Inference of Slope

Inference of Slope

When answering these questions, either copy and paste your JMP output into your response, or ‘Save script in data table’.

1.  Storks Deliver Babies

Don’t believe what you’re told in biology class. Storks deliver babies. To analyze this we have data from 17 countries. The variables we have are:

·  Country

·  Area (km²)

·  Pairs of storks

·  Human population (millions)

·  Birth rate per Year (thousands)

Dataset: http://castatistics.wikispaces.com/file/view/storks+deliver+babies.jmp

To find out if storks deliver babies, it would be reasonable to see if you can establish a relationship between pairs of storks and birth rate per year. Therefore, your task is:

a.  Run a test of slope using Pairs of storks to predict Birth Rate per Year.

(make sure you run through the entire ‘test of slope’ process.)

b.  Interpret the slope and intercept

c.  Using your model, predict the Birth Rate per Year for 1000 Pairs of Storks.

d.  Are you happy with your bivariate model in part a.?

Source paper: Storks deliver Babies, Matthews, R., Teaching Statistics Vol. 22, No. 2, 2000

2.  Income and Life Expectancy

Does your income predict how long you will live? The following dataset is of the fifty States and DC. The variables given are:

·  State

·  Percentage of Smokers who attempted to quit

·  Percentage of Smokers in the State

·  Life Expectancy (years)

·  Estimate Lung and Bronchus Deaths

·  State Population in 2007

·  Median Income 2007 ($)

http://castatistics.wikispaces.com/file/view/Smokers+Project.jmp

(2007 data collected by CA Students. Sources include CDC and US Census Bureau)

a.  Run a test of slope using Median Income to predict Life Expectancy (years)

b.  Interpret the slope and intercept

c.  Use your model to predict life expectancy for a Median Income of $40,000

d.  Is your model a useful predictive model? Justify your answer.

e.  So, to live longer we just need to earn more money (and not be a teacher), right? Justify your answer.

3.  Crying and IQ

Did you cry much as a baby? Ask your parents.

A paper was published in 1964 that investigated if there was a link between how much a baby cried and their IQ. Babies were stimulated to provoke crying (i.e. they were pinched) and their duration and volume of crying recorded. Their IQ was measured a few years later when it was possible to record it (you don’t give babies IQ tests). The data from that paper is here:

http://castatistics.wikispaces.com/file/view/Crying+and+IQ.jmp

The variables are:

·  Crying (not sure what the units are called, but it is a combination of duration and volume)

·  IQ (measured when the child was 5 years)

a.  Investigate if there is a link between how much a baby cries and their IQ. Make the crying the explanatory (independent) variable.

b.  Interpret the slope and intercept

c.  Predict the IQ of a baby who measured 28 on the crying scale.

d.  Comment on the appropriateness of the model in a.

Samuel Karelitz et al., "Relation of crying activity in early infancy to speech and intellectual development age three years," Child Development, 35 (1964), pp. 769-777.