Incidence Rate in Market Research: Formula and Screening Maths

• Incidence rate (IR) is the share of people who finish a screener and qualify. People to screen = sessions needed ÷ show rate ÷ IR ÷ (1 − screener drop-off).
• Each halving of the incidence rate doubles the number of screener starts you need.
• NFHS-5 (National Family Health Survey, fieldwork 2019 to 21) found 8% of Indian households own a car: 14% in urban India and 4% in rural India. A city and a village are different recruitment jobs.
• A planned IR is an estimate. A pilot with 100 finished screeners narrows it only to about plus or minus 7 percentage points.

Incidence rate in market research is the number of qualified respondents divided by the number of respondents screened, qualified and non-qualified together, usually shown as a percentage (Wikipedia: incidence rate (market research)).

IR = qualified ÷ screened × 100

A screener is the short set of eligibility questions people answer before they are offered a place in a study. We count "screened" as people who finish the screener. We count people who abandon it separately, as drop-off, so the two losses never get mixed up. This is our working convention.

Example: 150 people finish a screener and 21 qualify. IR = 21 ÷ 150 = 14%.

Screener starts = completes ÷ show rate ÷ incidence rate ÷ (1 − drop-off). For 12 sessions at an 80% show rate, 14% incidence rate and 6% drop-off, you need about 114 screener starts.

If some qualifiers cannot make your session slots, fold that share into the show rate.

The same 12 sessions, 80% show rate and 6% drop-off at other incidence rates (each figure rounded up from the unrounded chain):

Read the table as a ratio, not a verdict. Halving the IR doubles the starts. Going from 14% to 4% multiplies them by about 3.5 (399 ÷ 114). Whether 114 or 399 is acceptable depends on your channel, timeline and budget. For quota group sizes across age and gender groups, use our quota sampling calculator.

For ownership screeners, national household surveys give a usable starting point. NFHS-5 (fieldwork 2019 to 21) found that 8% of Indian households own a car, 14% in urban India and 4% in rural India. It also found that 54% own a scooter or motorcycle. The survey covered 664,972 households (ThePrint, 27 May 2022).

The later Household Consumption Expenditure Survey (HCES) 2023-24 points the same way: more than 60% of households own a motorcycle or scooter, and fewer than 10% own a car or jeep (Data For India).

Applied to our funnel, a "household owns a car" screener in an urban setting starts near 14% and needs about 114 starts for 12 sessions. The same screener in rural India starts near 4% and needs about 399. A two-wheeler screener sits above half the population and barely moves the numbers.

These are shares of households, not of adults who answer a recruitment screener, so they are a starting estimate and not a forecast. If you recruit through online channels, your pool will lean urban, so the urban figure is the better first guess. Add every other condition you want, such as "bought in the last 12 months", and the rate falls further.

Every extra condition in a screener removes people, so the real rate is usually lower than a rate planned on the headline criterion alone. If two conditions were unrelated, their shares would multiply. For example (hypothetical numbers), 14% who own a car and 50% who also meet a second condition would leave 7%. In practice conditions overlap, so the product is a rough guide, not an answer.

The cost of being wrong is large. At 7% instead of 14%, the same 12 sessions need 15 ÷ 0.07 = 214.3 finishers, and 214.3 ÷ 0.94 = about 228 starts. That is double the 114 you planned for.

Our recommendation: write the target profile as the fewest conditions that still define the person, and move "nice to have" conditions out of the screener into the discussion guide. Our research screener template shows how to separate the two.

Run a small batch of screeners first, then recalculate from what actually happened. Four steps:

1. Send the screener until at least 100 people have finished it (about 107 starts at 6% drop-off).
2. Compute actual IR = qualified ÷ finishers. Compute drop-off = abandoners ÷ starters.
3. Put both into the formula in place of the defaults.
4. Allow for noise. With 100 finishers and a 14% reading, the 95% margin is 1.96 × √(0.14 × 0.86 ÷ 100) = about 6.8 percentage points. The true rate could be anywhere from about 7% to 21%. That is 228 starts at the low end and 76 at the high end for our 12-session example.

A pilot gives you a range, not a point. Pass: the high-cost end of the range still fits your budget and timeline. Fail: it does not, so relax a condition, widen the geography or increase the number of cities before fieldwork starts. A larger pilot narrows the range. At 400 finishers the margin is about 3.4 points. This margin is our calculation using the standard normal approximation.

When your incidence rate is low, we check feasibility against your brief, recruit by city and target by household and demographic profile, then send a tailored proposal and quote. A free pilot lets you confirm the real rate before you commit.

Your brief should cover the research objective, method, target audience, participant profile, geography, sample size, timelines and recruitment criteria. We quote each project individually. A good place to start is our one-page market research brief template.

Once the project is commissioned, we check each participant's identity before inviting them. Sessions can run in languages other than English, and we can supply moderators.

To start, email your brief to support@qualfacto.io. Our hours are Monday to Friday, 10:00 AM to 6:00 PM Indian Standard Time, and we aim to reply within one business day.