Are you planning on a total bacterial count? If so, have you decided on the type of agar you are planning to use? Can you not use sterile swabs and then plate it on the agar?

Hi Jason,

This reminds me of an experiment that we performed when I was in college: special agar plates are available that allow you to press the agar against a surface that you'd like to investigate. They are somewhat smaller than a petri dish and really convenient for the use you're describing. I think they're also referred to as 'steri-keri' plates.

With regard to statistics: I'm not sure what method is best, I can only advice you to sample a lot of different toilets to enable sensible statistical analysis..

Good luck!

You must carefully think how far you wish to generalize your findings. The generalizability will depend on what "bathrooms" you will sample and at what times you will sample. Are these private and/or public bathrooms? What are the "clients" using these bathrooms (at the time of sampling)? Whener are the bathrooms (location; referring to cultural backgrounds, general habits, local distributions of pathogens, etc etc)?

After you have decided what bathrooms you will sample and when you will sample, you should think if it may be reasonable to expect systematic differences between particular subgroups (e.g. i'd expect public bathrooms be generally more dirty than private ones). You may use a startified sampling approach to get enough data from all subgroups you consider relevant.

Concerning the data, there are two possible aspects that come into my mind: (i) the number of colonies you will find and (ii) the number of different bacterial species you will find. It may be sensible to analyse (i) startified by the species.

The analysis of count values called for a Poisson regression model. A negative binomial model may be more flexible and fit the data better. In case you have a huge amount of data with high count numbers (several tens or hundreds) per agar plate and species, you may go for a "simple" linear model (based on a normal error distribution).

If you sample the same locations several times, you will have a repeated-sample design you can analyse with a linear mixed model (either with Poisson, negative binomial or normal error structure).

Thank you for your answer, the toilets will be public toilets within the university department. For the statistical part, I'm asked to use one of the following approaches ONLY:

The nature of variability

Populations, samples and sampling strategies

Types of data

Probability

Standard deviation

Normal distribution

Sampling from the normal distribution, standard error

Students t distribution, t tests

Analysis of variance

One way or Two Way ANOVA

Multifactorial analyses, blocking

Correlation and regression

Chi-square analysis, contingency tables

Non-parametric tests

Questionnaire technique and structure

Which one should I go for? Thanks for the help

Jason, I don't get your question.

Would you regard "Then go for 'Standard Deviation'" being a sensible answer?

As I noted before, *you* must decide is the distribution of bacterial species is of interest or not. And *you* must decide which covariables are of interest. And *you* must decide if you will sample the same bathrooms several times.

Depending on *your* choice, you may do a stratified analysis or a non-stratified analysis, and for the analysis a linear (possibly mixed) model with negative binomial error structure will supposedly be the best choice.

Hi,

Honestly, the choice of your statistical test will depend on the nature of your variables and your hypothesis. We know sex (men vs. women) would be a categorical variable but how do you aim to measure the bacterial colonies? Is it going to be a numerical variable ( bacterial count) or categorical (types of bacteria)?

Regards,

Gayatri

There is another problem I am facing, it's about a fair experiment. How can I manage to build up a fair experiment when for example men's toilet could be dirtier due to MORE men actually using it and women using their toilets less. Any idea how I could resolve this issue?

Clean the surfes before they are touched by a person. Wait for the next person to leave their germs, then take the sample.