Reporting Method - Cobrand
Reports by Gender
Description of Reporting Method:
There are different ways reports can be made. FixMyStreet.com represents reports made by FixMyStreet.com, Mobile reports are made on an app, and Cobrand are made through cobranded websites.
Description of Reports by Gender:
Gender is derived from name. Inconclusive names are dropped, so this represents less than 100% of the dataset.
The chart below shows the distribution of reports in Reporting Method - Cobrand for Reports by Gender.
This analysis has used a chi-square test to examine the differences between this distribution and how it would be expected to be distributed if behaved identically to the general dataset (all reports).
The test statistic is 31,244.62 and the p-value is <0.00001.
As each graph only explores the interaction of two variables analysis may be incomplete or misleading where other variables interact with both. A relationship can be statistically significant without necessarily implying one causes the other.
Columns in green and red have at least a 95% chance of a real difference from the global dataset and at least a 1% different from the expected value.
Green means a category has more reports than expected. Red means it has less.
Comparison with General Dataset
In the next chart, the transparent bar preceding each entry represents the expected values if counts were distributed in the same pattern as the general dataset.
This helps to demonstrate the scale of any deviation from the general dataset. When the amount of data available is large, a difference can be statistically significant while being relatively small.
Difference by Percent
This chart shows the percentage difference between the expected and actual result - representing the scale of difference between this category and the general dataset.
This chart shows the absolute difference between the expected and actual result - representing how many more or less reports were received in this category compared to what was expected.