If I understand the question, it's the same questionnaire, conducted before and after an intervention. The analysis would be an ANOVA with a fixed factor intervention, two levels: before and after.
Dear Ankur,
I do not necessarily understand the term continuous linear scale.
See definitions:
Linear Scale : Definition and Meaning A linear scale is a line which is divided into equal parts. It is used on maps to show the relationship between a particular distance on a map and the actual distance on the earth. A linear scale is also called a bar scale, scale bar or graphic scale.
What is CONTINUOUS RATING SCALE?
a measure upon which positions are allocated along a continuum instead of in accordance with classes. These kind of measures are produced by making a record on the measure to show the position of the measure or by allocating a numeric value to show the degree of reaction.
CONTINUOUS RATING SCALE: "Continuous rating scales are praised for giving raters more options and not limiting the available responses like other methods would, such as likert scales."
Psychology Dictionary: What is CONTINUOUS RATING SCALE? definition of CONTINUOUS RATING SCALE (Psychology Dictionary)
Can Likert Scale Data ever be Continuous?
by Karen Grace-Martin
A very common question is whether it is legitimate to use Likert scale data in parametric statistical procedures that require interval data, such as Linear Regression, ANOVA, and Factor Analysis. A typical Likert scale item has 5 to 11 points that indicate the degree of agreement with a statement, such as 1=Strongly Agree to 5=Strongly Disagree. It can be a 1 to 5 scale, 0 to 10, etc.
The issue is that despite being made up of numbers, a Likert scale item is in fact a set of ordered categories.
One camp maintains that as ordered categories, the intervals between the scale values are not equal. Any mean, correlation, or other numerical operation applied to them is invalid. Only nonparametic statistics should be used on Likert scale data (i.e. Jamieson, 2004).
The other group maintains that while technically the Likert scale item is ordered, using it in parametric tests IS valid in some situations. For example, Lubke & Muthen (2004) found that it is possible to find true parameter values in factor analysis with Likert scale data, if assumptions about skewness, number of categories, etc., were met. Likewise, Glass et al. (1972) found that F tests in ANOVA could return accurate p-values on Likert items under certain conditions.
Pre-post test is a valid and useful experimental design. However life happens to people during a couple of months, it can affect us but this is the risk we have to take and we control for this by means of statistics
I agree that we need to know more about the scale that you are using as well as your overall analysis goals.
As an alternative to ANOVA, you might also consider doing regression where your dependent variable is the time-2 score and your independent variables are the time-1 score, plus any other variables that you want to control.
Morgan,
I do not understand your reasoning with regression analysis in a pre-post situation
post as DV and pre as IV?
Yes, time-2 is the "post measure." In repeated measures ANOVA, the analysis is based on simply subtracting the pre from the post, whereas regression uses the coefficient from the IV to predict the degree of difference (i.e., it does not assume that the slope equals 1.0).
How much do you expect then that is explained of the variance by the pretest?
Ankur,
In terms of pre and post assessments if you are doing a paired sampled t-test you will be able to gather evidence related to the questions you present to the respondents. The limitations of any pre and post test structure related to not being able to draw individual thought because most pre and post assessments are structured through a series of questions. I recommend that you think about the outcomes or the hypothesis of the research to define why research methodology you will implement.
Morgan,
Can you give me an example how you write the result?
When doing this kind of two-wave analysis, you just report it the way you would any other regression, with a recognition that most of the variance in the post-measure will be accounted for by the pre-measure.
Here is a technical explanation difference between the regression and the repeated-measures approaches. If you imagine the simplest regression equation, where the pre measure is the only predictor, then you could subtract that from both sides of the equation. On the left hand side you would have the pre-post difference, which is what is analyzed in a repeated-measures t-test (or ANOVA). On the right hand side, all that would left is the intercept, which would be the mean value of the difference, so testing its significance would give you exactly the same thing as the repeated measures t-test. In other words, if you eliminate the b coefficient (slope) from the regression, what you have left is the repeated measures test.
OKey Morgan,
Thank you very much for your informative answer.
"It is better to appear stupid during 5 min than to remain stupid for the rest of your life "
- Badges
- Science method
- Similar topics
- Methodology
- Research Methodology
- Qualitative Research