Any help on the question below. I am really starting to panic now.
Thank you in advance.
I came across this question which best explain what a need in one of the forum, however it was not answered. Please assist based on the scenario below:
I have 3 variables (2 are independent and 1 is dependent). All three variables are measured with questionnaires consisting of multiple questions recorded on 5 point scale. I want to perform regression analysis. How do I go about doing that? My variables are explained below
Creativity = Q1 + Q2 + Q3 (Do I need to transform Q1-Q3 into single variable representing all three q's of creativity to run regression; if so how to do that?)
Climate = Q1 + Q2 + Q3 + Q4 + Q5 + Q6 + Q7 + Q8 (Q1-Q8 into single variable)
Motivation = Q1 + Q2 + Q3 (Q1-Q3 into single variable)
All the questions (Q1, Q2, ...) are measured on a 5 point scale, 1 for strongly disagree and 5 for strongly agree. I want to run below model.
Creativity = constant + β1 Climate + β2 Motivation + error
I am using excel
Any help on the question below. I am really starting to panic now.
Thank you in advance.
Not sure what your actual question is. If several items (e.g. q1, q2, q3)
jointly represent some construct (e.g. creativity), you can sum them up
in order to achieve a "creativity" score. You can perform some procedure
to check if items which are supposed to represent the same construct
are sufficiently interrelated, e.g. by calculating Cronbach's alpha.
Or does your question refer to how to perform a regression analysis
in Excel?
With kind regards
K.
Thank you for the reply. I have summed up the questions to get a single score for each construct. Thereafter I have performed regression on the summed up score.
I have attached the excel spreadsheet with the input data tab, Summed Data tab and output tab. May you please review and advise if i have done it correctly?
Please advise?
I'm afraid that I can't do it, since I am not accustomed
with Excel's statstical analyses.
With kind regards
K.
Mosima (05-14-2015)
Your regression model finds no significant linear relationship: the estimated coefficients of the two independent variables are not significantly different from zero, and adjusted R-squared is actually negative. Some software would just truncate this at zero. I've attached a file with the correlation matrix of the individual question scores. (This created with an add-in available here: http://regressit.com.) There are significant negative correlations between components of the same construct in a few cases. In particular, CLQ1 and CLQ2 have a correlation of -0.448, CLQ5 and CLQ7 have a correlation of -0.653, and MTQ2 and MTQ3 have a correlation of -0.446. Insofar as the construct score is defined as the sum of the relevant question scores, it seems as though you would expect the question scores to be positively correlated with each other. The fact that some of the correlations are negative suggests that the questions are not really measuring the same construct.
Mosima (05-14-2015)
Noted. Thanks Schalke
Thanks a lot a Fritz. it means I will need to review my questionnaire.
Tweet |