Reflection on Session 4

In today's session, I learnt about using SCLEQ as an instrument to take measurable data from a learning environment. SCLEQ will factor in areas like Student support, Affiliation, Professional interest, achievement orientation, staff freedom, participatory decision making, Innovation, Resource adequecy and work pressure. Comparing SCLEQ with Moo's schema (Personal Development, Relationships, System change) most of the items will fall under "Relationships". After using an instrument like SCLEQ, it is important for us to give meaning the data that we have gathered. One software that is able to do this is the SPSS software that can be used to help us analyse data.

Cronbach's alpha mean
Basically in order to give meaning to any of the data gathered, we should first test for Cronbach's alpha mean (alpha), then check to see if the items (questions) are unidimensional (measuring only one thing) using the "Factor" command. Next, group the items that are seemingly measuring the same dimension and test for reliability (alpha) again. A high reliability between items indicates correlation. A final test using the "correlations" command should confirm the correlation between items. Items with high correlation should be combined into the same scale.

We should note that if Cronbach's alpha mean is 0.70 or above, it means that the data for the items(questions) being measured is reliable. If it is o.70 or below, it implies that for all the items measured, the data is not reliable.

T-test
T-test is what is used for pre-post kind of data. It helps to measure the change in scores. It is especially useful when we have a lot of confound variables in the experiment/research.

How do we use the T-test:
First we need to look at the mean of the pre-data, mean of the post-data and the difference between them. Once we have some kind of intuitive meaning to the data , we can start to apply a T-test to check if what we intuitively think can be confirmed statistically by testing whether the change between the pre and post data is statistically significant.

When applying the T-test, we follow the typical statistical test procedures but with SPSS it should be much easier:
Step 1: State your Hypothesis/ Null Hypothesis
Step 2: State any underlying assumptions
Step 3: Run the paired sample statistics
Step 4: Run the paired sample correlation
Step 5: Run the paired sample test against hypothesis
Step 6: Run the paired sample test against null hypothesis
Step 7: Interpret data

What I have written so far is what was gathered from the internet about Cronbach's Alpha mean and T-test.