Nappy Washing Survey – Provisional Results


  • 3 washing/drying practices were associated with fewer reported smells (p<0.05).
    • Higher temperature washes,
    • hot tumble drying,
    • more nappies in wash
  • Hot tumble drying seems to help our respondents avoid rough nappies (p<0.05).
  • Eco washes were associated with rough nappies. (p<0.05).

A huge thank you to Emily Griffiths for designing the questionnaire and doing the analysis.

Thank you to everyone who completed our Nappy Washing Survey.
Below are some provisional results.

Emily also created this fabulous infographic summarising the results:

Infographic v3

How the data was analysed:

I [Emily] split responses into “reported smells” vs “no reported smells”.

I coded every question in the survey about washing/drying practices as simply as possible
 (time in bucket, detergent type, detergent amount, temperature of wash, extra rinses, number of nappies in wash, other items in wash, age of machine, use of ecocycles, machine cleaning, drying method, and drying time).

For each washing/drying practice in turn I did a logistic regression against reported smells. I then combined significant practices into a multiple repression against reported smells

Results Set 1 – Smelly nappies

In our live chat with Elaine and Greg, they told us that they use HA to concentrate bacteria (i.e. the bacteria love to stick to HA in high numbers). Therefore, we believe one sign that may indicate the presence of HA is smelly nappies.

Result 1: Higher temperature washes was associated with fewer reported smells (p<0.05)
(Bar plot of number of responses, light grey = smells, dark grey = no smells)

Temperature graph

Result 2: Hot tumble drying was associated with fewer reported smells (p<0.05
(Bar plot of number of responses, light grey = smells, dark grey = no smells)

 Hot tumble graph

Further questions:

  • Does hot tumble drying reduce the likelihood of HA?
  • Could heat and friction physically break up the HA coting on the fibres, releasing it from the fabric? Is this partial removal what causes the crusty texture?
  • If HA comes off in the tumble dryer, then the fluff that collects in the filters should contain HA? Would it be there after one load or would it build up over time?


Result 3: More nappies in wash was associated with fewer reported smells (p<0.05
(Bar plot of number of responses, light grey = smells, dark grey = no smells)

Number of nappies graph

When these variables were combined in a multiple regression, all the above associations had the same direction and had marginal (p<0.1) or significant (p<0.05) associations.

Results set 2 – Rougher nappies

We believe that HA could be making our nappies feel hard or crusty.
The original HA containing nappies that were sent to Shirley Technologies were described as feeling ‘pebbly’.

Result 4: Crinklier / rougher nappies were associated with not hot tumbling (p<0.05)
i.e. Hot tumbling seems to help avoid that rough crinkly feeling
(Bar plot of number of responses, light grey = rougher, dark grey = not rougher)

rough hot tumble

Result 5: Crinklier / rougher nappies were associated with using an ecowash (p<0.05
(Bar plot of number of responses, light grey = rougher, dark grey = not rougher)

What does ecowash mean?

It may depend how old your machine is. The ‘economy’ mode on older machines was designed to save money. The Eco[logical] mode on newer machines is a setting that is meant to be better for the environment. These settings are not the same.

What actually happens when you press these buttons also seems to vary from one machine to another. Here is a summary from the website 10:10:

“On most washing machines the eco mode is either an extra setting you can apply to any of the standard cycles, or a special cycle in its own right.

In both cases, you’re altering the way the heating element behaves – either running a longer cycle at a lower temperature, or heating the water to the same temperature but much more slowly. It’s not obvious that running something at lower power for longer would save energy, but in this case it seems to work.

The only exception to this is Samsung’s ‘ecobubble’ setting, which froths up the detergent so it can spread through your clothes more easily. Samsung claim that this vastly reduces the amount of heat you need to get things clean, giving you the results of a 40° wash at just 15°. Reviewers seem to be divided on how well it works, but as a new approach to energy saving, it looks pretty promising.

Worth using?

Maybe, but don’t expect a huge saving. It’s also worth bearing in mind that the eco setting isn’t necessarily the most energy efficient cycle you can get from your machine.

Washing at 60° in eco mode will use much more energy than if you’d picked a standard 30° cycle, so make sure you’re using the right temperature for the load in question before you worry about the eco mode.”

Example Ecowash

The manual for a hotpoint washing machine tells us that its eco-wash setting uses cold water, makes no difference to the spin speed, but the recommended maximum load is vastly reduced from 7kg to just 3kg.

It would be interesting to know how other manuals describe their ‘eco’ mode.

Result 6: Use of bio / non bio did not seem to be linked to smells or roughness

bio vs non bio graph

Next steps?

Emily did a few more stats on the three proxy measures for HA in nappies:

  • Smelliness and crustiness are positively correlated (r=0.2). Both are also positively correlated with reported change in absorbency (r just about 0.09).
  • I’ve run a multivariate regression for these three response variables, but the only predictor coming up as significant is use of ecocycles and it only accounts for about 1% of the variance.

Further stats?

  • Further statistical analysis or data visualization of reported smells?
  • Look at other HA-linked nappy characteristics in addition to smelliness and texture (i.e. reported change in absorbency or nappy material),
  • Delve deeper into the detailed responses within each practice (e.g. combinations of drying methods), and do more fancy stats.
  • Multivariate models? I’m thinking now is the time to move from simplified binary variables to factors that reflect the range of survey responses…



Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s