Monday, July 21, 2014

International Travel Survey (ITS) 2012

Questions

1) Looking at the CMA/CA coding, the documentation refers us to the 2001 Standard Geographical Classification. However, there are 5 codes used in the 1st quarter “co” file which aren’t in the 2001 classification (http://stds.statcan.gc.ca/sgc-cgt/2001/ss-rs1-eng.asp?criteria=35, etc.): 531/533/329/820/920. These codes do appear in the 2006 SGC. Has the documentation lagged behind the classification used, or do the codes on the ITS refer to some other areas?

2) I’ve begun working on the CO (Canadians travelling overseas) file (1st quarter), and have some questions, which I think will help with the other files in the set. First, I have a basic question about the structure of the quarterly files: I presume that when applied the weights create quarterly estimates, which can then be added together for annual estimates (and that therefore the four quarters could be combined into a single file for analysis)?

3) 
Accommodation variables - (ACCVS1, ACCVS2, ... ACCVS15)

o Each of these variables appear to record up to 6 different accommodation types, with spaces accepted in the field (“123456”, “1 3 56”) Correct

o The documentation states "The accommodation used in each location visited: Hotel, Motel, Home of friends or relatives, Camping, Cottage or cabin and Other." How is this applied?
- column 1 is either "1" or blank - is this hotel?
- column 2 is either "2" or blank - is this motel?
- column 3 is either "3" or blank - is this home of friends or relatives?
- column 4 is either "4" or blank - is this camping?
- column 5 is either "5" or blank - is this cottage or cabin?
- column 6 is either "6" or blank - is this other? 

o if a column is blank, does that mean that the accommodation type was not used, or that the accommodation type was not reported, or both? 

Carrier codes - (CARRC1, CARRC2, ... CARRC12)

o Is the coding on this to append three zeros to the national code (e.g., United Kingdom is coded 90077000)?

o Assuming that that is the normal coding, how are United States carriers specified beyond the 97433000 (e.g., 97433085, in the 1st quarter file for 2012)

- are the first 7 characters the US Destination code (e.g., in the case above, California)?

- what is the last column?

o Does a zero in this field indicate that there was no carrier, or that the carrier was not reported, or both?

Place of travel codes - (PLVSC1, PLVSC2, ..., PLVSC15)

o Is the coding on this to append two zeros to the national code (e.g., United Kingdom is coded 9007700)?

o Does a zero in this field indicate that there was no place of travel, or that the place of travel was not reported, or both? 

Various expense variables/person nights, etc. (fare1 to fare12, spvs1 through spsv15,

o Does a zero in this field indicate that there was no expense/nights, or that the field was not applicable (e.g., travel 8 didn’t occur, so expense8/nights8 is 0), or not reported, or any or all of the above?

RSCMA

o Does a 0 in this field indicate that the person did not live in a CMA/CA? 

Answers

1) Our mistake on that one. Yes, 2012 data is based on 2006 SGC. Good catch.
2) Yes, applying the weights as described in the Microdata User Guide will establish quarterly estimates, which can then be added together to create annual estimates, or yes combine all 4 files.

3) (Answers in italic red for this section)

Accommodation variables - (ACCVS1, ACCVS2, ... ACCVS15)

o Each of these variables appear to record up to 6 different accommodation types, with spaces accepted in the field (“123456”, “1 3 56”) Correct

o The documentation states "The accommodation used in each location visited: Hotel, Motel, Home of friends or relatives, Camping, Cottage or cabin and Other." How is this applied?
- column 1 is either "1" or blank - is this hotel? Yes
- column 2 is either "2" or blank - is this motel? Yes
- column 3 is either "3" or blank - is this home of friends or relatives? Yes
- column 4 is either "4" or blank - is this camping? Yes
- column 5 is either "5" or blank - is this cottage or cabin? Yes
- column 6 is either "6" or blank - is this other? Yes. Note: visits 1 to 15 can be multiple days per visit and thus any visit can have more than one type of accommodation during those days.
o if a column is blank, does that mean that the accommodation type was not used, or that the accommodation type was not reported, or both? Well, what we know is that a blank indicates that the accommodation type was not reported, which can be interpreted as not being used or as a response error (they used that type of accommodation but did not tell us!)

Carrier codes - (CARRC1, CARRC2, ... CARRC12)

o Is the coding on this to append three zeros to the national code (e.g., United Kingdom is coded 90077000)? Yes the carrier code should represent just a country code of the commercial carrier.

o Assuming that that is the normal coding, how are United States carriers specified beyond the 97433000 (e.g., 97433085, in the 1st quarter file for 2012) Seems to me like you have moved on to the variables PLVSC1 (places visited ) (7 digits posn 582-588) where for travel to US it represents the country//state codes and the 5 would be the first digit of 2 digits of the US Region Code (posn 589-590)
- are the first 7 characters the US Destination code (e.g., in the case above, California)? Yes, the appending of the US Country Code and State Code

- what is the last column? As above

o Does a zero in this field indicate that there was no carrier, or that the carrier was not reported, or both? Yes both. As above, there may not have been a commercial carrier (eg.Transport may have been by car) or the respondent indicated a plane but not a carrier name for us to code to a country.

Place of travel codes - (PLVSC1, PLVSC2, ..., PLVSC15)

o Is the coding on this to append two zeros to the national code (e.g., United Kingdom is coded 9007700)? For non-US, yes. For travel to US you will often see a state code if data was provided.

o Does a zero in this field indicate that there was no place of travel, or that the place of travel was not reported, or both? As above
Various expense variables/person nights, etc. (fare1 to fare12, spvs1 through spsv15,

o Does a zero in this field indicate that there was no expense/nights, or that the field was not applicable (e.g., travel 8 didn’t occur, so expense8/nights8 is 0), or not reported, or any or all of the above? Yes anything is possible at that point, there may not have been expenses, depending on type of transport, type of accommodation, etc.

RSCMA

o Does a 0 in this field indicate that the person did not live in a CMA/CA? Yes