James Frank Quotes

We saw that our customers required help beyond the data sets they had and that they could benefit from a wider opinion. So we built SurveyMonkey Audience, and we've now got 4 million users who signed up to take surveys. Our clients can choose the demographic they want to hear from, and we can provide that sample.

I really like the mission at SurveyMonkey, which is, we help people make better decisions. It's just a great thing.

Our polling methodology has gotten outdated, and, in fact, it's not really telling us what it needs to be telling us.

In order to create predictive algorithms, you need to have a training set. So, that training set is created through our quantitative surveys. Those surveys need to include either basic market research questions or basic political polling questions, which might be added to get your opinion on a brand or an issue or a candidate.

I am completely opposed to the concept of big data.

If you aren't taking a representative sample, you won't get a representative snapshot.

Certainly, last year we did an episode about the census and sampling versus a direct statistic. You just said the word 'census,' and people fall asleep.

The information gained from these techniques was valuable in some instances, but there is no way of knowing whether the same information could have been obtained through other means.

A great deal of creativity is about pattern recognition, and what you need to discern patterns is tons of data. Your mind collects that data by taking note of random details and anomalies easily seen every day: quirks and changes that, eventually, add up to insights.

At the beginning of all experimental work stands the choice of the appropriate technique of investigation.

I think the biggest problem in clinical trials is that they are underpowered. And that fundamentally, the studies are just too small.

All I do is look, listen and try to make sense of what I find, in biological terms.

It is important to consider whether the sample size selected by the Environmental Genome Project will provide sufficient power to discover most alleles relevant to gene-environment interactions.

If information is gathered, history shows that it will be used at some time.

Surveys give us a way of finding out what people across the country are thinking.

What is not a risk with a parts-per-billion test can always be exposed to a parts-per-trillion examination. If rats cope with .the heaviest dose of a chemical that can be soaked into their food and water, you can always gavage them. Or try mice or rabbits. Again, the only stopping rule is discovery of the sought-for effect, or exhaustion of the investigator (or his funds). Many of the risk assessment procedures used today are logically indistinguishable from those used by the Inquisition.

A great many decisions just can’t be made with confidence because the data just doesn’t exist …. [however] just because we don’t have the information doesn’t mean that we can’t get it.

To do successful research, you don't need to know everything, you just need to know one thing that isn't known.

Collect it all!

… information that has come in has not commenced an investigation …. I never made a lead from rhetoric that came from a Demographics report, and I'm here since 2006

Collect It All Process It All Exoloit It All Partner It All Snff It All Know It All

I'm a research queen.

I usually just try to do whatever's on the page because I've done research before - including a lot of analysis - but you end up with conflicting data. To me, the script is king.

Crazy old people are our entire source of polling information.

We sent our user research team out to sit with customers for periods of time and get some insight into how Slack is working. We also get tens of thousands of points of contact via Twitter and our customer support ticketing system every month and can synthesize those results.