Descriptive stats

Describing

-Organize, summarize + describe (raw) data
- Make initial sense of data
- Present meaningfully to others

Exploratory

Examines patterns + anomalies in more detail before conducting inferential tests

Frequency distributions

Examine values by ordering from lowest to highest, grouping same values together

Frequency distributions: Histograms

-Values (e.g. No. of words recalled) are grouped into ‘bins’ (intervals into which values fall)

-Ordinal/higher + strictly speaking, continuous variables

Frequency distributions: Box Plots

-Show min + max values + middle 50% of values (interquartile range).
-Used for ordinal data or higher

-Median (50% of values above and 50% below)

Measures of central tendency

Where on frequency distribution values tend to centre or cluster together:
-Mode
-Median
-Mean

Mode

-Most frequent value

-used for all levels of data

-Only measure of central tendency that can be used with nominalh nominal data

Mode Adv

Unaffected by extreme values

Mode Disadv

-Reps most frequent value only

-Not useful when multiple modes exist

Median

-Central or middle value when arranged in numerical order

-Used with ordinal/interval/ratio data,

-if ordinal → use only median/mode

Median Adv

Unaffected by extreme values

Median disadv

Only reps middle value

Mean

Sum of values/how many values there are

Mean adv

-Reps all values, lots of info

-Used to estimate population parameters in powerful stat tests

Dispersion

-Spread/variability of values in frequency distribution

Range

-Distance’ between lowest and highest values

-Range = max value – min value

-Used on ordinal (and interval/ratio) data, often used with median

Range adv

Easy to calculate/understand

Range disadv

-Doesn’t rep values between highest/lowest

-Sensitive to extreme values

Mean absolute deviation

Deviation score:

-deviation from mean → score – mean

Absolute deviation = same but ignoring sign

Mean absolute deviation disadv

-Not used as measure of dispersion

-Not as accurate as variance + standard deviation in estimating population parameters -Need accurate estimates of these things in certain inferential tests

-Instead use variance + standard deviation as measures of dispersion

Sample variance

Mean of squared deviations

Sample standard deviation

Square root of variance

Estimate population variance

Sum of squared deviation/(total number of answers - 1)

Estimate of pop standard deviation

Square root of est. pop variance

Standard deviation

-Approximately, average amount of deviation from the mean
-Used only with interval/ratio level data