RMB: T-TESTS WEEK 5

Question 1

What is a t-test?

Answer 1

• A test comparing two means, looking for difference
• A test involving two groups > one group and a population
• Looks at the difference between the two groups
• T-tests are a parametric measure

Question 2

Types of t-tests

Answer 2

• Unrelated/independent groups/between groups: comparing different groups. E.g. is there a difference between male and female scores on a numerical ability test?
• Related/repeated measures/within groups: comparing the same group. E.g. does recall improve after attending “Hypnotic Memory Training” with Paul McKenna?

Question 3

Conditions needed for a t-test

Answer 3

• Parametric assumptions should be met
• Requires interval or ratio data + should be normally distributed (for both t-test types)
• In an independent t-test, the scores are independent and they have homogeneity of variance

Question 4

Performing a t-test

Answer 4

• First assume the null hyp is true
• If we were to take many samples from the two populations each would have a mean > only need mean to do a t-test
• If we plot the distribution formed by all of the difference means we would obtain the sampling distribution of the difference between means
• T-tests basically relate the difference between OUR sampling means to what the sampling distribution should look like If the null hyp is true > if mean difference is 0 then the null hypothesis is correct, but if there is a difference the alt hyp is correct
• What is the probability of your results being part of the null hyp graph of means > if the difference is far at either end, then it is significant (5% or less)
• The associated probability estimate tells us whether or not our difference mean falls within the most extreme 5% of difference means that could be expected
• SPSS generates a t-statistic – t basically tells us how many standard deviations our difference is from zero in the middle of the theoretical distribution.
• When performing the calculations by hand we would then need to compare this in a statistical table, which would generate a probability statistic.
• SPSS however generates a p-value for us however, in order to reject the null hypothesis p-value needs to be less than 0.05. p < 0.05

Question 5

One sample t-test

Answer 5

• One sample t-tests compare your sample mean w/ a known given population mean > when we know the mean of a population
• e.g. are 1st year psychologists more intelligent than the norm? Average mean IQ is 100 and mean IQ of first year psychologists = 117

Question 6

Reporting a t-test in APA format

Answer 6

• To report a one-sample t test in APA format you need to know the df, t value and p value
  t(df) = t-value, p = p-value
• e.g T = 5.131 df = 9 P value = .001
  So our result would be reported as:
  T(9) = 5.131, p = 0.001 > our p value is significant
• To find significance of statistical value, focus on the P value and if this is less or more than significance level of 0.05