Gen Z Food Insecurity

https://consortiumnews.com/2022/08/10/young-americans-who-go…
https://ag.purdue.edu/cfdas/wp-content/uploads/2022/07/Repor…
The authors’ survey shows a significant portion of Gen Z – 30 percent – has relied on free groceries from a pantry, church or other charity.

We run the Center for Food Demand Analysis and Sustainability at Purdue University, and every month, through our Consumer Food Insights survey, we query over 1,200 Americans with the goal of tracking national food security as well as many other behaviors, attitudes and preferences related to food.

Food insecurity means having a lack of money or other resources for food. And when food insecurity surges, it can take a long time for affected populations to recover. After the Great Recession that ran from 2007 to 2009, food insecurity increased by 34 percent. It took a decade for food insecurity to drop to its pre-recession levels.

1200 seems to be a small sample size. But what do I know?

1 Like

1200 seems to be a small sample size. But what do I know?

Gallup typically uses a sample size of 1,000 for a national poll. The hard part is making sure that the 1,000 people are a representative sample of the population you’re trying to model.

intercst

2 Likes

1200 seems to be a small sample size. But what do I know?

When you statistically analyze a sample, there is always some uncertainty about how accurately the sample reflects what you’re sampling. The larger the sample, the smaller that uncertainty is. The catch is, the uncertainty shrinks much more slowly than the sample - and thus the cost of collecting and analyzing the data - grows.

Assuming a “normal” distribution, a well-chosen sample of somewhere between 1000 and 1200 instances (I don’t recall the exact number) gives you an answer that is 90% likely to be within 10% of what you would get by studying the entire population. For an infinite population.

Cutting either term of the uncertainty in half - 90% chance of being within 5%, or 95% chance of being within 10% - would require more than doubling the sample size. Then another halving of uncertainty would more than double it again.

(I hope I’m remembering stuff correctly from a statistics class I took many years ago.)

What’s a well-chosen sample? Well, here’s a famous bad example. One of the early telephone polls. Back when only wealthy people had telephones in their homes, so they got only wealthy people or household servants of wealthy people. Conducted during the day on a weekday, when most wealthy men would be at work, so they got wealthy women or household servants of wealthy people - and most in-the-house household servants were also women. The poll data was clear, and the headlines reflected that clarity: DEWEY DEFEATS TRUMAN!

1200 is plenty. Less than 500, less so.

From the link:

2 Sample sizes: Gen Z (n=450), Millennials (n=1,964), Gen X (n=1,894), and Boomers+ (n=3,211).

Beyond that, such a survey is like the survey on consumer sentiment. They are largely surveying opinion and not fact.

For example, I have serious doubts that over 50% of respondents either currently grow their own food or plan to.

Anecdotally, I have a potted strawberry plant that yields maybe a dozen strawberries a year but I would not respond by saying I grow my own food. I don’t recall the last time I saw a garden in my suburban community.