Reliability & validity issues in European web survey in Czech Republic

Slides:



Advertisements
Similar presentations
Item Writing Techniques KNR 279. TYPES OF QUESTIONS Closed ended  Checking yes/no, multiple choice, etc.  Puts answers in categories  Easy to score.
Advertisements

Heavy or problematic drug use, heavy and problematic to define? Iceland August 2010, Nordic Alcohol and Drug Researchers’ Assembly Ellen J. Amundsen,
Departments of Medicine and Biostatistics
Extended Project Research Skills 1 st Feb Aims of this session  Developing a clear focus of what you are trying to achieve in your Extended Project.
You Talkin’ to Me? An Overview of Disputed Terminology in AOD Prevention William DeJong, PhD Director U.S. Department of Education’s Higher Education Center.
ALEC 604: Writing for Professional Publication Week 7: Methodology.
Internet-based interventions — messages that click Dr. Peter Tossmann delphi-Gesellschaft für Forschung, Beratung und Projektentwicklung mbH, Berlin EMCDDA.
FINAL REPORT: OUTLINE & OVERVIEW OF SURVEY ERRORS
1 Emergency Infant Feeding Surveys Assessing infant feeding as a component of emergency nutrition surveys: Feasibility studies from Algeria, Bangladesh.
Cognitive Interviewing for Question Evaluation Kristen Miller, Ph.D. National Center for Health Statistics
Polydrug use challenges – European experience International Conference: New trends in drug use: facts and solutions, Parliament of the Republic of Vilnius.
Tutor: Prof. A. Taleb-Bendiab Contact: Telephone: +44 (0) CMPDLLM002 Research Methods Lecture 8: Quantitative.
Annual report 2010: the state of the drugs problem in Europe.
Trends and developments Selected trends and some insights in prevention Vilnius, 12 November 2014 Roland Simon.
Health Datasets in Spatial Analyses: The General Overview Lukáš MAREK Department of Geoinformatics, Faculty.
Alcohol Consumption Lydia Gisle Hélène Mimilidis Presentation in the framework of the EHIS workshop, Berlin, September 30, 2010 Improvement of the European.
HOW TO WRITE RESEARCH PROPOSAL BY DR. NIK MAHERAN NIK MUHAMMAD.
Chapter Fourteen Data Preparation 14-1 Copyright © 2010 Pearson Education, Inc.
MGT-491 QUANTITATIVE ANALYSIS AND RESEARCH FOR MANAGEMENT OSMAN BIN SAIF Session 16.
Joint OECD / European Commission workshop on international development in business and consumer tendency surveys Nov Task force on improvement.
Normative misperceptions about alcohol use in the general population of drinkers Claire Garnett 1, David Crane 1, Robert West 2, Susan Michie 1, Jamie.
CCT 333: Imagining the Audience in a Wired World Class 8: Quantitative Methods.
Just before we get started… Who are we? How questions will be handled Resources available after the webinar Key QILT Dates & the AGS Item Review 2.
The effects of Peer Pressure, Living Standards and Gender on Underage Drinking Psychologist- Kanari zukoshi.
Experimental Research Methods in Language Learning Chapter 12 Reliability and Reliability Analysis.
Mode effects in social surveys A mixed-mode experiment linked to the Safety Monitor Ger Linden, Leanne Houben, Barry Schouten (Statistics Netherlands)
Research Methodology Lecture No :32 (Revision Chapters 8,9,10,11,SPSS)
Overview of the cannabis use in Europe Paul Griffiths, Reitox academy, Berlin, 29th March 2007.
Chapter Fourteen Data Preparation 14-1 Copyright © 2010 Pearson Education, Inc.
The impact of using the web in a mixed mode follow-up of a longitudinal birth cohort study: Evidence from the National Child Development Study Presenter:
VARIABILITY IN DRINKING PATTERNS AMONG COLLEGE-AGED YOUTH Frances K. Del Boca Alcohol & Substance Use Research Institute University of South Florida.
Introduction to Survey Research
The Basics of Social Science Research Methods
Miami, Florida, 9 – 12 November 2016
3 Research Design.
The epidemiology of IPED users attending IEP services in Scotland
ESPAD Report 2015 Results from the European School Survey Project on Alcohol and Other Drugs ESPAD Group Lisbon, 20 September 2016.
Peter Linde, Interviewservice Statistics Denmark
Experience of Serbia in conducting pilot Time Use survey
Measuring Substance Use
Introduction to Marketing Research
Is the freedom from Cognitive Impairment really at hand?
Melanie Taylor Horizon Research, Inc.
Questions or comments:
Definitions of risky and problematic cannabis use: a systematic review
Facebook ads as recruitment for online drug surveys: the Holy Grail?
AXIS critical Appraisal of cross sectional Studies
Martínez-Loredo, V. 1, De La Torre-Luque, A. 2, Grande- Gosende, A
Health consequences of substance use in nightlife settings in the Netherlands Dr. Karin Monshouwer, senior researcher Department of Epidemiology.
Writing Survey Questions
Lecture 6 Structured Interviews and Instrument Design Part II:
H676 Meta-Analysis Brian Flay WEEK 1 Fall 2016 Thursdays 4-6:50
To use or not to use? An exploration of cannabis use motives and constraints Dr Liz Temple
Chapter 1 Getting Started Understandable Statistics Ninth Edition
Comparison of data from WBE with other sources
15% of Australians used an illicit drug (including using a pharmaceutical drug for non-medical purposes) in the previous 12 months.
2UNIVERSITY OF WATERLOO
Item 5.3 Feasibility studies
Chapter 12: Surveys Introduction 12.1 The method 12.2 Random samples
Elena Gomes de Matos, Tessa-Virginia Hannemann,
CLIENT RELATIONSHIP MANAGEMENT KEEPING TRACK OF REQUESTS THE EASY WAY
2nd meeting of the task force on survey based disability statistics
Secondary Data Analysis Lec 10
LAMAS October 2017 Agenda Item 5.1 Consultation to stakeholders
Input Presentation for Working Group 4
The Vermont Young Adult Survey: Methods and Some Initial Findings
Bob Flewelling Pacific Institute for Research and Evaluation
2.7 Annex 3 – Quality reports
RES 500 Academic Writing and Research Skills
Skopje, 21 November 2017, General Population Survey results– launch
Presentation transcript:

Reliability & validity issues in European web survey in Czech Republic Lisbon Addictions 2017

Conflict of interest statement. Work funded under the EMCDDA project CT.15.EPI.0112.1.0 (Assessment of the Methodology of internet web-surveys to collect amounts of drugs used: reliability and validity).

Aims - to trial online data collection and various solicitation channels - to assess test/re-test reliability of the individual items and of the derived consumption estimates - to improve questionnaire through the qualitative cognitive assessment

Conclusions: Methodological challenges It is crucial to know the target population. Use simple tools to measure complex and rather irregular phenomena.

Content and timeline Literature review (Autumn 2015) Translation, adaptation and testing of the common questionnaire (Winter 15 / Spring 16) Pilot data collection (Feb-April 16) *Follow-up data collection (re-test, April-May 16) Two-weeks daily recalls (diary method, May 16) *Cognitive interviews (May-July 16)

Overview of the data Pilot (N=610) - 65% male, Mage=28 (80% aged 18 to 34 years) - prevailing stable working and living conditions, 38% tertiary education - LYP: 88% cannabis, 29% cocaine, 42% XTC, 25% amphetamines Follow-up (N=153) - 61% male, Mage=28 - LYP: 93% cannabis, 33% cocaine, 55% XTC, 30% amphetamines Cognitive interviews (N=19) - recreational drug users, problem methamphetamine users, one dealer of methamphetamine Ads targeted – recreational drug users / party goers, cannabis users / cannabis legalization supporters, psychonauts, and problem users of met/amphetamines. (79% came through Facebook)

Test/re-test reliability (individual items) Moderate to high reliability of most items (w/sufficient N) Prevalence items single-question vs. arrays Traditional vs. unusual/unfamiliar drugs Frequency questions Ordinal scale vs. days of use (LMP/LYP) High reliability of clue-induced amounts (use, purchase) Only calculated for cannabis resin Low reliability for amounts used in “dry pipe” Low reliability for XTC/MDMA questions (powder/tabs) Low sample sizes for cocaine questions, no data on amphetamine tablets Follow-up data collection 2 weeks after pilot, recruited via email address request Same questionnaire (copy/paste in the Lime Survey) Reduced background and redundant questions   N= 153, RR=66.2% (out of 231 emails collected) kappa statistics (𝜅) for categorical and dichotomous variables interclass correlation coefficient (ICC) for ordinal and continuous measures modest (0.30 to 0.49), moderate (0.50 to 0.69), or high (0.70 to 1.00) reliability

Test/re-test reliability (typologies of users) Cannabis users: Other substances: Infrequent users or chippers = less than 11 days in the past year (≈ “less than once a month”). Infrequent users = less than 11 days in the past year (≈ “less than once a month”) Occasional users = 11-50 days (≈ “less than once a week but at least once a month”). Occasional users = 11-50 days (≈ “less than once a week but at least once a month”) Regular users = 51-250 days (≈ “once a week and ‘more than once a week, but not daily or almost daily”). Frequent users = 51 to more than 350days (≈ “once a week or more”). Intensive users = more than 250 days (≈ daily or almost daily”).

Test/re-test reliability (yearly consumption) Last 30 days not representative for last year. Number of assumptions and data manipulations for weed/skunk and resin. Consumption by type of user calculated. Last 30 days not representative for last year Number of assumptions and data manipulations for weed/skunk and resin Consumption by type of user calculated

Validity (cognitive interviews) Overall comprehensible, easy to read and understand Classical cognitive issues: Less effort in online data collection (“did not know precision is required”) Telescoping (e.g. accounting for distant heavy periods) Identification w/pattern (recreational vs. heavy vs. cannabis users) Exceptional experiences easier to recall (events/holidays, cocaine and resin) How many doses (joints/pipes) would participant prepare from one gram of cannabis or cannabis resin? low reliability for cannabis, moderate for resin (n.s.) No “usual” or “typical” dose/purchase/money spent for many users (“It depends… on weekday, available drug, quality, budget...”). Occasional extreme purchase and/or use. Daily cannabis users: grams/day vs. joints/day Irregular patterns in frequency questions (missing “once or twice”) Resin in frequent cannabis users Heavy methamphetamine users/binges Holiday season binges   Types of cannabis – buy vs. consume Amphetamine vs. methamphetamine (double-barrelled section) Local production of meth (precursors, exchange for goods) Specific terminology: ”one gram” = 0.7 gram = 10 syringe units Unclear amount of purchase: One dose/one line ”for one thousand”

Summary of the main points Moderate to high test/re-test reliability Unambiguity and effort increase reliability Frequency questions perform better when number of days is asked (compared to ordinal scale item) Individual consumption easier from direct questions (=less error) Recommendations from cognitive testing - never assume regular patterns and same understanding of items.

Thank you for your attention. Katerina Skarupova skarupovakat@gmail.com