The National Context Effect: An Empirical Test of the Validity of Cross-National Research Using Unrepresentative Samples  

Oleh:

Straus, Murray A.

Jenis: Article from Journal - ilmiah internasional
Dalam koleksi: Cross-Cultural Research vol. 43 no. 03 (Aug. 2009), page 183-205.
Topik: measurement validity; cross-national research; physical assault; crime; dating violence
Fulltext: 183.pdf (227.92KB)

Isi artikel

The objective of this study was to investigate whether results from crossnational studies using convenience samples which are not representative of the nation can provide valid cross-national comparisons. Analysis of data from the International Dating Violence Study (IDVS) of university students in 32 nations (n = 17,404) enabled 18 tests of concurrent validity and found an average correlation of .51 between variables measured by the IDVS measured and by nationally representative samples. Construct validity was also supported by 41 empirical tests. The concept of national context effects explains how samples that are not nationally representative can provide valid nation-to-nation differences. It was concluded that convenience samples that are not representative of the nation but are comparable across nations can provide valid tests of theories about differences between nations. Consequently, if a study can only be done using convenience samples, that should not deter proceeding.

Opini Anda

Klik untuk menuliskan opini Anda tentang koleksi ini!

Kembali