Fichas
RDA parametric assumptions
Fichas del conjunto (8)
- 1PARAMETRIC ASSUMPTIONS: SCALE (due to importance of the mean) > Dependent variable must be scale (interval or ratio) > Interval: temperature, IQ > Ratio: height, weight, age, reaction timePARAMETRIC ASSUMPTIONS: NORMALITY
> Normal distribution: hypothetical, mathematically derived distribution where mean, median and mode are the same
- symmetrical
> check skewness and kurtosis using the 2 x SE rule
*KOLMOGOROV SMIRNOV test - 2PARAMETRIC ASSUMPTIONS: VARIANCES (so small/large variances don't cancel each other out) > instead of range/IQR, it's better to consider all scores *LEVENE TEST (part of independent t-test) - if the levene test p value (or 'sig') is greater than .05, then we can assume quality of variances- t-test (independent): scale, normal distribution, eq. variances (Mann-whitney)
- t-test (paired): scale, normal distribution (Wilcoxon)
- Pearson coefficient: scale, normal distribution (spearman coefficient) - 3SAMPLE DISTRIBUTION: distribution of raw scores in a sample calculate: mean, median, skew & kurtosisSAMPLING DISTRIBUTION: distribution of a statistic across different samples
calculate: distance of means
- st.d error of the mean
- how far sample means differ from population mean & mean of sample means - 4STANDARD ERROR: results of small samples can be generalised to large populations by the use of standard errors - how far sample means differ from population mean - how far sample means differ from mean of sample means - smaller the SE = more representative sample is of the populationPARAMETRIC ASSUMPTIONS: based on the ability to calculate the mean
- more complex than non-parametric but allows to generalise populations more confidently - 5STANDARD SCORES: processed to include info on other scores to; - know position of the score - compare different kinds of scoresTYPES OF STANDARD SCORES
1. PERCENTILES: not strictly considered S.score
2. Z-SCORES: most frequently used
3. T-SCORES: tends to be used in certain areas - 6PERCENTILES > indicate position of score (or participant) by indicating percentage of the group below that scoreZ-SCORES
> Mean = main focus
- describes how each score relates to the mean. e.g. big Z value = score far from the mean
*Z = (Score-Mean score) / SD - 7T-SCORES > similar to z-scores, but removes negatives via T = (10*z) + 50FOCUS ON THE Z-SCORE
> compares relative score magnitudes
> combines different types of variables
> determines proportion of sample with given scores - 8Z DISTRIBUTION > mean = 0 > ST = 1
Información
- Descripción
- Fichas sobre RDA parametric assumptions, creado por amelia gates el 01/05/2019.
- Creado por
- amelia gates
- Visitas
- 1
- Valoración