U.S.
Department of Commerce
International
Trade Administration
Office
of Travel & Tourism Industries
Survey
of International Air Travelers (IFS)
Section 1: Introduction Page 2
Section 2: Background and
Description of Current Program Page 4
Section 2a: Questions on the
Current Program Page 24
Section 3: Ideas to Improve the
Current Survey Program Page 33
Section 4: An Alternative
Methodology of Data Collection Page 40
Section 4a: Questions for
Alternative Methodology Page 45
Section 5: Open Area for
Suggestions and Ideas Page 50
The U.S. Department of Commerce has conducted an international air travel survey since 1982. This program has changed dramatically over the years to adapt to changes in the global economy and the travel industry. Commerce requests that the industry review information on the current program, review proposed options to improve it, respond to alternative methodologies, and provide comments on the several proposed changes to this program. Your input will be used to assist Commerce with the development of the next contract for this program. Industry input is also being requested because in order to assist the industry, Commerce needs to know what industry needs and what concerns exist.
While the program has been in existence for 20 years, few in the industry have a complete understanding of the In-Flight Survey. For the purposes of this RFI Commerce has provided background on the program. OTTI expects that the industry will review the information and provide feedback, i.e. questions or comments will result. Commerce will respond to all questions and post questions and responses on the FEDBIZOPS at: www.fedbizopps.gov
It is important to know that all comments and responses and ideas should be submitted to the following points of contact no later than June 11, 2005:
Ron Erdmann, ron_erdmann@ita.doc.gov
Phone: 202-482-4554; Fax: 202-482-2887
Richard Champley, Richard_champley@ita.doc.gov
Phone: 202-482-4753; Fax: 202-482-2887
Rubie B. King, Contracting Officer, Rubie.B.King@noaa.gov
Interested parties are also encouraged to provide the Department of Commerce with information on their company along with detailed information on how they would approach conducting this program.
Any ideas the industry may have to assist the Government improve the data on international travel to and from the United States will be considered by OTTI before writing the next statement of work for its survey program on international travelers. There is also an open area for suggestions, comments, concerns, issues, questions, other methodologies OTTI should consider using, and/or any other ideas the industry may want to share with Commerce. Please provide them in this open area.
I. BACKGROUND
The U.S. Department of Commerce, U.S. Travel and Tourism Administration (USTTA) and the U.S. Department of Transportation all contributed to the development of an in‑flight survey of international air travelers, which was implemented in the fourth quarter of 1982. The initial technical survey planning, design and operations through calendar year 1983 were conducted by the Transportation Systems Center of the U.S. Department of Transportation. USTTA administered the program from 1984-1996 and in1996 the agency was closed by Congress.
The research functions of USTTA were moved to the U.S. Department of Commerce, International Trade Administration, and Office of Travel & Tourism Industries (OTTI). The primary purpose of the survey has been to provide the USA, via primary research methodologies, with travel and passenger fare export and import figures. Also, a goal was to provide timely, accurate, and detailed data on international air travelers to and from the United States. The data collected from this survey describes the characteristics and behavior of overseas and Mexican visitors traveling to the United States and of U.S. residents traveling abroad and to Mexico. Canadian travel data is not currently included.
This information is designed to assist the various agencies of the government (specifically, the International Trade Administration, and Bureau of Economic Analysis), participating airlines and the travel industry in defining international travel patterns and calculating the economic impact of international travel. Other uses of this data include developing appropriate target markets, to compare visitation to different destinations, to assist in planning and marketing strategies, and to develop an overall profile of the international traveler [see section F. Data Output]. The survey results also serve as a valuable technical assistance tool enabling the OTTI to reliably monitor U.S. international travel and tourism trends and respond to information requests.
The travel research program has been developed to support the major goals and sub goals of the Department of Commerce. The three salient Departmental goals include:
· Increasing America's competitiveness in the world economy
· Increasing U.S. exports of goods and services (international travel and tourism is a service export)
· Increase the U.S. market share of the total global travel market.
It also supports the Department goals related to the National Export Strategy report issued in October 1996 (and updated yearly) by the Trade Promotion Coordinating Committee (TPCC). Goals related to the TPCC include: “Providing timely and accurate international tourism statistics” and “Providing the foundation for economic policy decision making by assuring the travel and tourism industry is accurately measured.”
Data from this program are also used to support the U.S. Department of Commerce, International Trade Administration, Manufacturing and Services, Office of Travel & Tourism Industries. The other main Commerce user of this data is the Bureau of Economic Analysis. The Bureau of Economic Analysis uses the In‑Flight Survey (data files and reports ‑ inbound and outbound) to develop its "International Travel and Passenger Fares" estimates for travel and passenger fare exports and imports issued in the Survey of Current Business. For additional information on the Department of Commerce agencies, please see the addresses below:
http://www.trade.gov/td/td_home/tdhome.html
http://www.tinet.ita.doc.gov - OTTI website
To understand the overall statistical results of the In‑Flight Survey it should be kept in mind that the sample was originally designed around the specific airlines willing to participate in the survey. It evolved to include boarding area surveys. Determination of airlines and boarding areas was performed through an analysis of U.S. and international traffic to/from world regions and countries.
Survey data are gathered from passenger self‑administered questionnaires, which are primarily distributed and collected in‑flight by airline flight crew personnel. Certain participating carriers prefer not to use the in‑flight approach to the Survey. Thus, a boarding area methodology is used. Participation is voluntary, and thus the contractor and OTTI rely heavily on the cooperation of the airlines. The 1990-2003 carrier participation in this program is provided on the OTTI website. A great deal of emphasis is placed on developing a working relationship with the contacts at each airline to ensure they actively participate in this program. In return for their cooperation and participation, the carriers receive confidential reports, on a quarterly basis, profiling the characteristics of travelers aboard their flights as compared to the aggregate responses of all other carriers. The participating carriers also receive non-confidential inbound and outbound “national” reports on a quarterly basis, if requested.
II.
SYNOPSIS OF CURRENT SURVEY OPERATIONS/REQUIREMENTS
A.
Survey Population and International Airline Participation
The population of potential Survey respondents consists of all international revenue air passengers who travel on a participating airline’s departure leg from the United States. Overseas and Mexico air markets are included, however the Canadian travel market is currently excluded. OTTI uses the Department of Homeland Security (DHS), Office of Immigration Statistics I-94 (inbound non-residents) and I-92 (departing US-citizens) databases to develop the population estimates. It should be noted that non-residents are surveyed on their return flight and US-citizens are surveyed on their originating flight. Therefore the two segments are essentially surveyed on the same flight departing a US gateway. These databases are integral to the expansion of the survey results to ensure they are representative of the entire US-international travel population. See the travel population time line from 1990-2003. The April 2004 projections for 2004-2007 indicate a five to six percent growth rate each year.
The selection of an airline for participation in the survey includes the following factors: the airline's market share in the geographic area under consideration, the desirability of representing both U.S. and foreign flag carriers for each area, a carrier's willingness to participate and the necessity to keep administrative and operational costs to a minimum.
In 2003, over 60 U.S. and foreign flag carriers participated in the survey. While the survey is conducted monthly, airline participation varies by carrier and is based on individual verbal airline agreements with the ITA, Office of Travel & Tourism Industries.
Between 1996 or 2003, OTTI has requested a maximum of 6,449 flights to be sampled to a low of 2,522. The airline response rates (flights requested to be surveyed compared to actual flight kits returned with completed surveys) have been between 73% and 41%. Based on information from cabin crews the passenger response rates have been as high as 62%, and as low as 41%.
The
primary geographic units of analysis are both the “true” origin and
destinations (O&D) of the survey population. The primary world units are based upon nine world regions. The United States is primarily divided into
11 modified Census Divisions. Statistics are also developed for states, large
metropolitan areas, and selected major tourist attractions to the extent that
response frequencies for these places permit.
Mexico is treated separately, and Canada is not included in the survey
at the present time. Click on the links
to see the states that comprise each census
region, and the countries included in each world region
used for this program.
B.
Sampling
The sample universe consists of all international air travelers, but selections are made on a ‘cluster basis’ based on a random sample number of scheduled flights (clusters) from among all participating airlines that depart the USA to overseas or Mexican destinations. Certain charter carriers are included in the program at Orlando Sanford Airport, which reports 99% of its total traffic from charter airlines. (Charter carriers at other airports could be added in the future. Charter traffic accounts for less than four percent of all international overseas and Mexican arrivals to the USA.)
The flight clusters for each participating carrier are identified from the electronic Official Airline Guide (OAG) database of scheduled airline flights. A computer program draws a random sample of flights from among the survey week's flight list.
The survey period is defined as the seven-day week beginning on the third Monday of the month each month. The week of the month varies during holiday periods. OTTI and the contractor strive to ensure that the survey operation mirrors the airline’s own system as much as possible.
The total population consists of two strata. One stratum is comprised of U.S. citizens departing the United States on the originating leg of their flight. The other stratum is defined as non-resident visitors (per the I-94 record) on their return leg home who are aboard the same outbound flight. The number of flights sampled from each stratum is determined on an individual airline basis. Consideration is given to the number of passengers carried, volume of flights to foreign destinations, number of participating U.S. and foreign carriers serving the area, and airline cooperation. The stratum covers all airlines participating in the survey. An airline which is the sole participating carrier serving a particular country or region would have proportionately more of its flights selected than an airline which is one of several serving an area.
Therefore, the sample design is a stratified cluster sample. From the population of total international passengers the flight clusters (flights departing the U.S.) are randomly selected from the OAG list of participating carriers. Travelers in each stratum (resident v. non-resident) are asked to complete questionnaires, all passengers 18 years of age and older are eligible to respond to the survey. Since questionnaires are distributed to passengers throughout all aircraft cabins (first, executive and main) the responses are considered to be representative of all passengers on the flight.
Random selection by flight (cluster) and stratification of the population, based on residency, help to ensure that all participating airlines and all travelers, regardless of residency and carrier ‘flag’ have an opportunity for selection into the sample. Regarding the passenger mix by airline flag (i.e., U.S. or Foreign), OTTI observed that, not surprisingly, citizenship appears to be a factor in the choice of airline. For example, from 2003 I-92 data (Table IId.): 67% of all passengers flying on U.S. flag carriers were U.S. citizens (even though they comprised only 58% of the total market) and 50% of all passengers on foreign flag carriers were non-U.S. citizens (even though they comprised only 42% of the total market). This ‘share gap’ phenomenon needs to be factored in during the sample selection process.
Due to the voluntary nature of this survey, a pure random sample selection process of passengers on a flight would be logistically difficult for a flight crew to administer. Therefore, OTTI relies on the flight crew’s judgment in distributing surveys to eligible (over 18 years old) passengers. Also, since various participating airlines conduct their own in-flight surveys from time to time OTTI has adopted a standardized, but yet flexible approach with each carrier so as not to unduly burden the flight crew. OTTI works with each carrier to ensure the survey operation fits into its system. Most of the carriers distribute 50 to 125 questionnaires among all classes of service; the number of surveys provided is set by each airline.
Statistical reliability and validity of the data are dependent on many variables – the amount of bias in the selection process, sample sizes and the ability to quantify results. For example, how reliable are the estimates (degree of confidence, i.e. 95%) and with what precision (given an error rate of +/- percentage points). OTTI believes that the reliability of information and output data from the Survey is dependent on relative sample sizes and data validation and quality control. The latter process is described in detail in section E. ‘Data Entry, Editing,’ OTTI policies require that information will be released only if it meets the minimum acceptable standard for a sample size, i.e., 100 occurrences. Information presented at summary levels is assessed to be very reliable. For example, in 2003, of the 29,114 non-resident respondents from overseas, much can be said about the travel characteristics and travel patterns of the 46 percent who visited from Western Europe. However, there is little data and much less certitude about the characteristics of a visitor from Bhutan. In 2003, OTTI collected over 100 respondents for non-resident inbound travelers from 43 countries, 35 states/territories as destinations, 58 cities as destinations, and 20 U.S. airports. The sample collected for U.S. residents traveling abroad in which at least 100 respondents were collected were for 60 countries visited, 38 states of origin, 47 U.S. cities of origin, and 20 U.S. departure airports. To learn more about which states, cities, countries or airports are reported, visit the OTTI website and click on the “Inbound Travel to the U.S.” or “Outbound Travel from the U.S.” links.
The number of responses received those sampled among all travelers originating from overseas markets in 2003 was 29,114. This represents approximately 0.2 percent of the total population of travelers, which was over 18 million. The question is “How much is enough given a specific objective?” For example, how large a sample would need to be drawn to allow for a reliable conclusion to be made about the characteristics of a traveler from Paris who visited St. Louis? OTTI would appreciate the industry’s perspective of this issue as part of the response to the RFI.
Once the appropriate sample sizes are determined the real challenge will be to implement a more effective administration of the sampling process to: 1) dynamically adjust the number of survey kits distributed to the airlines, 2) improve the return rate from the carriers, and 3) increase the response rates from passengers. (More in section D.)
C.
Questionnaire Design and Content
The In‑Flight Survey is in the maturity phase of its program life cycle. The questionnaire is a government document controlled by the Office of Management and Budget (OMB). The survey instrument, OMB No. 0625-0227 must be submitted to OMB for renewal every three years. Its next expiration date is 7/31/2005. During this process OMB will also assess the viability of the program. It is expected that the resulting assessment will be positive.
The questionnaire developed was guided by industry standards of primary research survey instrument design to encourage accurate and maximum responses by the surveyed passengers. The airlines and other travel industry sectors have historically provided input on the content of this survey. The advice of airlines participating in the program is particularly appreciated, since without them there would be no in-flight survey research program. The survey instrument has been re-designed with travel industry input in 1990, 1993, and 1996. OTTI will work to obtain industry input on any changes made to the current survey instrument. When the survey was last revised, over 100 industry representatives provided input on the survey re-design. OTTI’s airline advisory board, several private sector representatives and staff from the U.S. Census Bureau Questionnaire Unit were used to develop and review the existing survey instrument.
To
reach the majority of non‑English speaking travelers, the questionnaire
was translated into eleven additional languages ‑ Arabic, Chinese
(Mandarin mix), French, German, Italian, Japanese, Korean, Polish, Portuguese,
Russian, and Spanish. There is an
English‑only version and eleven versions with English on one‑half
and a foreign language on the other.
Other language adjustments are possible.
The survey questionnaire is designed to obtain responses from both U.S. as well as non-U.S. residents and can be self-administered by the respondent. An announcement at the bottom of each survey informs the respondents of the availability of other versions. The language mix of questionnaires used on selected routes is determined by the destination served. The questionnaire design facilitates easy distribution and collection by eliminating the need to determine the citizenship of the passengers.
Click here to view the questionnaire.
The distinguishing features of the current questionnaire are:
· Trip itinerary - complete origin and destination (O&D) itinerary (questions 2-5 and 14) including U.S. and foreign port data. Destinations visited requested includes up to seven places visited in priority order (question 14). Many airlines have found this information useful for route planning purposes, as well as question two, main destination.
· Trip specific data - trip purpose (main and other), type of airline tickets purchased, aircraft cabin, reason for airline choice, travel party composition, nights away
· Expenditure detail (question 15 – 16) and
· Information on stays at specific hotel chains, type of credit cards used (note: credit card numbers are not asked), and name of the rental car agencies used. Currently, none of this data is released to the public. Should OTTI? Should limits be put on who could obtain this type of data? Would hotels, credit card companies, or rental car firms want their competitors or partners to have access to this type of data?
· Traveler characteristics: age, gender, family income, residency, and country of birth, previous international travel history
An industry standard coding system is used to identify airports, counties, cities, and states within the USA, as well as all international cities and countries. The system enables OTTI with the ability to use data-mining techniques to search and select-targeted data elements to identify a variety of destinations (cities, states, zip codes, countries, attractions, and National Parks) and ports (U.S. and foreign). The coding system for this program is extensive. There are over 130 airline codes, over 400 U.S. port codes, over 370 foreign airport codes, over 200 inbound and outbound country codes, there is a code for every U.S. state and territory, OTTI uses the Census FIPS codes for U.S. cities and counties, there are over 860 attraction codes, over 160 credit card codes, over 350 hotel codes, and over 60 rental card codes.
D. Field Administration
The Survey is administered in either one of two ways: on- board during flight (or ‘in-flight survey’) or in the airport gate boarding area.
The basic monthly survey cycle, starts 30 days before the actual survey week (generally the third week of the month), and is as follows: OTTI’s contractor selects the flights to be sampled, based upon a computer generated stratified random sample using the flight information from the Official Airline Guide. The proposed flights to survey for each airline are sent to an Airline Service Manager (ASM). The ASM is appointed by a participating airline and acts as the contractor’s primary contact for each airline. The objective of using an ASM is to minimize the costs and workload of airlines and to handle any problems as they arise (for example ‑ schedule changes affecting the flight selected for survey). The ASM will approve the flights and notify the contractor if any changes need to be made. Ideally, the ASM would also assist the contractor in tracking down flights that were requested to be surveyed that were not. The ASM also serves as the central contact points for each carrier should anyone at that airline have a question about the program.
The contractor then express mails the designated flight kits to each airport. At each airport, there is a gateway manager (GM). They are local contact used on the boxes of surveys sent to each airport. Within each box, a flight survey package containing the questionnaires, pencils, and instructions for the flight crew for each sample flight. Each flight kit is packaged as a nylon zippered bag that is 19 inches by 14 inches and weighing between six (6) to twelve (12) pounds, which has cost implications for the program.
The contractor follows up to ensure the kits were received and to handle any problems, regarding lost survey kits, or issues that may arise. In 2003, the loss of survey kits was approximately 12 percent. This was on par with previous years, but much less than in 1989 when the loss rate was 24 percent. (The experience of fewer losses appears to be a result of the shift of methodology from primarily in-flight to more of an airport intercept method).
The gateway manager notifies the crews of which flights are to be surveyed, ensuring the flight crews distribute the survey kits. The flight crews carry the flight survey packages to their airplane. Included in the flight kit package are instructions that the crew distribute the surveys to passengers throughout all cabins, answer any questions, provide pencils to passengers who may need them and collect the surveys back from the passengers. Flight crews are to place the surveys – both completed and not completed - back into the survey kit bag. They are asked to ensure that the survey kit is returned to the same USA gateway on the next flight (usually in the cargo hold of the plane) to the attention of the GM. The gateway manager then mails the kits back, postage pre-paid, to the contractor for processing
Currently, there are 59 ASMs, and 83 GMs. In 1995, there were 72 ASMs, 25 ASM/GMs, and 180 GMs. The lower number of contacts today is a result of the shift from an in-flight method to more airport intercepts being conducted. An airline may have more than one ASM if their company has divided the USA into regions, and they require that the contractor’s staff contact the regional ASM for approval of the sample, etc. Since each airline has different gateways and procedures for administration of the survey, the number of GMs per airline varies greatly.
Currently, over 60 carriers voluntarily participate in the program. A list of the airlines participating in the In-Flight Survey can be seen on the OTTI web site
http://www.tinet.ita.doc.gov/research/programs/ifs/carriers.html.
In 1995, the two survey administration methods were divided as follows: forty-seven (47) U.S. and foreign participating carriers used the in‑flight method and thirteen (13) airlines required the surveys to be administered to their passengers in the pre-departure boarding areas. While 22 percent of the carriers used the boarding area method, only six percent of the total surveys for the program were collected using the boarding area methodology. This difference reflects the fact that many of the larger carriers are still cooperating with the in-flight method.
By 2003, the survey distribution method was divided into 20 U.S. and international carriers participating in this program use the in‑flight method, and 40 airlines the surveys to be administered to their passengers in the boarding areas. The shift has had major cost implications. There are positives and negatives for both methodologies, and OTTI has used both methodologies since 1987. OTTI’s goal is to keep the survey representative of world travel patterns, and if an airline will not participate using the preferred. In-flight survey method, the boarding area survey method is then used. .
OTTI’s goal is to survey an appropriate number of travelers that will allow reliable statistical inferences to be made about destinations visited. In actuality the sample size will probably have to be increased to accomplish this goal, i.e., from the current 0.2 percent to something approximating one percent sample of the travel population. However, that percent is yet to be determined. In the near future steps will be taken to improve on the efficiency of the current process by:
· Increasing the return rate of the survey kits
· Increasing the response rates of the surveys distributed
Accomplishing these steps will effectively increase the sample size from the current 0.2 percent. However, funding has been a major issue for this program, and until a dramatically larger funding base is obtained, OTTI will have to survey the same, or even a lesser, number of travelers as has been the experience in the past.
The number of flights surveyed by the in‑flight methodology or the boarding area (or airport intercept) methodology has varied over time, as has the number of flights surveyed. A table with the number of flights surveyed per year and the airline response rate is provided, along with the number of surveys collected each year. The number of flights surveyed per carrier on a yearly basis it ranges from 24 to 280. To obtain the target sample each carrier’s sample frame will be adjusted on an individual airline basis. Over the years, approximately 17,500 to 22,500 (plus or minus 2 percent) questionnaires are considered completed and usable for analysis per quarter, although the goal is to collect questionnaires from 70-90,000 or more respondents over the year. Because of the seasonal travel patterns of international travelers, the third quarter tends to be the largest survey period. The ratio of resident to non‑resident questionnaires has historically been 1.2 to one.
Since the inception of the boarding area methodology in 1987, OTTI has surveyed at the following airports: Chicago, Cincinnati, Dallas, Guam, Houston, Las Vegas, Los Angeles, Miami, New York (JFK), Newark, Orlando’s Sanford, Portland, San Francisco, and Washington’s Dulles airport. In 2003, flights were surveyed at the departure gate for specific fights following airports: Chicago, Dallas, Guam, Houston, Miami, New York (JFK), Newark, Orlando’s Sanford, and San Francisco’s airport.
After the acts of terrorism on September 11, 2001, airport security has become a major issue in terms of obtaining access to airports so that the surveys may be conducted at the international departure gates. Additional clearances from the airport authorities and the DHS, Transportation Security Administration had to be obtained. This increased both the time to gain access and the total costs of conducting the surveys at the airport. It is expected that OTTI will continue to work closely with the above-mentioned government agencies; the airlines and other industry sectors to better facilitate the surveying process at the airports. This will also require time and effort from the private sector, which has proven difficult to obtain.
E. Data Entry, Editing, Verification, Weighting & Processing
Once the survey packages are returned to the contractor, they are processed. It is the contractor's responsibility to verify collection, code, edit, and input data from questionnaires. The contractor has personnel proficient in translating the responses in the different languages from the questionnaires for coding and editing purposes. Besides the data entry program, the contractor verifies the processing and integrity of the database. Finally, steps are performed to check the survey returns that have been processed. These steps include identifying the survey returns by control number and cross-tabulating returns as they are entered on the computer. Returns are crosschecked by day, airlines, and flight number to ensure the tally for all control numbers. The contractor has implemented an approved method to verify data entry and processing. The monthly data processing and review cycle includes the following tasks: Check the completed survey packages against the flight survey list and log them in; notify ASMs of any flight surveys not returned; manually edit and enter responses on questionnaire coding sheets, enter the data into the computer, perform verification of the data and wait for the I-94 or I-92 data so the quarterly data may be weighted. The entire process, excluding waiting for the control totals, normally takes two months.
The next step in the processing of the data is to expand the survey estimates by giving each respondent a weight that corresponds to the known number of international travelers entering the country, or the number of U.S. residents traveling abroad. To compute estimates of total international travel to and from the USA, OTTI uses information from the Department of Homeland Security (DHS) forms I‑92 and I‑94 flight records. The I-92 form is the only source in the world that provides information on the citizenship mix on international flights to and from the USA. The I-94 data collects information on non-residents and non-citizen inbound to the U.S. When OTTI processes the raw I-94 data, the data output is based upon the residency of the traveler. The World Tourism Organization’s recommendation is that travel to and from all countries be reported by country of residency. OTTI assumes the I-92 citizen data is equivalent to residents. The individual airline reports contain unweighted data, and therefore only include the items that do not require the expansion process.
The I-94 is the DHS Non-Immigrant Arrival Data Base. Tabulations of the I‑94 entry cards by DHS provide total arrivals into the United States by residence of other countries. They receive a monthly hard copy report (entitled Summary of International Travel to the U.S.) and data tapes from this database. OTTI ensures that the tapes/e-mail and reports are delivered to the contractor. The I‑92 is a DHS form that is also processed by OTTI. The tabulations of international air travel volume are based upon the I‑92 cards submitted by the airlines, providing numbers of U.S citizens and foreign citizen arrivals and departures between international gateways. The contractor receives a data tape/email from OTTI on a monthly basis. Only the data on U.S. citizens is used from the I-92 database. This data is published in a report entitled the U.S. International Air Travel Statistics Report. OTTI works with DHS to process the I-92 data. When OTTI reports U.S. outbound travel from the USA, OTTI assumes citizens equals residents. The contractor sends the OAG file to the Commerce Advanced Data Processing contact each month. A schedule of when the contractor expects to e-mail the OAG file is developed and sent to OTTI. This file is required by DHS so their contractor who keypunches the I-92 data can complete its work. Timely delivery of the OAG is crucial to the program.
These sources provide the input to the weighted ratio estimating procedures that expand the sample information to represent all visitors to the United States and all U.S. travelers to Mexico and overseas countries, and to ensure it is representative of world travel patterns. The use of both the survey responses and these independent data sources in the estimation process yields more reliable estimates than if only the survey responses and sampling rates were used. This is extremely critical because of the voluntary nature of this research program whereby certain carriers and world regions/countries have more respondents than their share of arrivals or departures.
The survey responses are the primary data source for computing estimates. The survey responses also provide information on distributions of variables and relationships among survey items as well as information relating the country of debarkation to the residence of the passenger.
A weight is calculated for each survey respondent. This weight represents passengers departing from the United States via scheduled international air carriers. Calculation of the weight for the U.S. residents based on the I-92 is a multi-step process. Briefly, it entails:
a. The initial
weight of a respondent is one, unless children are part of a travel party. If children are included, the initial weight
has a value greater than one, depending on the number of children and the size
of the travel party.
b. Although 100% sampling is performed on a flight, there is usually non-response. The respondents are considered a random sample of the passengers, and each weight is increased to cover the non-response on the flight.
c. Each weight of a respondent in a stratum is increased to represent all travelers on all flights on the stratum.
d. The I-92 data are incorporated into the weights by port of debarkation to represent not only the participating, but also the nonparticipating, airlines in the survey.
The weights are then used in standard weighted ratio estimation formulas for calculating the distributions, means, and medians found in the published tables.
The survey responses are also the primary data source for computing estimates for the non-resident data. Here, information developed from the Immigration and Naturalization Service Form I-94 forms the basis for the expansion process. Similar to the responses weighted using the I-92 data; the calculation of the weight is a multi-step process. Briefly it entails:
a. The initial weight of a respondent is one, unless children are part of a travel party, in which case, the initial weight has a value greater than one, depending on the number of children and the size of the travel party.
b. Both the I-94 data and survey responses are sorted and summarized by country of residence and port of customs information.
c. The weight computed for individual survey responses is the result of directly proportioning the I-94 data to the surveys.
d. The weights determined by the limiting variables in the survey responses match the corresponding control totals from the I-94 data summarized in the same manner.
The weights are then used in standard weighted ratio estimation formulas for calculating the distribution, means, and medians found in the published tables in the OTTI National Reports, as needed. The contractor is responsible for ensuring statistical reliability, validity and integrity of the survey data. The contractor is responsible for implementing this methodology, so careful attention should be given to this attachment.
Processing: IT System Requirements
The following is a description of the computer system and software that have been developed for the current program to load, edit, and process the survey data (this equipment is not required, but is provided to illustrate the bidders of the magnitude of this program and to ensure they have the computing capability to handle this project):
1. For the sampling procedure, a series of FORTRAN and DCL programs are used on a HP DS10 AlphaServer.
2. Currently, the coded questionnaires are keypunched interactively using an error and editing program written in FORTRAN. This program checks for permissible ranges, codes, skip patterns, logical consistency, etc.
3. After data quality assurance, the necessary tables are generated using a VAX/ALPHA 3600 computer using VMS 6.2. Statistical processing, cross tabulations, and weighted estimates are performed using the SAS version 6.12. To process one quarter of data, with year‑to‑date tables (i.e., third quarter, 1995 nonresidents), at least 300 to 400 megabytes of disc storage and 35 hours of CPU time are required. In contrast, in 2003, after data quality assurance, the necessary tables are generated using an HP DS10 AlphaServer computer using VMS 7.3. Statistical processing, cross tabulations, and weighted estimates are performed using the SAS version 6.12. To process one quarter of data, with year‑to‑date tables (i.e., third quarter, 1995 nonresidents), at least 300 to 400 megabytes of disc storage and 25 hours of CPU time are required.
F.
Information/Data Outputs
OTTI is able to make the information/data, processed from the In-Flight Survey program, available in a hierarchical approach: 1) Summary level profiles, published on the OTTI website which are a ‘roll-up’ of information found in 2) standardized “reports” and a 3) database that can be “mined” on an ad-hoc basis for specialized inquiries.
OTTI annually publishes the following two summary profiles on its website in html format, available to the public at no charge:
· Profile of the Overseas Travelers to the U.S.- Inbound (2003) http://www.tinet.ita.doc.gov/view/f-2003-07-001/index.html
· Profile of U.S. Travelers to Overseas Countries – Outbound (2003) http://www.tinet.ita.doc.gov/view/f-2003-101-001/index.html
OTTI also publishes a much more detailed version of the above, following a standardized format and available in both print and Excel versions, for a fee. Both national reports provide information about international visitors to the U.S. and American residents traveling abroad, respectively. These reports are produced quarterly and annually with the same format. Confidential information is only released to the appropriate parties (see below). Currently, only annual reports are made available to the public.
Click on the links underlined in this sentence to view the introduction and appendices the Non-resident 1 & 2 and Resident 1& 2 reports. Examples of an annual country and the inbound and outbound national reports are available on the OTTI website. Both the non‑U.S. resident and U.S. resident reports are offered as separate reports on travel to/from Mexico.
Each participating airline receives a ‘Confidential Airline Report’ presenting information specific only to the respective carrier with all other participating carriers grouped and their summary data presented as “all other”.
A customized report is a non‑standard request for special data profiles based upon a subset of, or a specified data relationship within the In-Flight Survey database. It varies in length and specifications depending on the needs of a user. A sample of some of the column headings selected for custom reports can be found on the OTTI website. The custom reports are generally similar in format to the published national reports and the quarterly OTTI reports, although they can take any form. Usually, the task starts with a sample size query of the contractor, and then the client selects the 12 columns they wish to use based upon the sample and their data needs. Individuals requesting a special run are charged at a rate determined by OTTI.
Currently, the main clients of custom reports are the state tourism offices and city convention and visitors bureaus. On occasion certain airlines look to expand their understanding of an origin-destination (O&D) market through use of the custom reports. It is possible to design a custom run for a hotel chain. OTTI always keeps specific competitor data confidential. The types of clients for the OTTI In-Flight Survey have been posted to the OTTI web site.
To learn more about obtaining custom reports view here.
http://www.tinet.ita.doc.gov/research/programs/ifs/customized.html
OTTI expects to expand its reach to other businesses, including other federal government agencies. The goal is to increase the number of users so that OTTI has a broader base of support for this program. Included could be: US-VISIT, Customs & Border Protection, State Department, National Park Service, Department of Transportation, Federal Aviation Administration, and others.
The Contract Structure
For information on OTTI’s contractor, CIC Research, Inc., view this link:
http://www.cicresearch.com/newhome2/AviationAndAirlines/OTTI_Info.asp
Base Program
All of the information provided above comprises the base contract, with the exception of custom reports/data files, which are described in the information that follows. Since 1990, the program sample has ranged from 59,000 surveys to more than 94,000 surveys per calendar year.
In addition to the base program, there are numerous contract options that may be purchased if funding is available and OTTI determines there is a need to implement them. Below is a summary of the contract options available. All of the items below are to assist OTTI in dealing with changes to the program or to implement options to improve it. Their implementation is solely dependent upon the availability of funds to conduct the program. Included in these options are:
1. Sample Size Increases
As a result of discussions with different government agencies, airlines, travel industry users, there is a strong interest in obtaining additional survey responses. To accommodate this, OTTI has built in a series of options that allows for the purchase of additional respondents. OTTI’s goal, as well that of the airline advisory board, when active, was to obtain a sample of 1% of total overseas and Mexican air passengers. In 2003, a one percent sample was approximately 484,000 travelers. To achieve the one percent (1%) sample goal today OTTI would have to dramatically increase the Survey of International Air Travelers respondent base, which has never exceeded 95,000 surveys. OTTI would like to systematically increase the sample size on an annual basis over several years to reach this goal. Increasing the sample size will enable OTTI to provide all interested parties with more accurate and reliable data. Tied to this increase would be the need to survey additional international airlines. The additional airlines would improve world regional representation.
2. Industry Investment Program
OTTI has been approached by a number of industry groups to enhance the sample for their state, city, airline, airport, etc. In the past, OTTI implemented this program as a pilot test with Ohio, Minnesota, and Illinois State Division of Tourism offices, as well as another attempt with the Capital Region, USA. OTTI expanded this program to increase the sample so other travel businesses/groups can obtain the sample required to provide quality estimates for visitation to their destinations. In addition, the increased sample would also allow these businesses to better understand the international traveler so they can improve the services and products offered to the visitor. A more detailed description of the investment program may be found on OTTI’s web site. This program was implemented in 2001 and 2002.
3. Custom Reports/Data Files
OTTI makes available special reports and data files for users based upon a fixed price scale for each request. Orders are placed for the custom reports and/or data via email files. These reports/files are sub-sets of the entire database as requested by individual clients. The sale of custom reports was also described in the data output section. Sales of the reports are used to generate revenue for this program to increase the sample or implement improvement options for this program.
Due to declining sample sizes for the overall program, OTTI has started to pool multiple years of the data to increase sample sizes for a particular destination, airline, airport, etc. It is also possible to pool like regions to gain an improved sample size. The pooled data is weighted to the current year, or the year requested. The benefit of the pooled data comes from larger sample sizes that tend to provide more statistically reliable estimates. There is no additional cost for combined year data. Custom reports have been purchased every year.
4. Adjustments to the Weighting Procedures
The Department of Homeland Security (DHS) I-92 and I-94 databases have had problems over the years. Adjustments to the weighting procedures have been or may be required to correct problems that occur with these two data bases. In addition, DHS is currently looking into developing an electronic collection of the international traveler data. As DHS changes the information collected, OTTI will adjust the methodologies to accommodate these changes. OTTI staff will provide guidance and support to the contractor to work through the corrections that may be needed. OTTI will approve the final corrections and defend work required to ensure accurate total population data is issued to weight the survey data collected.