data management seminar, 8-11th july 2008, hamburg 1 summary common sources of error
TRANSCRIPT
Data Management Seminar, 8-11th July 2008, Hamburg1
SummaryCommon Sources of Error
Data Management Seminar, 8-11th July 2008, Hamburg2
InstrumentPreparation
Sampling and Preparing
Survey Administration
Data Entryand Verification
SURVEY
Codebook Adaptation
Rece
ivin
g M
ate
rial
Data
Sen
dou
t
Scoring
Data Management Seminar, 8-11th July 2008, Hamburg3
InstrumentPreparation
Sampling and Preparing
Survey Administration
Data Entryand Verification
SURVEY
Codebook Adaptation
Rece
ivin
g M
ate
rial
Data
Sen
dou
t
Scoring
Data Management Seminar, 8-11th July 2008, Hamburg4
Errors During Instrument Preparation
Errors in translations (especially: careful if cooperate with other countries)The same terms in different questions translated differentlyComments from translation verification not implementedNational adaptations not or incorrectly documented on NAFDeletion of international questions and categories without strong reason and approvalInternational question stem changed
Data Management Seminar, 8-11th July 2008, Hamburg5
Errors During Instrument Preparation (ODC)
Conversion in SurveySystem started before paper instruments are reviewed start AFTER international verificationCopy and paste to wrong text, numeral, or instruction check paper version against previewPaper questionnaires and ODC questionnaires do not match in terms of disabled questions, variables or additional elements verify within country before submitting to IEA DPC
Data Management Seminar, 8-11th July 2008, Hamburg6
InstrumentPreparation
Sampling and Preparing
Survey Administration
Data Entryand Verification
SURVEY
Codebook Adaptation
Rece
ivin
g M
ate
rial
Data
Sen
dou
t
Scoring
Data Management Seminar, 8-11th July 2008, Hamburg7
Errors During (Within-school) Sampling
School sampling frame incorrect / major parts of the population not includedSample not approved by the IEA DPCReplacement of ineligible schools (frame errors like closed schools)Missing selection of participation in Regional Module or on-line data collection when creating new project in WinW3SSchools did not list all eligible students and teachers or left out eligible but excluded them follow-up
Data Management Seminar, 8-11th July 2008, Hamburg8
InstrumentPreparation
Sampling and Preparing
Survey Administration
Data Entryand Verification
SURVEY
Codebook Adaptation
Rece
ivin
g M
ate
rial
Data
Sen
dou
t
Scoring
Data Management Seminar, 8-11th July 2008, Hamburg9
Errors During Codebook Adaptation
Length of a national variable specified to small (one digit for a variable with 6 or more options)Wrong labeling/numbering of categories or dimensions when they were deleted or addedDeletion of variables for questions that are not administered set to „hide“ in WinDEMCodebooks for different data entry computers are differentIf ODC is used, codebooks for WinDEM do not match with ODC data (changes made after export from SurveySystem)
Data Management Seminar, 8-11th July 2008, Hamburg10
InstrumentPreparation
Sampling and Preparing
Survey Administration
Data Entryand Verification
SURVEY
Codebook Adaptation
Rece
ivin
g M
ate
rial
Data
Sen
dou
t
Scoring
Data Management Seminar, 8-11th July 2008, Hamburg11
Errors During Administration
Instruments distributed to wrong person (only to designated individuals)Sampled student/teacher replaced do not substitute!Participation status not tracked (Tracking Forms) or monitored (On-line Monitor Report)Spare Booklet assignment not documented (Tracking Forms)
Data Management Seminar, 8-11th July 2008, Hamburg12
InstrumentPreparation
Sampling and Preparing
Survey Administration
Data Entryand Verification
SURVEY
Codebook Adaptation
Rece
ivin
g M
ate
rial
Data
Sen
dou
t
Scoring
Data Management Seminar, 8-11th July 2008, Hamburg13
Errors During Material Receipt
Receipt of materials not tracked appropriatelyIncomplete materials contact school coordinator and try to resolveIncomplete materials because of low on-line participation consider following up to schools with paper questionnairesIf ODC used, “Monitor” reports not taken into account continuously
Data Management Seminar, 8-11th July 2008, Hamburg14
InstrumentPreparation
Sampling and Preparing
Survey Administration
Data Entryand Verification
SURVEY
Codebook Adaptation
Rece
ivin
g M
ate
rial
Data
Sen
dou
t
Scoring
Data Management Seminar, 8-11th July 2008, Hamburg15
Errors during Scoring
Single items are all scored by the same scorerReliability booklets are marked by the same scorer in the instrument and in the reliability Scoring sheetMain scorers enter codes directly in reliability booklets before reliability Scoring has been done
Data Management Seminar, 8-11th July 2008, Hamburg16
InstrumentPreparation
Sampling and Preparing
Survey Administration
Data Entryand Verification
SURVEY
Codebook Adaptation
Rece
ivin
g M
ate
rial
Data
Sen
dou
t
Scoring
Data Management Seminar, 8-11th July 2008, Hamburg17
Errors During Data Entry in WinW3S
If ODC was used, questionnaire type not tracked correctly in WinW3SParticipation status not entered in WinW3SSpare booklet assignment not performed in WinW3SITLANG in NAF and WinW3S not set if more than one language was used
Data Management Seminar, 8-11th July 2008, Hamburg18
Errors During Data Entry in WinDEM
Not all materials enteredWrong or duplicate IDs, column shifts, invalid values(All avoided by using WinDEM) train staff thoroughlyWrong treatment of out-of-range values enter data “as is”!Missing values used incorrectly (use ‘9’ for any omitted response, ‘7’ for any invalid response and valid values for all interpretable responses)
Data Management Seminar, 8-11th July 2008, Hamburg19
Errors During Data Entry in WinDEM
Data entered on different computers not combined or combined multiple times (leading to duplicate IDs)Values for Check-All-That-Apply questions used incorrectly (use 1 for CHECKED, 2 for NOT CHECKED)Filters interpreted enter data “as is”!
Data Management Seminar, 8-11th July 2008, Hamburg20
Errors during Verification
WinDEM checks not performedChecks performed but data not correctedManipulating the data files with MS Excel (data file will be cut of after 255 variables)
Data Management Seminar, 8-11th July 2008, Hamburg21
InstrumentPreparation
Sampling and Preparing
Survey Administration
Data Entryand Verification
SURVEY
Codebook Adaptation
Rece
ivin
g M
ate
rial
Data
Sen
dou
t
Scoring
Data Management Seminar, 8-11th July 2008, Hamburg22
Errors during Data Send-out
Data submitted, but without documentation (Tracking Forms, NAF)NAF or instruments not sent for all languages usedCompleted Tracking Forms are printed from WinW3S send paper copies or scans of original forms WinW3S file not exportedDouble-punching data missingDocumentation incomplete or does not match dataAfter data is submitted, national centers are not prepared to answer questions from the IEA DPC Although data processing is exhausting, it’s not time for vacations, yet ☺
Data Management Seminar, 8-11th July 2008, Hamburg23
SummarySummaryData Processing at the IEA DPC Data Processing at the IEA DPC
Data Management Seminar, 8-11th July 2008, Hamburg24
ContentContent
• Data and Documentation Submission• Data Processing at the IEA DPC
– Adapting the File Structure– Cleaning Steps
• Further Schedule• Contact Addresses
Data Management Seminar, 8-11th July 2008, Hamburg25
In general 3 month after last day of administrationLatest possible dates– Southern Hemisphere: March 27, 2009– Northern Hemisphere: August 28, 2009
Material to be Submitted to IEA Material to be Submitted to IEA DPCDPC
Data Management Seminar, 8-11th July 2008, Hamburg26
Set of Final Instruments
National Adaptation Forms(NAF) version IV
Exported WinW3S database
National Codebooks
Verified Data sets
Reliability Scoring Data
Data Submission
Documentation Data
Tracking Forms Double Punching Data
Data Management Seminar, 8-11th July 2008, Hamburg27
Please use the WebFTP sites to exchange materials between countries and the - ISC at ACER and/or- IEA Secretariat and/or - IEA DPC Please inform the recipients via email as soon you have uploaded any materialsFurther information on the IEA DPC WebFTP-Server has been provided in a separate WebFTP-Server Manual
File-Exchange using the IEA File-Exchange using the IEA DPC WebFTP-ServerDPC WebFTP-Server
Data Management Seminar, 8-11th July 2008, Hamburg28
ContentContent
• Data and Documentation Submission• Data Processing at the IEA DPC
– Adapting the File Structure– Cleaning Steps
• Further Schedule• Contact Addresses
Data Management Seminar, 8-11th July 2008, Hamburg29
Data Processing at the IEA DPCData Processing at the IEA DPC
Data & Codebook
sInstrument
s & Forms
INP
UT
IEA DPC
Country
1. Import to SAS and adapting the
file structure
Data Management Seminar, 8-11th July 2008, Hamburg30
Adapting the File StructureAdapting the File Structure
WinW3S database and WinDEM data files are merged and converted to SAS data filesUnique system numbers (CSYSTEM, TSYSTEM, SSYSTEM) are added– System numbers usually correspond to the
position of the record in the original WinDEM data files
– If a data set is missing either in WinW3S or in WinDEM, higher system numbers will be assigned
Data Management Seminar, 8-11th July 2008, Hamburg31
Structure of the Merged Data Structure of the Merged Data SetSet
Variable ValueIDBOOK 2
IDPUNCH 111
IDSCORA 101
...,ITPART1..2,IDSCHOOL,IDCLASS,...
3...1001
100101
IDSTUD 10010101
IDCHECK 270
TOKEN01 3
CB2PDO1 1
SSYSTEM,... 000001
Addedduring
cleaningprocess
Example: Student Achievement Data File
Information:
TrackingForms
Information:Student
AchievementBooklets
WinDEM(ISAxxxC2.DBF)
Data Management Seminar, 8-11th July 2008, Hamburg32
Data Processing at the IEA Data Processing at the IEA DPCDPC
Country Databas
eNew Data
Structure
Reports, Statistic
s & Docu.
INP
UT
OU
TP
UTStructur
e Check
Background
Cleaning
IEA DPC
Country
ID Cleaning
Linkage Cleanin
g
Data & Codebook
sInstrument
s & Forms2. Cleaning
Country communication will continue after data send-out!
Data Management Seminar, 8-11th July 2008, Hamburg33
ContentContent
• Data and Documentation Submission• Data Processing at the IEA DPC
– Adapting the File Structure– Cleaning Steps
• Further Schedule • Contact Addresses
Data Management Seminar, 8-11th July 2008, Hamburg34
Further ScheduleFurther Schedule
Software: – July 21, 2008: WinW3S– August 18, 2008: SurveySystem– August 18, 2008: WinDEM
Documentation in Survey Operations Procedures Manuals:– July 21, 2008: SOP (MS), Unit 2 (WinW3S I)– August 18, 2008: SOP (MS), Unit 3 (WinW3S II,
WinDEM, SurveySystem)
Material Submission for Main Survey: March 27, 2009 (SH); August 28, 2009
(NH)
Data Management Seminar, 8-11th July 2008, Hamburg35
ContentContent
• Data and Documentation Submission• Data Processing at the IEA DPC
– Adapting the File Structure– Cleaning Steps
• Further Schedule • Contact Addresses
Data Management Seminar, 8-11th July 2008, Hamburg36
Contact AddressesContact Addresses
tel: +61 3 9255 5555
e-mail:[email protected]
Web: http://iccs.acer.edu.au
tel: +31 20 625 3625 e-mail: [email protected]
InternationalStudy Center (ISC)
IEA Secretariat
Data Management Seminar, 8-11th July 2008, Hamburg37
Contact AddressesContact Addresses
tel: +49 40 48 500 750
e-mail: [email protected]
IEA DPC Sampling
IEA DPC tel: +49 40 48 500 611
e-mail: [email protected]: https://webftp.iea-
dpc.org
Data Management Seminar, 8-11th July 2008, Hamburg38
Thank you very much for your Thank you very much for your attention!attention!