Longitudinal Data Analysis Contest

In addition to usual communications, the conference will also feature a longitudinal data analysis contest. Participants will be asked to run their own analysis on a provided data set using methods of their choice and to submit results as a poster. Two prizes will be awarded during the conference, one by the conference attendees and one by the organizing committee.

Data and Issue

The data come from a longitudinal study of adolescents (8th graders) in the canton of Vaud, Switzerland. These adolescents were observed 5 times at 6 months interval from spring 2012 (T0) to spring 2014 (T4). The aim of the original study was to understand the use of Internet by young people. The provided dataset is a subset of 364 adolescents. The data were collected through Internet and there are many missing data. Only 185 adolescents answered to the 5 waves of the study. This sample is not directly representative of the studied population, but the use of the provided “Weights” variable improves the representativeness.

Objective

The goal of the contest is to provide an original longitudinal analysis of these data. Originality and adequation between substantive goal and methods will be rewarded! Particular attention should be paid to the treatment of missing data.

Rules

  • It is allowed to focus on a subset of the observation or a subset of the variables.
  • The results must be presented as a poster of a maximum size of 100 x 150 centimeters or less. The poster can be in landscape or portrait format.
  • An electronic version of the poster must be submitted no later than March 15th, 2016. The notification of acceptance will be sent on April 4th, 2016. Upon acceptance, it will be the responsibility of authors to print a paper version of their poster and to bring it to the conference. As for usual communications, at least one author will have to be registered to the conference for the poster to be included in the contest.

Variables (codebook provided in the dataset)

The dataset is available in both Stata 12 and R formats. In the R data file, the underscore (_) character is replaced by a dot (.). Variables measured only once at T0:

  • ID: Numerical ID of the respondents
  • T0_Age: Age in years at T0
  • T0_Gender: Gender at T0
  • T0_Parents: Parental situation at T0
  • T0_SES : Socio-Economic Status of the family at T0
  • T0_Residence: Place of residence at T0
  • T0_Track: School track at T0

Variables measured on each wave (x varies from 0 to 4 in the dataset)

  • Tx_Academic: Academic performance (not available at T3)
  • Tx_Internet: Time spent on the Internet on weekdays
  • Tx_IAT: Internet Addiction Test score
  • Tx_BMI: Body Mass Index
  • Tx_Sport: Sport practice
  • Tx_WHO5: WHO emotional well-being index (not available at T2)
  • Tx_Sleep: Sleep problems
  • Tx_Drunkenness: Drunkenness during the last 30 days

Weighting

  • Weights: Sampling weights computed at T0

Download

Submission

The submission should be done on the LaCOSA 2 easychair page by selecting the "Longitudinal data analysis contest" track : https://easychair.org/conferences/?conf=lacosa2

Deadline for submitting the poster: March 31, 2016