SPSS Review

Frequency

number of cases at a particular values

Percent

the frequency as a percent of ALL cases

Valid percent

the frequency as a percent of all cases not missing not missing on that variable — typically most reported when reporting frequency distributions

Cumulative percent

the percent of cases at or below a particular value. May be useful with an ordinal or interval variable

Pie charts

used to represent the distribution of a nominal variable

𝝁 (mu)

mean of variable of interest in a population

x̄ (x bar)

mean of variable of interest in a sample

nominal variables

unordered categories

ordinal variables

ordered categories

bar charts

used to represent the distribution of an ordinal variable

interval variables

number variables with no true zero (ex. temperature)

ratio variables

number variables with a true zero (ex. age)

histogram

demonstrates the spread of an interval or ratio variable

dichotomous variable

variable with two options (ex. yes or no)

look at percentages or counts when comparing dichotomous variables on a crosstabulation?

percentages

Single population mean

testing the mean of one population against a specific number (ex. I believe all men in Richmond are 6 feet tall)

null hypothesis

the assumption that the speculated average is correct in a population test (ex. null hypothesis: mu for Richmond men=6 ft)

alternative hypothesis

challenge to the assumption that the speculated average is not correct in a population test (ex. alt hypothesis: mu for Richmond men <6 ft or mu for Richmond men DNE 6 feet

Use the 1 sided p value

If alternative hypothesis is directional (greater than/less than)

Use the 2 sided p value

If alternative hypothesis is non-directional (not equal to)

95% confidence interval

range constructed for the true value of mu, centered on x bar

how to find traditional confidence interval using of the difference confidence intervals from SPSS?

add each number to test value (6 feet)

difference of means hypothesis test

compares two means of some variable from two different populations

which row of p value do you look at and why?

Always second row - more conservative estimate

fail to reject the null if p value is

greater than 0.05

single population proportion hypothesis test

hypothesis test that measures the proportions of a single dichotomous variable

Chi square test

aka two proportion z -test: compares two proportions