You are here:
- Home »
- Using evidence for learning »
- Working with data »
- Effect size
Effect size
A good way of presenting differences between groups or changes over time in test scores or other measures is by ‘effect sizes’, which allow us to compare things happening in different classes, schools or subjects regardless of how they are measured.
An effect size in educational measurement is the difference in assessment outcomes measured in standard deviation or standard error units. Essentially, calculating an effect size corrects for differences in the spread of data.
2007 | Actual beginning of Year Results | Actual End of Year Results | Mean Diff | Confidence Interval for Effect Size | Effect Size |
|||||
---|---|---|---|---|---|---|---|---|---|---|
Mean | N | SD | Mean | N | SD | Lower | Upper | |||
Year 9 Reading | 505 | 81 | 56 | 602 | 68 | 67 | 97.0 | 1.21 | 1.95 | 1.59 |
Year 10 Reading | 554 | 83 | 59 | 616 | 86 | 93 | 62.0 | 0.48 | 1.10 | 1.06 |
Year 9 Maths | 507.7 | 82 | 65 | 613 | 67 | 106 | 105.3 | 0.87 | 1.58 | 1.23 |
Year 10 Maths | 609.8 | 76 | 90 | 713 | 83 | 110 | 103.2 | 0.69 | 1.35 | 1.02 |
The mean is a measure of a central tendency. It is simply the average of all the values in a sample. To compute the mean, all the values are added and then divided by the number of values.
The standard deviation (SD) is a measure of the spread of the data. A smaller standard deviation would indicate the majority of results are close to that of the mean. A larger standard deviation would indicate that the results are more spread out, both below and above that of the mean.
The mean difference is the difference between the end of year results and the beginning of year results for each class.
The confidence interval for effect size is a measure of the significance of the effect size taking into account the spread of the data and also the number of observations. A confidence interval that doesn’t include zero indicates that there is a significant difference between students expected end of year score and students actual end of year score.
Effect size is a way of quantifying the difference between two groups. In this instance we are quantifying the difference between the end of year and beginning of year results.
An effect size of 1.0 indicates that a particular approach to teaching or technique advanced the learning of the students in the study by one standard deviation above the mean, typically associated with advancing children's achievement by one year, improving the rate of learning by 50%, or a correlation between some variable (for example, amount of homework) and achievement of approximately .50. When implementing a new programme, an effect size of 1.0 would mean that approximately 95% of average students receiving that treatment would exceed 84% of students not receiving that treatment (Hattie, 1999).
According to Hattie (1999) normative comparison points of effect sizes are:
- 0.1 Student maturation
- 0.24 Any teacher in front of class
- 0.4 Innovations in schooling
Anything above 0.4 would imply that the innovation is working better than expected.
The Professional Learning and Development BES suggests 0.2 – 0.4 is a small but educationally significant impact, 0.4 – 0.6 is a medium educationally significant impact and greater than 0.6 is large.