Tests Flashcards
Null hypothesis
No change or no difference, this is the arrival or opposite view of what the researcher believes and what they want to test
One sample t-test
Calculates the difference between a sample mean and the known mean of a population.
Null hypothesis is no difference between sample and population means
Alternate is there is a significant difference between sample and population means
Example
The average BMI of Detroit, to see if it is significantly different from Michigan’s overall average BMI
Two sample t-test
Compare the main differences between two different samples, these are two independent samples
Matched or paired t-test
Testa to compute if there is a statistical difference between two sets of data that are paired together, example is pre-and post test
Null hypothesis is no difference between means of paired samples
Alternate is paired sample means are not equal
Example
Pair a pre-test average math score and a post test average math score
One way ANOVA
A test to compare three or more samples to one another, uses F distribution
Chi Square goodness of fit test
Measures the difference between what is observed any sample and what is expected
Chi Square test
Used for two categorical variables. Measures if any of deserved a difference in data set is due to chance
The larger the result is, the larger the difference. If it’s close to zero there is little or no difference
Null hypothesis is variables are same as expected
Alternate is that observed is different than expected
Example
According to a data set, 20% of students regularly smoke, 30% smoke socially, and 50% never smoke. Students at MSU responded to a survey saying 22% of them smoke regularly, 15% smoke socially and 63% never smoke
Type one error
When we reject the nbull hypothesis when it is true
Type two error
Fail to reject the null hypothesis when it is false
P value
Probability of reserving the obtained data given the null hypothesis is true
Compare of P value with the Alpha level chosen, usually 0.05 or 95% confidence
Hypothesis testing procedures, five
- Make both hypothesis, choose alpha level
- Collect sample data, select test statistic
- Set of decision rule
Can either compare P values with alpha, Or compare observed Ziva value with critical Z value - Compute test statistic
- Draw conclusion and summarize significance
When we use Z value versus T value
When N is greater or equal to 30 we use Z score
When N is less than 30 we use T distribution
One tailed test versus two-tailed test
One tailed
Opera tailed or lower tailed
Example, Cholesterol mean is greater than 203
Two tailed
Uses both upper and lower tailed sections, left and right side of graph
Example, cholesterol mean is not equal to 203, meaning it is greater than or lower than 203
