Haworth Claire M A, Harlaar Nicole, Kovas Yulia, Davis Oliver S P, Oliver Bonamy R, Hayiou-Thomas Marianna E, Frances Jane, Busfield Patricia, McMillan Andrew, Dale Philip S, Plomin Robert
Social, Genetic and Developmental Psychiatry Centre, Institute of Psychiatry, King's College London, London, United Kingdom.
Twin Res Hum Genet. 2007 Aug;10(4):554-63. doi: 10.1375/twin.10.4.554.
Quantitative and molecular genetic research requires large samples to provide adequate statistical power, but it is expensive to test large samples in person, especially when the participants are widely distributed geographically. Increasing access to inexpensive and fast Internet connections makes it possible to test large samples efficiently and economically online. Reliability and validity of Internet testing for cognitive ability have not been previously reported; these issues are especially pertinent for testing children. We developed Internet versions of reading, language, mathematics and general cognitive ability tests and investigated their reliability and validity for 10- and 12-year-old children. We tested online more than 2500 pairs of 10-year-old twins and compared their scores to similar internet-based measures administered online to a subsample of the children when they were 12 years old (> 759 pairs). Within 3 months of the online testing at 12 years, we administered standard paper and pencil versions of the reading and mathematics tests in person to 30 children (15 pairs of twins). Scores on Internet-based measures at 10 and 12 years correlated .63 on average across the two years, suggesting substantial stability and high reliability. Correlations of about .80 between Internet measures and in-person testing suggest excellent validity. In addition, the comparison of the internet-based measures to ratings from teachers based on criteria from the UK National Curriculum suggests good concurrent validity for these tests. We conclude that Internet testing can be reliable and valid for collecting cognitive test data on large samples even for children as young as 10 years.
定量和分子遗传学研究需要大样本以提供足够的统计效力，但对大量样本进行亲自测试成本高昂，尤其是当参与者在地理上分布广泛时。越来越多的人能够使用廉价且快速的互联网连接，这使得在线高效且经济地测试大样本成为可能。此前尚未有关于互联网认知能力测试的可靠性和有效性的报告；这些问题对于测试儿童尤为相关。我们开发了阅读、语言、数学和一般认知能力测试的互联网版本，并研究了它们对10岁和12岁儿童的可靠性和有效性。我们对超过2500对10岁双胞胎进行了在线测试，并将他们的分数与在这些孩子12岁时对其一个子样本进行的类似在线互联网测试的分数进行比较（超过759对）。在12岁在线测试后的3个月内，我们对30名儿童（15对双胞胎）亲自进行了阅读和数学测试的标准纸笔版本。10岁和12岁时基于互联网测试的分数在这两年间平均相关性为0.63，表明具有相当的稳定性和高可靠性。基于互联网测试与亲自测试之间约0.80的相关性表明有效性极佳。此外，将基于互联网的测试与教师根据英国国家课程标准给出的评分进行比较，表明这些测试具有良好的同时效度。我们得出结论，即使对于年仅10岁的儿童，互联网测试在收集大样本认知测试数据方面也可以是可靠且有效的。
