Psychometrical Properties of the Russian 15FQ+

Reliability of the 15FQ+

Reliability is the property of measurement which assesses the extent to which variation in measurement is due to true differences between people in the trait being measured or to measurement error. The reliability coefficient shows whether the questionnaire is able to produce stable, reliable results.

Test-retest reliability (stability) shows how much the respondent’s results change depending on his mood, well-being, fatigue and other similar factors. To compute test-retest reliability, a sample of respondents completes the questionnaire twice (usually, the second testing is organized 2-8 weeks after the first one). The test-retest reliability coefficient is then computed by determining the correlation between the first and second test results. For personality questionnaires that measure stable traits (such as the 15FQ+), the test-retest reliability coefficient should exceed 0.6-0.7. The table shows that 15FQ+ meets these requirements.

15FQ+ Scale Test-retest Reliability fA 0.79 β 0.88 fC 0.82 fE 0.82 fF 0.85 fG 0.88 fH 0.86 fI 0.86 fL 0.84 fM 0.87 fN 0.79 fO 0.77 fQ1 0.85 fQ2 0.86 fQ3 0.84 fQ4 0.89 SD 0.74

Table 1. Tests-retest reliability of the 15FQ+scales

Internal consistency of a questionnaire scale shows the ability of questions on a scale to measure the same trait. Chronbach's Alpha was used to compute the internal consistency of Russian 15FQ+ scales. Table 2 shows that these coefficients are quite high, which means that questions on each scale are consistent among themselves. Moreover, internal consistency coefficients do not exceed the critical level of 0.9-0.95, which means that 15FQ+ factors measure broad source traits, rather than narrow (highly homogeneous) surface traits.

15FQ+ Scales Alpha Coefficient fA 0.74 β 0.79 fC 0.80 fE 0.61 fF 0.74 fG 0.79 fH 0.79 fI 0.72 fL 0.79 fM 0.65 fN 0.77 fO 0.64 fQ1 0.80 fQ2 0.76 fQ3 0.66 fQ4 0.83 SD 0.69

Table 2. Internal Consistency of the Russian 15FQ+ scales (computed using Chronbach's Alpha)

Standard error of measurement (σ_m) indicates how accurately questionnaire scales measure personality traits. The laws of statistics tell us that for 68% of respondents their “true” result lies within ±1 σ_m of their actual result and for 95% it lies within ±2 σ_m of their actual results. The standard error of measurement for the Russian 15FQ+ scale is approximately 1 Sten, which is standard for multiple-factor personality questionnaires.

15FQ+ Scales σm (in Stens) fA 1.02 β 0.92 fC 0.89 fE 1.25 fF 1.02 fG 0.92 fH 0.92 fI 1.06 fL 0.92 fM 1.18 fN 0.96 fO 1.20 fQ1 0.89 fQ2 0.98 fQ3 1.17 fQ4 0.82 SD 1.11

Table 3. Standard errors of measurement of the Russian 15FQ+ scales.

Validity of the 15FQ+

Validity is the ability of a scale score to reflect what that scale is intended to measure. If we choose shooting as an analogy, then reliability can be compared to a series of shots, while validity can be compared to overall accuracy of fire.

Construct validity assesses whether the characteristic that a test is actually measuring is psychologically meaningful and consistent with the test's definition. The construct validity of a test is assessed by demonstrating that the scores from the test are consistent with those from other major tests that measure similar constructs and are dissimilar to scores on tests that measure different constructs. The original 15FQ and the 16PF, which both have the same scale structure as the 15FQ+, can be used to demonstrate the latter's construct validity.

As the validity of a questionnaire is limited by its reliability, the correlation coefficient for the results of two questionnaires will be lower then correlation coefficient for “true” values of personality traits. That's why a special correlation coefficient, which is corrected for attenuation due to measurement error, is used in computing construct validity.

Table 4 shows correlations between 15FQ+ and the original 15FQ. Except for the scale в, which has a different meaning in the 15FQ+ and the 15FQ (“Intellectance” and “Intellect”, respectively), all other 15FQ+ scales correlate significantly with 15FQ.

15FQ+ Scales 15FQ Scales fA 0.43 β - fC 0.75 fE 0.93 fF 1 fG 0.97 fH 1 fI 0.98 fL 1 fM 0.84 fN 0.77 fO 0.95 fQ1 1 fQ2 1 fQ3 1 fQ4 0.4

Table 4. Correlations between the 15FQ+ factors and the original 15FQ.

Table 5 shows correlations between the 15FQ+ scales and various versions of Cattell's 16PF. As the 15FQ+ is a business analogue of 16PF, these correlations should be fairly high. In fact, table 5 data enables us to confirm construct validity of the 15FQ+.

15FQ+ Scale 16PF 16PF5 fA 0.37 0.70 β - - fC 1 1 fE 0.99 1 fF 0.98 1 fG 0.89 0.75 fH 0.99 1 fI 0.75 0.56 fL 0.52 0.79 fM 0.65 1 fN 0.70 0.31 fO 0.99 1 fQ1 0.43 0.84 fQ2 0.85 1 fQ3 0.50 1 fQ4 0.94 0.89

Table 5. Correlations of the 15FQ+ factors with 16PF (Form A) and 16PF5

Tables 6 and 7 shows correlation coefficients between “Big 5” scales in the 15FQ+ and the 16PF. All correlations are highly statistically significant.

PF4 EX PF4 AX PF4 TM PF4 IN PF4 SC Extraversion 0.76 -0.29 -0.01 0.41 -0.03 Anxiety -0.22 0.84 -0.04 0.08 -0.17 Openness 0.27 0.1 0.48 0.25 -0.02 Agreeableness -0.28 0.14 0.16 0.71 -0.05 Control -0.05 0.14 0.09 -0.12 0.59

Table 6. Correlation between 15FQ+ and 16PF4 global factors

PF5 EX PF5 AX PF5 TM PF5 IN PF5 SC Extraversion 0.88 -0.27 -0.12 0.45 -0.29 Anxiety -0.22 0.87 -0.04 -0.05 -0.03 Openness 0.11 0.14 0.65 0.29 -0.29 Agreeableness -0.03 0.08 0.29 0.81 0.19 Control -0.08 0.13 0.43 -0.21 0.79

Table 7. Correlation between 15FQ+ and 16PF5 global factors