Data Analysis Flashcards
(17 cards)
Nominal Data (definition, example, central tendency measure, measure of dispersion)
Data is qualitative, no ranking or order.
Example - Hair colour, region
Central Tendency - Mode
Measure of Dispersion - NA
Ordinal Data (definition, example, central tendency measure, measure of dispersion)
Data that has a sequence but irregular gaps between levels
Example - Ages of people, body mass
Central tendency - Median
Measure of Dispersion - Inter-quartile range
Interval Data (definition, example, central tendency measure, measure of dispersion)
Steps in the scale are evenly placed but zero does not mean zero.
Example - Temperature
Central tendency - Mean
Measure of Dispersion - Standard deviation
Ratio Data (definition, example, central tendency measure, measure of dispersion)
Steps in the scale are evenly placed and zero means zero.
Example - Length, time
Central tendency - mean (if no outliers)
Measure of dispersion - Standard deviation
Standard Deviation (Variance)
The average amount all scores deviate from the mean. Calculated as the root of variance(s).
E.g. A SD of 0.32 means a greater dispersion/variation of data points than an SD of 0.12.
Standard Error of the Mean (SEM)
Uses the SD to find the variation from the mean. A large SEM means more variability from the mean.
E.g. A greater SEM indicated more variation between data points.
Confident Intervals
A CI is 2 x SD from the mean. It shows where 95% of the data falls within that range.
E.g. If two CI’s do not overlap, they are likely to be statistically different.
P-Values
Refers to the probability that any difference is due to chance. It indicated if there is a statistically significant difference between two conditions.
E.g. IF P < (or equal to) 0.05 THEN you can accept that a difference in statistically significant. IF P > 0.05 THEN the conditions are not statistically significant.
R-values
Correlation coefficients (r) describe the relationship between two variable.
E.g. r = +1 = positive correlation
r = 0 = no correlation
r = -1 = negative correlation
What type of data should be used for parametric tests?
Normally distributed, large sample size , interval/ratio data
What type of data should be used for non-parametric tests?
nominal/ordinal data, unclear distribution (skewed/outliers), small sample size (<20), less statistical power.
Which correlation test should be used if it’s interval/ratio data with a large sample size?
Pearson’s Correlation
Which correlation test should be used if it’s nominal/ordinal data with a small sample size?
Spearman correlation
Which test should be used for independent groups with normal distributed, interval/ratio data?
Unpaired T-test
Which test should be used for independent groups with a small sample size and outliers?
Mann-Whitney U Test
Which test should be used for repeated measures or matched participants with normally distributed interval data?
Paired t-test
Which test should be used for repeated measures or matched participants with a small sample that has unclear distribution?
Wilcoxon signed-ranks test
