1016 quant and qual data analysis

what is a discrete variable

it is counted or categorised not measured, whole numbers, eg, gender, political affiliations or number of marbles in a jar

what is a continuous variable

it is measured not counted, scores lie along a continuum therefore often include decimal places, eg., distance travelled to work, height or weight

what type of data is a histogram used for

continuous variable, always have a frequency

what do histograms offer

examine shape of distribution, show min/max score and the mode, estimate location of mean and median and assess skew and kurtosis

what is the difference between between a histogram and a bar chart

histogram is used for a continuous variable and bar charts are used for discrete variables

what about data visualisation can be misleading

a modified or edited scale that makes the difference between scores seem smaller or larger than it is

what do bar charts offer

comparison of values across different subgroups (categories within a discrete variable)

what type of data is displayed on a line graph

continuous - most appropriate for showing growth, progression or decline

what is the difference between a bar chart and a line graph

bar chart compares values associated with separate categories and a line graph emphasises change in a dependent variable due to an independent variable

what sort of data is displayed on a pie chart

discrete/categorical data - used to display relative frequencies - depicts how large sub categories are relative to each other

what info is shown on a box and whisker plot

distribution of a variable - median, lower and upper quartile and extreme values or outliers - used to combat skew of outliers

what is skew

when the distribution is not symmetrical and has one long tail

positive vs negative skew

positive skew - when ‘tail’ points towards larger values on x axis - can be seen in mean test scores if test was too difficult negative skew - most values clustered around top end of distribution - seen if test was too easy

what is kurtosis

when the distribution has a flat or pointy peak

what is meant by a platykurtic distribution

wide spread of scores and a very shallow peak - occurs when similar numbers of people achieve every possible score

what is meant by a leptokurtic distribution

a narrow spread of scores and a very pointy peak - occurs when most people achieve a narrow range of possible scores

what is meant by a mesokurtic distribution

a normal distribution

where can skew be seen on a box and whisker plot

skew can be seen by the position of Q2 - nearer the bottom of the box is positive skew and nearer the top of the box is negative skew

non parametric tests

chi-squared, mann-whitney U - wilcoxon signed ranks

what does it mean if a data set is parametric

it means it has met the assumptions of that statistical test eg must be normally distributed

what are the 3 key measurement properties of variables

magnitude, equal intervals and true absolute zero

what is magnitude

one score can be said to be higher or lower than another score - scores can be logically ordered or ranked eg., age or weight

what is equal intervals

the distance between measurement points on the scale is always the same, no matter where they are on the scale eg., distance or length - centimetres

what is true absolute zero

means that the zero point on the measurement scale is the point where nothing of the variable exists eg weight - no weight less than 0 grams

what are the 4 levels of measurement

1= nominal - none of the measurement properties, 2= ordinal - has only magnitude, 3= interval - has magnitude and equal intervals, 4= ratio - has magnitude, equal intervals and absolute zero

what is nominal data

discrete categorical data with no natural order eg. nationality or occupation

what is ordinal data

all properties of nominal data except the categories have a natural order - difference between each point is not consistent eg., feedback (best to worst) height (tallest to shortest)

what is interval data

top level of measurement - precise due to having same distance between each adjacent point - continuous eg., temperature in celsius

what is ratio data

interval data but with an absolute zero eg., time in seconds weight in grams