Hoyle (1995) who recommends a sample size of at least 100 to 200
Kline (2005: 15) who considers less than 100 to be ‘untenable’ other than for
very simple models
Schumaker and Lomax (2004:49) who surveyed the literature and found
sample sizes of 250 to 500 to be used in "many articles" and "numerous
studies ..that were in agreement" that fewer than 100 or 150 subjects was
below the minimum
Hair et al. (2006: 741) who write that minimum sample sizes for maximum
likelihood estimation are 100 to 150 and that 200 provides a sound basis for
estimation
Hair et al. (2006: 742) add that where factor loadings are in the range of 0.67
to 0.74 then the sample size should be ‘more in the order of 200’
Other authors base their guidelines not on absolute sample size, but on sample
size relative to the number of observed variables. Examples of such guidelines
include:
Kline (2005: 178 ) who suggests at least 10 cases per observed variable
Stevens (Stevens 1996) who suggests at least 15 cases per observed variable
Garson (2011b) who suggests that sample size should be at least 50 more than
8 times the number of observed variables in the model
Based on the guidelines for absolute sample size described above, this sample, at
greater than 200 cases, is adequate for SEM analysis; in relation to the guidelines
for the ratio of cases to observed variables, based on the minimum guideline of
10 cases per observed variable, 213 cases allows for the estimation of a model
containing 21 observed variables.
A further model-based assessment of sample adequacy is available when using
the AMOS software for SEM analysis. Hoelter’s Critical N (CN) (Hoelter 1983)
specifically concerns the adequacy of sample size for calculating the χ^2 value for
that model. Hoelter’s recommendation was that where CN is ≥200 then the
sample size is large enough to for χ^2 to be calculated satisfactorily. Garson
(2011b) notes that a CN of less than 75 is considered insufficient to accept a
model by chi-square. Hoelter’s CN is reported for each SEM model in Chapter 7
detailing the analysis of the data.
A further check on the adequacy of the sample was undertaken by comparing the
current sample size with reported sample sizes for other leadership-focused