LEQ

Analyzing LEQ

Effect Sizes & the LEQ

James Neill
Last updated:
30 Jan 2006

Notes about LEQ & Effect Sizes

  • Effect size (Cohen's d) is basically (mean2 - mean1 / standard deviation1), where:
    • pre-test is 1

    • post-test is 2

  • The ideal standard deviation (sd) is from the population (not the sample).

  • Therefore, it is generally recommended that you use the standard deviations from the cumulative LEQ database (Neill, Marsh & Richards, 2003), based on approximately 3000 Australia participants aged 13 to 65 years.  The larger or more different your sample is, the more important it may be to use standard deviation estimates derived from your own sample:

    • Time Management (1.25)
    • Social Competence (1.16)
    • Achievement Motivation (.98)
    • Intellectual Flexibility (.97)
    • Task Leadership (1.24)
    • Emotional Control (1.24)
    • Active Initiative (1.18)
    • Self Confidence (1.12)
  • Sample doesn't really impact on calculation of effect sizes (e.g., you can calculate effect sizes for small samples), but note:

    • the smaller the sample, the less reliable the results

    • particularly for small samples, reporting a confidence interval around the estimated effect size is important (e.g., try a 90% confidence interval)

FAQ

Q: I would like to know few things about LEQ evaluation process. For computing effect sizes (ES) of the LEQ, is it necessary to count standard deviation (SD)? And, if so, are the SDs the constant numbers posted on wilderdom web site or can someone can use it calculating it from SPSS analysis?

A: For the Effect Sizes (ES), yes, you need to use a standard deviation (SD). Ideally, the SD used should be the best possible estimate of the SD in the population to which you wish to generalise the results. This means you could either rely on the existing SDs based on ~3000 mostly Australian people aged ~12 to ~60 years (average in the 20's) or use SDs based on existing ADRA data. If you have sufficient LEQ data from your own sample (~say 100+), it probably makes most sense to calculate your own SD.  You can calculate SDs by getting descriptive statistics in SPSS. ES calculation does take some extra commands, though e.g. the SPSS syntax on the website.