats this is substantial in number, although few, if' any, of those not noted would have importance to the remainder of' the town, probably largely coming f'rom the north shore of' Long Island Sound or the vicinity of' Montauk. Shore fishermen. Data on daily periodicity of shore fishermen partici- pation was less uni:form than :for swimmers. On three days when any :fishing occurred at continuous observation sites peak :fishernan numbers occurred at 11:15 a.m. (and 11:30) in one case, 3:15 p.m. in another, and 7:30 p.m. in the third. The maximum percentage o:f shore :fishermen observable at any instantaneous count between 2:00 and 5:00 p.m. on those three days were 20, 44 and 0 percent respectively -- an average of' 21 percent. The average proportion of' shore :fishermen observed in any two-hour period between 2:00 and 5:00 p.m. was 20 percent. Numbers o:f shore :fishermen observed were increased by a :factor of 5 to yield daily participant :figures. Picnickers. Lacking more detailed estimates, all picnicker numbers and park users were increased by a factor o:f 4. Clammers. The timing of stages of the tide seemed to have marked in:fluence on numbers of clammers, there:fore observed clammer numbers were doubled. Hunters. Dates of observation of hunters included the first three days o:f the--upland game season on Long Island in 1965 (November 1 - 3). The goose and "sea duck" season was also open during this period. In 1964 observations were run November 26, 27, 28 and December 27,29, and 30. On Long Island small game and "sea duck" seasons were open during these periods but the general waterfowl season was open only during the December observations. -7- . Duck hunter activity (in boats or on exposed points) typically could be noted by driving roads or from observation points, for normally a car was visible nearby. Upland game hunters were less noticeable and often walked to hunting areas. Cars located in areas known or believed to be used by hunters were recorded. Combining all segments, an average of five cars loere noted on weekend days, and an average of 11 cars noted the first three days (weekdays) of the upland game season. Data from other portions of this study indicate an average of two hunters per car. Thus 10 and 22 hunter days respectively are indicated. Personal interviews indicated 29 percent of the hunters did not use a car to get a hunting site. Expansion by thIs fac';or leads to a dai17 average of 13 and 29 hunters for these two periods. In addition, certainly some hunters in cars would be unobserved, perhaps an equal number to that noted above. This leads to an expansion to 26 and 58 hunters for weekend days in the season and the opening three days respectively. Similar calculations for weekdays but not at the small game season opening reveal an average of only 16 hunters afield on an average day. A 6l-day season with 19 weekend (and holiday) days, 3 "season opening" and 39 regular weekdays would yield 2 ,584 hunter days. Assuming that instantaneous counts account for only half of the hunters afield, approxi~ mately 2,500 (2,584) hunter days would occur in the town. Hunting seasons on various minor species are open in this town from September 25 to January 31. A minimum of 500 more man-days are doubtless spent hunting during those periods, giving a total of hunter days in excess of 3,000. Discussion of Techniques ~ Figures on participation used in this study are based upon direct enumeration plus expansion by appropriate factors. Under conditions existing I believe these can yield satisfactory estimates of m~3t activities studied in this town. The accuracy of such estimates needs to be, could have been, improved had observations been more adequately stratified and sampled by various weather patterns and if additional "peak hour" observations had been possible. The error possible when expanding counts by factors as great as five is apparent. The need for sampling inclement and "good" weather separately seems apparent when one concludes that a rainy, windy day will entice fewer outdoor recreation participants than a sunny, calm day. This factor was not considered in any of the studies of recreation participation reviewed by this author. Presumably the historic tendency for recreational research to build upon studies of fishermen, hunters, and campers is partly responsible for this glaring omission. These activities are probably among the least affected by a particular day's weather. Fishermen numbers in this study were less depressed on inclement weather days than participant numbers in other activities. . -8- For activities in which those not members of the labor force (students, housewives, and other on extended vacations) form a large portion of the participants, weather probably causes as great a variance in the extent of participation as weekend days vis-a-vis weekdays. In determining total participant days it is evident that some activities (fishing from rental rowboats, charter boats, and open boats) typically occupy persons for extensive portions of the day. For these activities, instantaneol1s counts at approjJriate time periods do provide a valid measure of vi2"<;ua11y ,,11 partic:i.j::mts. For activities typically involving only a few h.,'..,rs of the day (dl othl':' activities studied in this project), determi.;:"tion~:J' all partidpan\:J ::-e'l'.::,"'es determining proportions of total participants active at varied time periods. For those activities having rather predictable distribution of participation over time (either unifol':n peG.ks or steady levels) the total participants can be determineJ if the proportion represented at given time periods is known. This appears to be the case with swimmers in this study. IVhen participation fluctue.-:"'8 widely through the dace with no predictable pattern, total partici.pation in an activity i& More d.i.fficu1t to deter- mine. Shore fishermen and boaters appeared to repre3ent this pattern in the present study. Recording boats moving into or out of observable areas would probably yield a better estimate of total boater trips than the instantaneous counts used in this study. This could be accoPlYJlishc1 over a broad area through use of time 1cpse cameras. One such c,ccera U;m.-ardin and Ashe, 1965) was used on a single day at a beach. Se'" to tnice one picture every 15 minutes, the resultant pictures can include a rath"r large area (Figures 3 and 4). Set to record at shorter time intervals this device could replace an observer at many sites. In the on. attempt at its use by the author the camera housing was so large as to~e noticeable. It was apparent that swimmers avoided locating within its lateral range of coverage. I feel this device in a minie~urized, well concealed form has great potential for deriving information on participation in many activities, for a limited area. Participation in recreational activity has been estimated by expansion from car counts (James, 1966). In three activities considered in this study, 30 percent or more of the participants did not use cars to arrive at a site. Estimation of participants by expansion from car counts would have been subject to substantial errors in these activities, if uncorrected. In other activities, participant cars made up a minor percentage pf the total number of cars present. Equally useless estimates would have reSUlted. State Park use was the only activity in which over 90 percent of the participants arrived by car, and the cars of these participants comprised virtually all of the cars at the site. The superintendent of the park (Tabor, personal conversation) reported that errors and operational problems had resulted in discontinuing the use of pneumatic car counters in that park. For these reasons, uncorrected car counts do not appear 9 Figures 3 and 4. Examples of photographs taken with time-lapse camera on a public beach, Town of Southold. ~lO- usefUl as measures of participants in most of the activities studied, under the conditions existing in this town. . Aerial observation consisted of flights along the shoreline areas at an elevation of 500 feet in a fixed-wing aircraft. Minimum air speeds were maintained; all persons on the shoreline or in the water, and all manned boats were recorded by two observers. Binoculars were used to scan shoreline areas. Cost of the pilot and plane was $25 per hour. The entire mainland town's shoreline of 65 miles could be observed in one hour. Boats were readily perceived from the plane. Participants on shore were equally obvious and where sparse, readily recorded. Fishermen in these areas, too, were quite noticeable. Problems were encountered in accurately enumerating persons where large numbers of participants congregated. Use of aerial photographs of high resolution doubtless could reduce that problem. In addition, persons in the water (swimming or clamming) were difficult to notice. Although not attempted, aerial observation would appear unsuited to enumerating hunters under conditions existing in this town. An additional severe limitation -- inclement weather -- makes flights difficult or hazardous to undertake. . Primary benefits of aerial observation were speed of coverage, indication of minor areas of use, and boat enumeration. Speed of coverage would have permitted 20 observations of the entire shoreline area at peak activity hours on 20 days at a cost of $500. One hundred twenty days would have been involved in developing the same observations from land. Even then, activity on many low intensity use areas unobserved from land would have been unrecorded. The value of obtaining larger numbers of observations (at reduced costs) and of stratifying observations by weather can be clarified by exploring the confidence levels one can put on the participant day estimates derived in this study. Table 11 gives confidence limits (.95 level) for the point estimates previously presented. The wide limits are a function of the wide fluctuation in use on a given type day, the double stratification used, and the restricted number of observations utilized. -11- Table 1. Confidence limits* of participant day estimates for seven outdoor recreBticn.activities, avera~e for the summers of 1964-65. Town of Southold . Participant Dll3's in Thousands Participants Point Estimate Confidence Limits Swimmers at public beaches 304 Swimmers at private beaches 190 Boaters 181 State Park users 29 Shore fishermen 18 Picnickers outside the State Park 13 Clammers 3 t 125 t 102 % 103 :!: 13 % 8 :!: 12 t 3 * At .95 level. The wide variation in use on a given type day reflects great variability in daily use, as influenced to a substantial degree Qy weather within the "good" and "inclement" weather days. Thus the value of further strati- fication on the basis of weather is evident. This would lead to another problem -- a problem reflected in the present study where four type days were used. The use of these, while permitting greater accuracy, forces an expansion of the confidence intervals, for there are fewer observations in any cell. As a result of experience gained in this study it is suggested that a general procedure to follow in determining the level of recreation participation should include the following: 1. Determine from knowledgeable persons the activities of major prominence in the community and the location of high intensity use areas 2. Determine, through use of time lapse camera or observation, the patterns of activity over a 24-hour or dll3'light period on those areas during several days 3. Establish land or boat observational routes, time lapse camera stations or aerial observation routes, the method chosen -- depending on the intensity and predictability of levels of participation 4. Develop instantaneous counts, using a sampling scheme stratified on the basis of weather, and by weekday or weekend day . -12- Discussion of Participant Interviewing The primary problem encountered in determining socio-economic and attitudinal attributes of participants was the selection of a representative or random sample. For very limited areas (one beach, a marina or pier, one picnic area) this problem would be relatively easily overcome. Personal interviews stratified by time of day, day of week, and weather would yield an unbiased estimate. For participants in one activity in a larger area (a town), serious complications evolve. Stratification of sampling by time of day, day of week, and weather, as noted above, must be carried out and, in addition, stratification by intensity of use should be established. Thus all beaches receiving a predicted number of swimmer days in a year (for example, 1 - 999, 1,000 - 2,999, 3,000 - 9,999, and 10,000 or more swimmer days) should be stratified and sampled independently. Interviewing of participants in other recreational activities should be handled in a similar manner. The complexity of this would seem to require computer techniques for drawing a sample. Lacking these refinements, the sample drawn in this study more closely represents a judgment than a random sample. It thus does not permit meaningful statistical tests of the data derived. If one were to derive a random sample, statements could be made with statistically significant associations. Presuming random samples were obtained, extremely large samples would have been needed to compute chi-squares within residency groups for the many categories used. This large size would be needed to make a chi-square an adequate test, five cases needing to be present in each cell in order to use that tool (Kish, 1965). The samples derived are sufficiently large if one collapses categories. Thus, considering the "X-2" tables, one could speak of persons of age 40 or less vs. those 41 or over. The sampling level used in this study (if it were a random sample) does permit rejection (at .95 level) of a hypothesis that these age groups are equally represented among residency groups among boaters or among swimmers at public beaches. A hypothesis of equal representation of those age categories between boaters and swimmers at public beaches would also be rejected. The number of interviews which need to be taken in order for these statistically significant differences to be stated varied with the distribution of the responses. Considering the primary town attractant, for instance, where one item -- Rustic Aura, in the case of State Park picnickers -- is identified by 65 percent of the sample as of primary importance; a random sample of 87 persons is needed in order to have a limit of error of 10 percent (at .95 confidence level). As many as 96 -13- . persons would need to respond to attain this same limit and level of confidence if the responses were equally represented (50-50). If the proportion in our sample giving a specified answer to our question increases to 90 percent (such as rental rowboat fishermen -- 92 percent of whom are visitors --) we need only derive the information (on residence) from 35 persons to attain the same level and degree of confidence. In the case of rental rowboat fishermen (with 143 interviews) we can say (with .95 level of confidence) the actual proportion that are visitors is 87 _ 97 percent of the total. Note by increasing the sample size we are able to narrow our confidence interval (to 5 percent in the rental rowboat fishermen example) (Parten, 1950). With the information derived in this study, others presumably can establish more discrete, yet restrictive, categories for analysis. It should be emphasized that the categories used would vary between activities. It is my feeling that the data derived are reliable, particularly where substantial differences are indicated within or between residency groups in an activity, or between activities. When a group was encountered in my study only the decision-maker was to be interviewed. The person chosen was the apparent decision-maker (!.~., if a family group it would be the father). On occasion some other individual in the group would attempt to preempt the respondent's role and this was taken as indicative that she (or he) was actually the decision-maker (thus a "revealed decision-maker"). This occurred in fewer than 5 percent of the interviews. Perhaps appropriate techniques to substitute the "revealed decision-maker" should be used. This study clearly revealed that participants are generally willing to respond to an interview. Fewer than 3 percent of participants in any activity refused to be interviewed. Typically these were individuals just about to leave a site and "late already." Others willingly partici- pated in the three to ten minutes the interview required. It was the interviewer's feeling that ten minutes was near the maximum time most respondents would have spent answering the questions posed. Those interviewed were typically in a group, often having nonfamily members present. It was my feeling that when "outsiders" were present, responses critical in nature were often softened and at times not stated. I thus feel the true level of critical replies cannot be ascertained from the technique used. A possible means of minimizing this error may be through use of a card containing numbered probable responses. This approach was used to obtain family income levels and was successfUl in deriving income approximations from 90 percent of the adult respondents. It was evident this high level would not have been obtained if a direct statement of income had been required. Interviewers were dressed in casual street wear, feeling that this was appropriate for most activities and that this lent a business-like atmosphere to interviews on beaches. -14- Personal interviews with hunters were considered unsuccessful due to the low level of participation encountered. Mail questionnaires (Appendix) may provide a more satisfactory approach to these participants. Personal interviews were capable of eliciting information on users of this town's outdoor recreational resources (except for hunters). This type of interview was superior to household interviews in its ability to derive information on visitors out for only a day, a substantial number of the participants in most activities studied. Further, these interviews lend themselves to deriving comments on a specific area from a large proportion of users of that site, a difficult task in household interviews. It is believed information such as that derived by this study can be confidently used by groups planning for a community's future development. -15- LITERATURE CITED Cowardin, L. M., and J. E. Ashe. 1965. An automatic camera device for measuring waterfowl use. J. Wildl. Mgmt., 29(3):636-640. Kish, L. 1965. Survey sampling. John Wiley and Sons, Inc., New York. 643 p. Parten, F. S. New York. 1950. Surveys, polls, and samples. 624 p. Harper and Brothers. Wilkins, B. T. 1967. Outdoor Recreation and the Commercial Fishery in the Town of Southold. Dept. of Conservation, Cornell Univ., Ithaca, N.Y. 129 p. . -17- Confidental Confidental Recreation Impact Study Conducted by Cornell University In Cooperation with the Township of Southold Dear Sir: This questionnaire is part of a research project intended to help the town determine the extent, and impact, of hunting in this community. The enclosed questionnaire, completed by the driver of your car, can be a big help in answering this question. . Two minutes of your time spent filling out or checking the appropriate answers, sealing the sheets inside the envelope and dropping it in a mail box, may help yield more and better hunting for us all in the future. Thank you for your help. Be our guest, keep the pencil! Sincerely yours, Bruce T. Wilkins Department of Conservation Cornell University BTW : sh 11/65 . -18- . Confidental Southold Recreation Study (To be filled out by driver) 1) Did you come here mainly to hunt today? Yes No. Why are you here? if a "no" answer and return in the envelope. (Stop here Thank youl) 2) What were you mainly hunting today (check one)? . Pheasants _Ducks _Quail _Squirrel Rabbits Other (Please describe) 3) How many of each did your group get? _Pheasants _Ducks _Quail _Squirrel Rabbits Other 4) What time did you begin hunting today? 5) How long did you hunt today? Hours 6) How many shots did your group fire? 7) How many days did you hunt last year? o 1 - 3 4 - 6 7 - 14 15 or more . 8) Do you live in the town of Southold all year long? , No. Yes. (If yes go to question 14 on next page). -19- 9) How many weeks will you or your family spend in the town of Southold this year? · less than 1 week 1 - 2 weeks 3 - 10 weeks 11 - 24 weeks - Over 24 weeks 10) Please check where you spent last night. Permanent home Seasonal home Relative or friend's home - Motel, hotel, etc. ===== Other (please describe) 11) About how far is your );ermanent home from this spot? . o - 9 miles 10 - 50 miles 51 - 125 miles ===== 126 or more miles 12) To determine this trip's cost, please indicate how much your group has spent on this trip and that part spent in the town of Southold. , Amount Part spent in Southold (all, 1/2, none, etc.) Meals Lodging Shells other (Please Describe) 13) How many years have you been coming to the town of Southold? 14) How many years have you hunted in the town of Southold? These next questions are simply to give us a clearer idea of who hunts in this town. . 15) In what year were you born? . 16) What is your occupation? 17) How many were in your car hunting today? 18) Of those in your car how many are members of your family? THANK YOU FOR YOUR HELP!