DataTile Metrics

A complete list of metrics using in DataTile

Updated over a week ago

To manage metrics in DataTile navigate to Project settings → Settings → Metrics.

Project settings

Here you can:

  • switch metrics on/off;

  • adjust the decimals for each metrics separately;

  • drag the metrics within the list to change the order.

Metrics Settings

List of available metrics (in alphabetical order)

  • Aff.idx – affinity Index is a measure of the relative similarity in composition of two samples;

  • Aff.idx (Unfiltered) – same as Affinity Index, but calculated without applied filter;

  • Base% – the proportion of respondents in the cell to Total in the crosstab;

  • Cases [for rotated data only] – number of unique cases (represented by several records). A counterpart of Count metric for non-rotated bases;

  • Correl – Pearson correlation coefficient. Read more here;

  • Count – number of respondents in the cell;

  • Count MR – count for MR-set category;

  • First Label – populate a cell with the first category label;

  • Horiz% – horizontal percentage – the proportion of respondents in the cell to the Total value in the line;

  • Horiz Base – horizontal base – the number of respondents in the row excluding sysmis (missing values);

  • Horiz SOV – horizontal share of voice – is the same as Vert SOV, but is calculated as a proportion in the row;

  • Max – maximum value;

  • Mean – average. For non-numeric variables calculations are made with digits which may be changed in the Meta-Editor;

  • Mean MR – mean for MR-sets;

  • Median – value that divides the array in half by frequency. For non-numeric variables calculations are made with digits which may be changed in the Meta-Editor;

  • Min – minimum value;

  • Pen [for rotated data only] – penetration, share of cases for the column. A counterpart of Vert% metric for non-rotated bases;

  • Prob Share [for rotated data only] – probability share;

  • Sample Count – unweighted number of cases;

  • Top Label – populate a cell with the top category label;

  • Total UQ Base [for rotated data only] – all unique cases in the sample;

  • Total pen [for rotated data only] – share of cases among all unique cases;

  • Unweighted Base – vertical sampling base without weighting, coincides with the usual vertical base for unweighted queries;

  • Unweighted Horiz Base – the same as HorizonBase but without weighting;

  • UQ Base [for rotated data only] – all unique cases in the column. A counterpart of Vert Base metric for non-rotated bases;

  • Vert% – vertical percentage – the proportion of respondents in the cells among the total number in the column;

  • Vert% MR – vertical percentage for MR-sets.

  • Vert Base – vertical base – the number of respondents in the column minus sysmis (missing values);

  • Vert Base MR – vertical base for MR-sets;

  • Vert SOV – vertical share of voice – share of the Vol in the cell to Vol in the column. The analog of the vertical percentage, but for numeric values;

  • Vol – volume – the sum of the numeric values in the cell. This metric can be considered as Count - only instead of 1, as in Count, an equivalent numeric value is substituted;

  • Volume MR – volume for MR-sets.

Metrics for rotated databases

As indicated above, some metrics are available only for rotated databases. In order to make them available you need to tick the 'Enable rotating' checkbox at the end of the list.

Enable Rotating

MR-set related metrics

MR-set related metrics are specific calculations for special types of numeric MR-sets introduced in DataTile only. The basic difference between MR-set metrics and conventional metrics is that the calculations are based on the number of responses in the entire MR-set array rather than the number of respondents. Basically the meaningful calculations for multi-response questions can be made only with MR-set related metrics.

How differ MR-set related metrics

Did this answer your question?