Ranking Data in IBM SPSS Statistics

This article speaks about the Ranking data in IBM Statistics.

Ranking can be defined as the relationship between a set of items such as two or more data. In the ranking, it is always the first data that is ranked higher or lower. In other words, if the two items are ranked in such a way that the first is either “ranked equal to” or “ranked higher than” or “ranked lower than” to the second item.

IBM SPSS Statics ranks the cases in such a way that the cases start to automatically define the new variables to contain ranks, savage score and normal, and the percentile to the variables to the selected numeric.

If you want to rank the cases, then you need to direct to the transform and then to Rank Cases {**Transform>Rank Cases} **from the menu. The Rank Cases dialogue provides you multiple options like selecting one or more variables to rank, Rank cases in ascending or descending order, organize ranking into subgroups, selecting one or more grouping variables for the “By list”.

There are multiple ranking methods in IMB SPSS Statistics. They are

- Rank: Simple rank. It is the value of the new variable equals its rank.
- Savage score: The new variable contains Savage scores based on an exponential distribution.
- Fractional rank: The value of the new variable equals rank divided by the sum of the weights of the non-missing cases.

- Fractional rank as a per cent: Each rank is divided by the number of cases with valid values.
- Sum of case weights: the sum of the case weight is always equal to the value of the new variable. If it is in the same group, then the new variables are constant for all the cases.
- Ntiles: Ranks are always based on the percentile groups. Here each group containing approximately the same number of the cases can be found. For example, rank one will be assigned to the cases that are below 25th percentile, and rank two is assigned to the percentile between 20 to 50, and so on.
- Proportion estimates: calculation of the cumulative proportion of the distribution corresponds to a very partial ranking scale.
- Normal scores: The cumulative proportion estimates z scores.
- Total preorder: A total preorder is one which defines a ranking when two pair is found incomparable.

The Rank case Tile monitors the process of assigning a ranking to cases with the same value on the original variable. There are four different methods, and they are Mean, Low, High, Sequential ranks to unique values. They can be assigned based on the values of the variables.

…

Read More