How to install packages when you know the name

require(PACKAGE.NAME)

 epicalc

Package used to determine reliability of groups of items

foreign

package to read data entered in other statistical packages, including SPSS

lmtest

package to conduct linear regression analyses

nortest

package used to test data for normal distributions.

outliers

package used to test for the presence of outliers in a dataset

PASWR

package used to conduct additional statistical analyses not in the base package (eg., z-tests).

citation()

package used to get information about the developer of R

citation("NAME")

used to get information about the developer of a package

GUI preferences

at the bottom of the Edit menu,  used to configure the GUI (graphical user interface) 

What not to do when creating a variable

Do not start your variable name with a number

Do not name your variable something that is a command in R

Don't forget variable names are case sensitive

what to do when creating a variable

 An option for entering data in R

The syntax to create a variable 

name=c(x1,x2,x3,...)

After you enter your syntax, press Enter

If you have entered the syntax properly a > sign will appear

to be sure you have entered your data properly simply type the name of the variable

(it will look like this)  [1] x1 x2 x3 

Another option for entering data is scan()

 name=scan() 

if you use space the data will look like this

1: 4 1 7 7 9

if you use enter the data will look like this

1: 4

2: 1

3: 7

4: 7

name=c(name, x1,x2,x3)

command you use to enter additional dat after you've created your original variable.

NA

what you enter if someone is missing a value for a particular measure in your data set

.txt or .cvs

file extension used for files you want to use in R

Save as Tab Delimited Text (.txt)

or

Comma Separated Values (.cvs)

to import data to R

First 

use the File menu option  > Change Dir 

Before importing data

Create a working file where you data is located. It is similar to creating a variable. Name it something simple like x or one

to import a file

x=read.table(file="FILENAME.txt",header=T)

x is the name of your working file

read.table tells R you want to import a data file from another program

file="FILENAME.txt" change the FILENAME.txt to your file name.

header=T indicate first row of table has data

x=read.table(file="FILENAME.cvs",header=T,sep=",")

To use spreadsheets that aren't in excel, the syntax is similar with two exceptions:

1 the file name now ends in .csv rather than .txt 

and

2 the sep="," syntax at the end. This tells R that the file you stored your data in uses commas to indicate separate values of data

Again x is your name of working file and your file's name for FILENAME

>attach(NAME)

attach tells R you want to use a dataset entered in another program as if the data were entered directly into R.

t.test

to run a t-test if we attach our data syntax is

t.test(variable1~variable2,var.equal=T)

if we do not attach data syntax is

t.test(x[['variable1']]~[['variable2']],var.equal=T)

by using attach command we can use the 1st t.test syntax because R knows where the data is located

data.entry(NAME)

How to edit data that has been directly entered into R

data.entry- opens a spreadsheet containing the data stored

(NAME) - specifies which variable you want to edit

fix(NAME)

opens up a test file with your data if you want to change an entry but only one varibale can be changed using the

fix command

if you realize you have made a mistake entering your data into another spreadsheet, it is strongly recommended you edit the spreadsheet and reimport the data into R

car

package required to recode data

Recoding data

Is used when you need to use reverse scoring.

require(car)

command to load car package for use

NEWVARNAME=recode(OLDVARNAME,"recodes")

NEWVARNAME - creates new variable that contains the reversed scores

recode - tells R that you want to create different values using information contained in parentheses

OLDVARNAME - the actual name of your variable you are reversing

"recodes" - contains the information on how to recode your original data in OLDVARNAME

NEWVARNAME=recode(OLDVARNAME,"0='3';1='2';2='1';3='0'"

syntax for recode is contained inside one set of quotation marks-"", each pair of data is separated by a semicolon-;, and original number followed by equal sign-=, then new value is enclosed with single quote marks-'

Categorial Variables

for instance depressed and non depressed

total=c(x1+x2+x3+x4 . . . x20)

to creat a variable that totals scores

lo

hi

else

for categorical variable

tells R to start recording range at the lowest, observed score

for instance 

categoryname=recode(totalname,"lo:16='0';17:hi-'1'")

this will transform all scores in the totalname into 0's and 1's and store in categoryname

lo

hi

else

categoryname=recode(totalname,"lo:16='Not Depressed';17:hi='Depressed'")

follows the same format as reverse scoring. 

name=scan()

name - is what you want to call your variable.

When you want to enter data without haveing to enter comma to separate values. 

hist(NAME)

command to make a histograms

brk=c(x50,x55,x60,x65,x70,x75,x80,x85,x90,x95,x100)

 1.to create a variable that contains each of the values you

want on the x-axis¹.

 2.to save time you could use the command 

brk=c(seq(50,100,5))

 seq function tells R to crate a sequence of numbers using the information in the ( ). 1st and 2nd numbers are min and max values and final number tells how far apart you want the sequence to be. 

hist(test.pct,breaks=brk)

Command you can use once you crate your break variable.

hist(test.pct,breaks=brk)

Once you create your break bariable this is the syntax you enter to tell R that you want a different set of values on the x-axix, and tells R where to find these values.

freq

to change the data reported y-axis

freq=T

Tells R that you want the absolute frequencies listed on the y-axis

freq=F

This tells R you want the density (not the frequency)

hist(test.pct,breaks=brk,freq=T)

syntax to put absolute values on the y-axis

hist(test.pct,breaks=brk,freq=F)

syntax used to tell R that you want the density on the

y-axis

boxplot(test.pct)

to create a boxplot simply use the command boxplot followed by the name of the variable containing the data we wish to look at. 

qqnorm(test.pct)

another option to check normality of your datais the Q-Q plot. This plot involves plotting observed vs. theoretical quanities on the x-and y-axes (hence Q-Q)

plot(Y~X)

plot -  tells R to create a scatterplot

Y - is the outcome variable (replace Y with name of the variable where your outcome data is stored)

X - is the predicator variable (replace X with the name of the variable wher your predictor is stored)

summary(NAME)

This option will provide a wide range of information 

It will provide you with the

Min and Max scores 

Mean and Median

1^stQu and 3^rdQu (1st and 3rd Quartiles) 

mean(NAME)

tells R to create a histogram with the data of a given variable

(NAME) specifies the variable to be analyzed (you will need to change NAME to the name of the variable)

median(NAME)

tells R to report the middle score for a given variable

(NAME) specifies the variable to be analyzed (you will need to change NAME to the name of the variable)

var(NAME)

this is the syntax used to produce information about the variance

(This is information that is not included in the output produced by the summary opiton)

sd(NAME)

this is the syntax used to produce the information about the standard deviation

(This is information that is not included in the output produced by the summary opiton)

na.rm=T

  command used to handle missing data points, used in conjuction with the 

mean(NAME,na.rm=T)

or

sd(NAME,na.rm=T)

tapply(NAME,group.var,desc.stat)

tapply - tells R that you want informaton from a variable broken into groups

NAME - the name of the variable you want decriptive information on (replace with actual name of the variable where the data is stored)

group.var - the name of the variable containing the group information(replace with actual name...)

desc.stat - the specific descriptive statistic you want reported(replace with actual name...)

F or f

it is recommended if you ever want to create a factor for a variable you should use F as in FNAME or f as in fsex.

FNAME=factor(Name,labels=c("label1","label2"))

FNAME - is the factor name for a give numeric variable(replace with actual factor name)

factor - tells R you want to create a factor for a variable

NAME - specifies where the numeric data is(are) from (replace with actual numerical variable name)

labels=c - tells R that you are defining the labels for each numerical value.

"label1","label2" - labels for each numerical value (replace with your desired labels.)

cor(x,y)

syntac to calculate Pearson's r is fount in the stats package 

cor - tells R to report th ecorrelation coeffieient between 2 numeric variables

x - is the predictor variable (use actual name of your predictor variable)

y - is the outcome variable (use actual name of your outcome variable)

cor.test(x,y)

to find t-, p-value and additional information

lm(y~x)

 This is the syntax used to find linear regression

the main difference is cor.test is replaced with 

lm - tells R that we are interested in constructing a linear mofel for our data.

although x and y have the same meaning in this syntax, they are entered in reverse order and separated by a ~ (tilde)

lm command is located in the stats package. 

summary(lm(y~x)

Used to get a more complete report of your analysis

At the bottom of output will be

F-statistic: x.xxx on X and X DF, p-value:  z.zzzz

x,X, and z are really numbers

p-value

Remember if your p-value is lower thatn .05, the model is a good fit. 

If higher than .05 then one should accept the null hypothesis