In elinw/lehmansociology: Tools for Sociology Students Using R

In this template we are going to use our two random samples (previously created and saved) and the statkey web application to create confidence intervals.

When we take random samples of a population, we have seen that most of the time the estimates are pretty good, sometimes they are terrible, and sometimes they are almost exactly right. We have seen this thanks to working with an artifiical situation in which we know the populaton values. However, this is not a real world situation. Usually we do not and cannot know what the population values are. We are using random samples to estimate the population values and we are uncertain whether we have one of the pretty good estimates, one of the great estimates or one of the really terrible ones.

Because of this uncertainty, when we report our results we generally want to report a range of values called a confidence interval rather than a single value. We also want to link our range to a level of confidence. So when you report results you should have three parts: + the calculated value (called the point estimate) + the confidence interval (called the interval estimate) + the level of confidence that is somewhere between 0 and 1 (or 0 and 100 if you put it in terms of percentages)

You can create confidence intervals for any statistic but in this labe we will focus on univariate statistics: the mean, median and standard deviation.

To start off with this we are not going to use R, we are going to use an online application called (statkey)[http://lock5stat.com/statkey/bootstrap_1_quant/bootstrap_1_quant.html]

To use statkey we need a column of data. Let's start with PCTPOVALL_2013.

To make a column of data you can do the following code.

Please note, knit is not going to work for this right now.

# You may need to reload your sample data set, in that case, uncomment the line below.
load(file = "data/mysample1.RData")
write.table(mysample1$PCTPOVALL_2013, file="data/mysample1poverty.txt", 
            row.names = FALSE)

Look in your data folder and you should see the new file. Click on it to open. You should see a column of numbers, a header and nothing else.

Repeat the same procedure for mysample2.

Go to the Statkey site.
Click on the edit data button and replace the data that is in the file with the numbers from your file.

Statkey uses a "boostrap" or "resampling" approach to estimating confidence intervals.

Following the instructions in Blackboard, usee Statkey to create bootstrap confidence intervals for your first sample.

Answer the questions in Blackboard.

elinw/lehmansociology documentation built on May 16, 2019, 3 a.m.