chi square would be suitable as it could show for example the difference in high order/low order shops with distance from the CBD. As these shops would be counted then the chi square could work because it uses frequencies ? ...

chi square would be suitable as it could show for example the difference in high order/low order shops with distance from the CBD. As these shops would be counted then the chi square could work because it uses frequencies ? ...

It is the expected frequency that has to be greater than 5. Advantages for using Chi Square should have been that it can be used on data that can't be ordered numerically, it is easy to calculate and can be used with more than one variable. Disadvantages would be that it can be unreliable with small frequencies and data in the form of percentages can't be used. More advantages for Spearmans rank would be that it is simple and quick to calculate, data can be ranked in either ascending or descending order so less likely to make a mistake, and can be used with a scatter graph. Disadvantages would be that one small mistake would completely throw the result and can be difficult to spot and it just shows that a relationship exists but not exactly what it is.

## 6 Comments:

Reasons for the use of the different stats tests we suggested. EW

So, why, for instance, would chi square be a suitable statistical tool?

chi square would be suitable as it could show for example the difference in high order/low order shops with distance from the CBD. As these shops would be counted then the chi square could work because it uses frequencies ? ...

chi square would be suitable as it could show for example the difference in high order/low order shops with distance from the CBD. As these shops would be counted then the chi square could work because it uses frequencies ? ...

It is the expected frequency that has to be greater than 5. Advantages for using Chi Square should have been that it can be used on data that can't be ordered numerically, it is easy to calculate and can be used with more than one variable.

Disadvantages would be that it can be unreliable with small frequencies and data in the form of percentages can't be used.

More advantages for Spearmans rank would be that it is simple and quick to calculate, data can be ranked in either ascending or descending order so less likely to make a mistake, and can be used with a scatter graph. Disadvantages would be that one small mistake would completely throw the result and can be difficult to spot and it just shows that a relationship exists but not exactly what it is.

The last comment was RW

Post a Comment

<< Home