Part 4: Data Recording, analysis & presentation

Raw Data

Raw data hasn't been processed/analysed in any way. (e.g. list of all of your pp's individual scores)

Most common way to present raw data is in a raw data table. Simple table with your pp's raw scores. When collecting raw data in an experiment your raw data table split in two columns, one for each condition.

Descriptive Statistics

To make sense of raw data it needs to be described/summarised in some way. Descriptive statistics.

Mathematical skills

  1. Decimals
  1. Fractions
  1. Percentages
  1. Ratios
  1. Make estimations from data collected
  1. Order of magnitude
  1. Standard Form
  1. Significant figures

Measures of Central Tendency

Mean- arithmetic average. Add all scores together and divide it by number of scores. Mean best measure of central tendency when data roughly symmetrical and no anomalies/extreme scores. Mean takes into account every piece of data, giving overall summary. Anomalies skew the mean.

Median- middle value. Median is calculated by placing all values of one condition and find midpoint. Median best measure when there are anomalies. This is because 'extreme' make mean unrepresentative. An anomaly wouldn't skew the mean.

Mode- most common value. To calculate simply find number that occurs the most in set of data.
Best measure of central tendency when dealing with categorical data. Because takes into account frequency of data, identify most common score or most frequently observed behaviour.

Measures of dispersion

Range- Subtract smallest number from largest number. Easy to calculate but affected by anomalies.

Variance- 1. Work out mean.

  1. Then for each number: subtract mean and square result.
  2. Add all scores together.
  3. Divide by number of scores -1.

Standard Deviation- Work out variance and calculate square root. More precise calculations of distribution less affected by anomalous results. More difficult to calculate than range.

Frequency tables

Used to display how often event occurred. E.g. how many times pp's displayed range of behaviours when waiting at a bus stop.

Graph

Easily create visual display of descriptive statistics. Must have title and labelled axis.

Histogram

Similar to bar chart except for continuous data. Vertical axis must start at 0 and horizontal axis must be continuous, no gaps.

Bar chart

Summary of data. Height of each bar represents frequency of each item. Gaps between bars show not continuous. Bar chart good for showing means in 2 conditions or frequency of behaviour in observation.

Line Graph

Histogram, continuous data on x-axis. Dot marked at top of bar and line connects dots. Allow 2 sets of data to be displayed.

Pie Chart

Another way represent frequency. Each piece represents proportion.

Scatter diagram

Graph display correlational data. Shows whether positive, negative or no correlation been found.

Skewed distribution curves

Normal distribution curve occurs when certain variables measured, such as IQ. Variables such as these are distributed so most of scores clustered around mean, median and mode. Measures of central tendency at mid-point. Curve is 'bell-shape'.

Skewed distribution curve where mean, median and mode not the same. Positive skew scores bunched towards left. Negative skew most scores bunched to right.