ACADEMY · FREE MODULE 7 OF 11

M7 · Basic Statistics for Improvers

7 min video with captions · one-page job aid · short quiz

Job aid

Let the data speak

Numbers summarize; charts show the shape. Start with the question, then the type of data, then pick the chart.

Number How What it tells you
Mean Add the values, divide by how many there are The balance point. One outlier pulls it
Median Sort the values, take the middle one (or the average of the middle two) The typical value. Outliers barely move it
Range Largest − smallest The full spread, set by the two extremes
SD √( Σ (value − mean)² ÷ (n − 1) ) The typical distance from the mean

In a spreadsheet, use the built-in average, median, max − min and sample standard deviation functions.

Build the charts

  • Histogram: split the scale into equal bins → count the values in each → draw a bar per bin.
  • Pareto chart: count by category → sort largest first → add the running total → find the few bars that make up most of it.
  • Run chart: plot in time order → join the points → add a line at the median.

Read a control chart: a centre line, plus limits at about 3 standard deviations, calculated from a stable period (individuals chart: mean ± 2.66 × average moving range). Signals: a point outside a limit · 8 in a row on one side of the centre · 6 in a row going up or down.

  • No signal = common cause: improve the system, don't chase single points.
  • A signal = special cause: find what was different that time.
  • Control limits show what the process does. The tolerance (specification) says what the customer needs. They are not the same thing.
Question Data Chart
Which categories matter most? Categories (counts) Pareto
How are the values spread? Measurements Histogram
Is it changing or stable over time? Measurements in time order Run or control chart

Watch out for: the mean when there's an outlier · tolerance drawn as control limits · reacting to every point · a Pareto of measurements (it needs categories) · an outlier left inside the limits it widens.

Examples

Industry Data Four numbers Chart and reading
✈ Airline MRO A-check TAT, 10 checks (h) Mean 11.6 · median 11.2 · range 4.4 · SD 1.28 Control chart (limits from 9 typical checks): check 6 at 14.9 h is special cause, a hangar door failure. Record it; don't redesign the check
⚙ Manufacturing Bore Ø, 8 shafts (mm) Mean = median 25.005 · range 0.05 · SD 0.016 No special cause, but limits 24.93–25.08 are wider than the 24.95–25.05 tolerance: change the system (tool life). Pareto: top two scrap reasons = 70%
🏦 Banking/Services Team AHT, 12 days (min) Mean 6.41 · median 6.15 · range 3.6 · SD 0.96 Histogram: 11 days in 5.5–7.0 min. Day 6 at 9.4 min is above the 8.75 UCL: a system outage. Record it; plan a fallback
✚ Healthcare Weekly median clinic wait, 8 weeks (min) Median of the weeks 32.5 Week 5 at 52 min (IT outage) is above the 39.5 UCL set from the 7 typical weeks (supplemental example)
📦 Logistics Orders packed per hour, 10 shifts Mean = median 53.5 · range 9 Values spread evenly around the centre (supplemental example)

All examples are fictional.

Quiz

Answer each question first; the answer and why open after you choose or press “Show the answer”. Nothing is saved or sent.

Pass mark: 5 of 6 questions, and Q4 must reach 4 of 5. Q6 counts as correct at 5 of 6 checklist points. Calculated answers are accepted within ±5% of the key.

Q1 · Airline MRO · Mean or median?

Scenario: A fictional hangar logged the turnaround time (TAT) of 10 A-checks, in hours: 10.5, 11.0, 12.5, 10.8, 11.6, 14.9, 11.1, 10.9, 11.4, 11.3.

Question: Which number best describes a typical check?

Q2 · Manufacturing · Reading a Pareto chart

Scenario: A fictional machining cell scrapped 100 parts last month: oversize bore 48, burrs 22, surface scratches 14, wrong length 9, other 7.

Question: What share do the top two reasons make up, and where should the team start?

Q3 · Banking/Services · Pick the chart

Scenario: A fictional call center has 12 days of its team's average handle time (AHT, the average minutes spent on each call). The manager asks: "Is daily handle time stable, or did something change?"

Question: Which chart answers that, and why not a Pareto chart?

Q4 · Control-chart exercise (mixed industries)

Task: For each chart, say whether it shows only common cause or a special cause, and why.

  1. Airline MRO: 10 A-check TATs; limits from the 9 typical checks: centre 11.2 h, UCL 13.2 h, LCL 9.3 h. Check 6 is 14.9 h; the rest are 10.5–12.5 h.
  2. Manufacturing: 8 bore diameters, 24.98–25.03 mm; limits 24.93–25.08 mm.
  3. Banking/Services: 12 days of team AHT; limits 4.06–8.75 min. Day 6 is 9.4 min; the rest are 5.8–6.5 min.
  4. Logistics: orders packed per hour over 10 shifts, 49–58; limits 40.5–66.5. The points go up and down with no run.
  5. Generic process: 20 points, all inside the limits, but points 11 to 19 are all above the centre line.

Q5 · Healthcare · Transfer question

Scenario: A fictional outpatient clinic tracks the median wait each week (the middle wait of all patients that week), in minutes: 32, 35, 30, 34, 52, 33, 31, 29. In week 5 the booking system was down for two days (an IT outage). Control limits set from the seven typical weeks are 24.5 to 39.5 minutes.

Question: What's the best response to week 5?

Q6 · Build exercise (learner's choice)

Task: Use data from your own work, or this fictional table from a service desk:

  • Resolution time (hours), 12 tickets in the order they were closed: 4.2, 3.8, 5.1, 4.6, 3.9, 4.4, 6.8, 4.1, 4.5, 3.7, 4.9, 4.3
  • Reasons tickets were reopened last month: missing information 14, waiting for approval 9, system error 5, other 2

Build a histogram and a run chart of the resolution times, and a Pareto chart of the reopen reasons. Write one line saying what each chart tells you.

All examples are fictional.