8 Paired samples T-Tests
This activity introduces the paired samples t-test. It is also known as the ‘within participants’ or ‘related’ t-test. It is used when your design is within participants, i.e. where participants contribute data for the dependent variable in all conditions. If participants were to contribute to two conditions, for example an experimental and control condition, then a paired samples t-test is used.
Remember that there are two variants of the t-test:
paired samples t-tests.
Data files are set up differently according to which t-test is chosen. In addition there are separate menus in SPSS for each type of t-test. Finally, each t-test will produce different outputs.
In this activity you will examine data from a training evaluation known as pre-post design. A pre-post design measures how effective a training course can be by taking a base line measure before the training commences (pre). Training then takes place. Another reading is then taken (post) by the same instrument used during the baseline measurement. Comparisons are then made between the pre and post data.
The actual dataset is similar to that found in a study by Grant (2007). He looked at the improvement in coaching skills and emotional intelligence in a group of students training to become life skills and personal coaches. Grant assessed these variables using:
Goal-focused Coaching Skills Questionnaire (GCSQ);
Shuttle Emotional Intelligence Scale (SEIS).
The experimental hypothesis for this activity is that:
There will be a significant increase in coaching skills (as measured by the GCSQ) following the coaching training.
Once again, it is important to remember that there will be one independent variable – with two conditions, for example:
pre and post;
experimental and control.
Please observe that the same participants are providing the data for each condition. This is a requirement for using the paired samples t-test.