Add Maple
← All help articles

Numeric Factors

Numeric Factors let you combine related columns into a single score.
They are useful anywhere you want one composite metric from multiple inputs, including:

  • Rating-style questions (e.g. satisfaction or agreement items)
  • Frequency/behavior measures (e.g. usage across several actions)
  • Readiness or maturity indices (e.g. process, capability, adoption)
  • Experience or risk scores built from multiple indicators

You create a new calculated column, choose the source columns, and control how the factor is calculated.

1
Numeric factors walkthrough

Where to find it

  1. From the dashboard, click More.
  2. Click Calculated Column.
  3. Choose Numeric Factor.
  4. Click Continue.

Before you start

Numeric Factors work best when:

  • Your selected columns measure the same concept.
  • Higher values should generally mean "more" of that concept.
  • You select at least 2 source columns.

Supported source column types include opinion scale columns, numeric columns, and categorical columns that have assigned numbers.

How to create a Numeric Factor

  1. Enter a column name (for example, Combined Ratings).
  2. Under Source Columns, select all questions you want to include.
  3. In Reverse scoring, enable reverse scoring for items where higher values are actually negative.
  4. Choose Aggregate method:
    • Mean: sensitive to small differences.
    • Median: more robust to extreme responses.
  5. Set Minimum answered columns per row:
    • Require all selected: strict, most complete.
    • Require at least N columns: allows partial completion.
  6. (Optional) enable Normalize result to 0-100.
  7. Review Live reliability (Cronbach's alpha).
  8. Click Continue, check the preview, then click Create Column.

Understanding the key options

Reverse scoring

Use this when a question is phrased in the opposite direction.

Example:

  • "I find this process easy" (higher is positive)
  • "I find this process frustrating" (higher is negative)

If both are included in one factor, reverse-score the negative item so all columns point in the same direction.

Aggregate method (Mean vs Median)

  • Mean is best when you want every response shift to influence the score.
  • Median is best when you want a stable center score that is less affected by extreme values.

Minimum answered columns

This controls missing-data behavior:

  • Require all selected gives stricter, cleaner factor scores but may exclude more rows.
  • Require at least N keeps more rows by allowing partially answered item sets.

Normalize result to 0-100

Enable this when source columns have different scales (for example, one item is 1-5 and another is 0-10). AddMaple rescales each selected item before combining, so the final factor stays on a consistent 0-100 scale.

Reliability check (Cronbach's alpha)

Numeric Factor setup shows a live Cronbach's alpha estimate:

  • Higher alpha generally means the selected items move together consistently.
  • Very low alpha can indicate your selected columns are not measuring the same underlying concept.

Use this as a quick quality check before finalizing the column. In plain terms, if this score is low, your selected questions may not belong in one combined score.

Example workflows

You have 6 experience ratings:

  • Ease of use
  • Reliability
  • Speed
  • Value for money
  • Support quality
  • Overall satisfaction

You can combine all six into one factor called Combined Ratings, normalize to 0-100, and then use that new numeric column in pivots, charts, filters, and comparisons like any other numeric field.

Another common use is a product-adoption score:

  • Weekly usage frequency
  • Feature breadth used
  • Self-reported confidence
  • Task completion speed

Even when these come from different scales, you can normalize to 0-100 so each input contributes comparably.

Key points

  • Numeric Factors create one reusable numeric score from multiple related columns.
  • Reverse scoring is important for negatively worded items.
  • Missing-data rules (Require all vs Require at least) can materially change coverage.
  • Cronbach's alpha helps validate whether your selected items belong together.